View in EDS HTML Full Text PDF Full Text

Teaching Bias? Relations between Teaching Quality and Classroom Demographic Composition

Saved in:
Bibliographic Details
Title: Teaching Bias? Relations between Teaching Quality and Classroom Demographic Composition
Language: English
Authors: Cherng, Hua-Yu Sebastian, Halpin, Peter F., Rodriguez, Luis A.
Source: American Journal of Education. Feb 2022 128(2):171-201.
Availability: University of Chicago Press. Journals Division, P.O. Box 37005, Chicago, IL 60637. Tel: 877-705-1878; Tel: 773-753-3347; Fax: 877-705-1879; Fax: 773-753-0811; e-mail: subscriptions@press.uchicago.edu; Web site: http://www.journals.uchicago.edu/journals/aje/about
Peer Reviewed: Y
Page Count: 31
Publication Date: 2022
Document Type: Journal Articles
Reports - Research
Descriptors: Teacher Effectiveness, Teacher Characteristics, Minority Group Students, Racial Composition, Educational Quality, Racial Differences, Ethnicity, African American Students, Hispanic American Students
DOI: 10.1086/717676
ISSN: 0195-6744
Abstract: Purpose: Prior work has drawn consistent conclusions about systematic racial disparities in the allocation of high-quality teachers in US public schools, such as classrooms with more students of color having teachers with fewer credentials and less experience. However, these fixed characteristics of teachers are only proxies for the quality of teaching, which may vary within teacher by the different classrooms they teach. Research Methods/Approach: Using data from the Measures of Effective Teaching (MET) project, we consider various sources of within-teacher, across-classroom variation of teaching effectiveness using a teacher fixed-effects modeling approach and highlight those that may, on average, disadvantage youth of color. Findings: We find that (1) about half of the variation in classroom teaching efficacy is within teachers, (2) classrooms taught by the same teacher with higher percentages of Black and Latinx students receive lower quality of teaching, and (3) these patterns are consistent across teacher racial/ethnic groups. A number of plausible explanations of this association are considered, including rater biases on observational measures and differences in teaching practices; we also consider how these variations differ by teacher race. Implications: Our findings highlight the importance of teachers' in-classroom practices, not simply their credentials, in educational research, policy, and practice targeted at reducing racial inequality.
Abstractor: As Provided
Entry Date: 2022
Accession Number: EJ1325802
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwG_5eMykXc4TpYxVUfSo1CGAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDKceKomNy9FVgArwmQIBEICBm04_QGe-PQxGC2x44Ie5JClNQuOwB3x-fsUwxFrGyTQAKWjD3v5obcihvpTwDpMrzLVky59kL0YX9VBIdYplmHAHFLIZvtYy2gK-Xl-zAToe9yr3s9obM-5t_s04TZ4gqaNoiBOFg-nGZJePA4A8G1oa0hsd2_VgfnxmsU5eGqOWXf_771JMtHGnJpXHRhGFskmwMdpnoGeV-GZQ
Text:
  Availability: 1
  Value: <anid>AN0155256492;jrd01feb.22;2022Feb17.02:06;v2.2.500</anid> <title id="AN0155256492-1">Teaching Bias? Relations between Teaching Quality and Classroom Demographic Composition </title> <p>Purpose: Prior work has drawn consistent conclusions about systematic racial disparities in the allocation of high-quality teachers in US public schools, such as classrooms with more students of color having teachers with fewer credentials and less experience. However, these fixed characteristics of teachers are only proxies for the quality of teaching, which may vary within teacher by the different classrooms they teach. Research Methods/Approach: Using data from the Measures of Effective Teaching (MET) project, we consider various sources of within-teacher, across-classroom variation of teaching effectiveness using a teacher fixed-effects modeling approach and highlight those that may, on average, disadvantage youth of color. Findings: We find that (<reflink idref="bib1" id="ref1">1</reflink>) about half of the variation in classroom teaching efficacy is within teachers, (<reflink idref="bib2" id="ref2">2</reflink>) classrooms taught by the same teacher with higher percentages of Black and Latinx students receive lower quality of teaching, and (<reflink idref="bib3" id="ref3">3</reflink>) these patterns are consistent across teacher racial/ethnic groups. A number of plausible explanations of this association are considered, including rater biases on observational measures and differences in teaching practices; we also consider how these variations differ by teacher race. Implications: Our findings highlight the importance of teachers' in-classroom practices, not simply their credentials, in educational research, policy, and practice targeted at reducing racial inequality.</p> <p>A range of studies across a variety of contexts have drawn consistent conclusions about systematic racial disparities in the allocation of high-quality teachers in American public schools (Borman and Kimball [<reflink idref="bib7" id="ref4">7</reflink>]; Clotfelter et al. [<reflink idref="bib17" id="ref5">17</reflink>]; Freeman et al. [<reflink idref="bib35" id="ref6">35</reflink>]; Goldhaber et al. [<reflink idref="bib42" id="ref7">42</reflink>]; Jackson [<reflink idref="bib56" id="ref8">56</reflink>]; Jerald and Ingersoll [<reflink idref="bib57" id="ref9">57</reflink>]; Kalogrides et al. [<reflink idref="bib58" id="ref10">58</reflink>]; Lankford et al. [<reflink idref="bib69" id="ref11">69</reflink>]; Tennessee Department of Education [<reflink idref="bib96" id="ref12">96</reflink>]). Data released by the US Department of Education's Office for Civil Rights in 2011 affirm the national scope of a prominent racial/ethnic educational inequality: classrooms with more students of color are taught by less qualified teachers (US Department of Education Office for Civil Rights [<reflink idref="bib97" id="ref13">97</reflink>]). Across the literature, these analyses involved a number of indicators of teacher quality, including teaching experience, level of education, and salary. Every indicator led to the same conclusion: youth of color are more likely to have lower-quality teachers. This finding has been consistently affirmed by prior and subsequent research on US public schools, which shows that less qualified teachers are systematically selected into districts, schools, and classrooms with larger proportions of students of color (Clotfelter et al. [<reflink idref="bib17" id="ref14">17</reflink>]; Goldhaber et al. [<reflink idref="bib42" id="ref15">42</reflink>]). Although highlighting an important racial inequality in the allocation of teacher human resources, a primary limitation of this work has been the reliance on teacher qualifications (e.g., credentials, years of work) as proxies for the quality of teaching experienced by students. Several studies have found that teacher qualifications are tenuously linked to students' academic outcomes, with teacher subject certification serving as the most viable predictor of student learning outcomes (Barnett [<reflink idref="bib5" id="ref16">5</reflink>]; Buddin and Zamarro [<reflink idref="bib10" id="ref17">10</reflink>]; Early et al. [<reflink idref="bib27" id="ref18">27</reflink>], [<reflink idref="bib28" id="ref19">28</reflink>]; Goldhaber and Brewer [<reflink idref="bib43" id="ref20">43</reflink>]; Hill et al. [<reflink idref="bib51" id="ref21">51</reflink>]; Kane et al. [<reflink idref="bib61" id="ref22">61</reflink>]; Preston [<reflink idref="bib82" id="ref23">82</reflink>]; Rivkin et al. [<reflink idref="bib84" id="ref24">84</reflink>]; Smith [<reflink idref="bib89" id="ref25">89</reflink>]; Zeichner [<reflink idref="bib102" id="ref26">102</reflink>]). Moreover, qualifications are relatively static characteristics of teachers that fail to capture classroom-to-classroom variation in the quality of teaching—the primary responsibility of teachers—exhibited by the same teacher. This latter point is especially relevant in light of research showing that the effectiveness of teachers varies over classrooms (Ho and Kane [<reflink idref="bib52" id="ref27">52</reflink>]; Kane and Staiger [<reflink idref="bib62" id="ref28">62</reflink>]).</p> <p>Scholars studying teacher quality have sought to address this by utilizing a construct of teaching quality as defined by value-added measures (VAMs) estimated through student test score growth. The wave of research exploring variation in teaching effectiveness as estimated through VAMs has consistently found that teaching quality is highly variable and inequitably distributed across school environments (Steele et al. [<reflink idref="bib92" id="ref29">92</reflink>]) as well as within schools across classrooms that differ by the socioeconomic and demographic composition of students (Goldhaber et al. [<reflink idref="bib42" id="ref30">42</reflink>]; Isenberg et al. [<reflink idref="bib55" id="ref31">55</reflink>]). In addition to VAM, classroom observations have become increasingly prevalent in the assessment of teaching effectiveness in the public education system. Primarily propagated by state efforts to bid for the federal Race to the Top grant competitions beginning in 2010, several states integrated both the use of VAMs and annual classroom observations for the purpose of facilitating teacher development and accountability (National Council on Teacher Quality [<reflink idref="bib78" id="ref32">78</reflink>]).</p> <p>However, measures of teaching effectiveness are also subject to sources of variation occurring over classrooms taught by the same teacher. This within-teacher variation can be from classroom to classroom within the same year, or from year to year. Stated another way, current conceptualizations of "effective" teaching do not adequately acknowledge the contextual nature of teaching, which differs across classrooms. There is also reason to suspect that teaching may vary across a teacher's classroom in ways that, on average, disadvantage youth of color. Prior research finds that teachers are biased in their perceptions and judgments of non-White versus White students (Cherng [<reflink idref="bib13" id="ref33">13</reflink>]; Ladson-Billings [<reflink idref="bib68" id="ref34">68</reflink>]; Riegle-Crumb and Humphries [<reflink idref="bib83" id="ref35">83</reflink>]). Therefore, the question remains whether teachers teach differently when working with classrooms with more students of color.</p> <p>The present article contributes to our understanding of these issues by studying two relatively direct data sources on teaching quality, namely in-classroom observations of teachers' practices and student growth on annual state exams. We focus on different possible channels by which within-teacher variation can be reflected in measures of teaching effectiveness and focus on how the racial/ethnic composition of classrooms is related to these measures of teaching quality using the Measures of Effective Teaching (MET) database, a large but nonrepresentative sample of US public school teachers from 2010 to 2012.</p> <hd id="AN0155256492-2">Background</hd> <p>The lack of consensus on the importance of teacher qualifications for student outcomes has important implications for the study of racial bias in the allocation of teachers—if teacher qualifications are not strongly and consistently related to student performance, findings concerning the distribution of teacher qualifications may overstate the severity and consequences of racial inequalities. The literature on teacher value added largely finds that conventional measures of teacher qualifications are not strongly related to estimated teaching effectiveness. For instance, new teachers are usually associated with lower value-added estimates of teaching effectiveness than their more experienced counterparts, but the returns to teacher experience taper off in later years (Ladd and Sorensen [<reflink idref="bib65" id="ref36">65</reflink>]; Papay and Kraft [<reflink idref="bib80" id="ref37">80</reflink>]). Furthermore, on average, teachers with master's degrees have similar estimated effects to teachers with only bachelor's degrees and, likewise, teachers with alternative certification have similar estimated teaching effectiveness as traditionally certified teachers (Aaronson et al. [<reflink idref="bib1" id="ref38">1</reflink>]; Buddin and Zamarro [<reflink idref="bib10" id="ref39">10</reflink>]; Kane et al. [<reflink idref="bib61" id="ref40">61</reflink>]; Koedel and Betts [<reflink idref="bib63" id="ref41">63</reflink>]).</p> <p>To address the concern with regard to the low predictive power of teacher qualifications on teaching effectiveness, recent research on teacher quality has increasingly focused on measures of effective teaching, rather than teacher qualifications alone (see, e.g., Kane and Staiger [<reflink idref="bib62" id="ref42">62</reflink>]). The effects of teaching are consistent evidence that teachers' VAMs are predictive of their students' annual gains on state exams in later years (Garrett and Steinberg [<reflink idref="bib36" id="ref43">36</reflink>]; Wenglinsky [<reflink idref="bib100" id="ref44">100</reflink>]). Less well studied are measures of teaching practice that use in-classroom observations. Unlike classroom "checklists" used in the past, newly developed measures of classroom practices are deeply rooted in theoretical perspectives of teaching quality, and they have been shown to predict student gains on state exams when the observations are conducted by trained research staff (Danielson [<reflink idref="bib20" id="ref45">20</reflink>]; Pianta and Hamre [<reflink idref="bib81" id="ref46">81</reflink>]).</p> <p>Focusing on teaching effectiveness rather than static credentials also captures an important aspect of teaching practice: that teachers may strategically vary their teaching practice across classrooms. Variation in a teacher's instructional quality over classrooms is well documented (Brimijoin [<reflink idref="bib8" id="ref47">8</reflink>]; Cohen and Hill [<reflink idref="bib18" id="ref48">18</reflink>]). Teachers are explicitly trained to adapt, or differentiate, their training based on student need and background (Darling-Hammond [<reflink idref="bib21" id="ref49">21</reflink>]; Snow et al. [<reflink idref="bib90" id="ref50">90</reflink>]; Spillane [<reflink idref="bib91" id="ref51">91</reflink>]). Research has also found that adaptive instruction is linked with academic gains, with effect sizes ranging from 0.22 to 0.45 (Brühwiler and Blatchford [<reflink idref="bib9" id="ref52">9</reflink>]; Fraser et al. [<reflink idref="bib34" id="ref53">34</reflink>]; Scheerens and Bosker [<reflink idref="bib86" id="ref54">86</reflink>]). Measures of teaching effectiveness likely also vary across classrooms taught by the same teacher (representing within-teacher variation). It is useful to consider different sources of this variation, including the potential role of classroom composition. Below, we highlight multiple channels by which within-teacher variation can be reflected in measures of teaching effectiveness, one of which is the student racial composition of a teacher's classrooms—the primary concern of our study.</p> <hd id="AN0155256492-3">Within-Teacher Variation in Estimated Teaching Effectiveness Due to Random Noise</hd> <p>One potential explanation of within-teacher variation in VAMs and classroom observations is simply random "noise" that affects a whole class of students simultaneously—a rowdy kid that disturbs their peers or disruptive noise on the day of an exam, an outbreak of the flu the week of the state test, and so forth. Such variation is unsystematic and therefore not associated with characteristics of teachers or students. This explanation is often implicit in the teaching effectiveness literature, where the teacher is the primary unit of analysis and other sources of variation are considered extraneous. Research suggests that a sizable proportion of the within-teacher variance in VAMs is indeed random noise. For example, the MET project sponsored by the Bill and Melinda Gates Foundation reported reliabilities ranging from 0.18 to 0.38 for VAM estimates obtained for two different course sections taught in the same year to different groups of students and 0.20–0.40 for estimates obtained between sections across years (Kane and Cantrell [<reflink idref="bib59" id="ref55">59</reflink>]). As implied by between-section and between-year correlations well below 0.50, more than half of the observed variation in teacher value added is associated with nonpersistent differences between teachers. McCaffrey and colleagues ([<reflink idref="bib74" id="ref56">74</reflink>]) similarly reported that about 30%–60% of the intertemporal variation of VAMs was due to sampling error, depending on grade, subject, and district. However, of the remaining systematic variation, 30%–54% was within teachers; moreover, this variation was not explainable in terms of observable time-varying teacher covariates like experience, attainment of advanced degrees, or in-service training.</p> <p>Research on in-classroom observational measures has led to similar conclusions, but with some notable differences. Whitehurst and colleagues ([<reflink idref="bib101" id="ref57">101</reflink>]) found higher intertemporal consistency among year-to-year classroom observations (0.65) compared with year-to-year VAM (0.38), particularly when conducted by administrators from within teachers' schools. This may be due, in part, to school administrators developing preconceived notions about a teacher's effectiveness. For example, if a principal has a negative perception toward a particular teacher, that teacher may receive a lower observation score than the teacher would have received if the principal were unfamiliar with them prior to the observation. Nevertheless, previous studies based on the MET project suggest that up to 70% of the variation in these measures is within teachers, with remaining systematic sources of variation including lessons, classroom sections, and raters (Ho and Kane [<reflink idref="bib52" id="ref58">52</reflink>]; Kane and Staiger [<reflink idref="bib62" id="ref59">62</reflink>]). Taken altogether, it is reasonable to conclude that random noise is not a sufficient explanatory factor of within-teacher variation in observation- and test-score-based measures of teaching effectiveness, which begs the question of other possible contributors, which we enumerate below.</p> <hd id="AN0155256492-4">Within-Teacher Variation in Estimated Teaching Effectiveness Associated with Student Classroo...</hd> <p>Estimated teaching effectiveness can vary based on the demographics of students served across particular classes. The question becomes whether teaching effectiveness differs across different demographic student subgroups net of academic ability. One recent paper using data from the MET study found classrooms with higher densities of Black students received lower teaching scores (as measured using the Classroom Assessment Scoring System, or CLASS)—notably, the findings were drawn from comparisons made across both teachers and classrooms (Osei-Twumasi and Pinetta [<reflink idref="bib79" id="ref60">79</reflink>]). A modest literature has found suggestive evidence that value-added estimates of teaching effectiveness, which condition on students' prior achievement, do not vary considerably by student subgroup (Fox [<reflink idref="bib32" id="ref61">32</reflink>]; Loeb et al. [<reflink idref="bib71" id="ref62">71</reflink>]). By estimating value added specific to subgroups of students based on gender, ability, race, and free lunch status, Fox ([<reflink idref="bib32" id="ref63">32</reflink>]) found that, across math and reading, correlations of estimated effectiveness within teacher by student subgroup were high (i.e., above 0.9). In a separate study, Loeb and colleagues ([<reflink idref="bib71" id="ref64">71</reflink>]) similarly found that teachers' estimated teaching effectiveness was stable when estimated specifically for English learner subgroups. However, the possibility of sorting student subgroups to classrooms based on unobservable student traits may have attenuated the estimated differential effectiveness for student subgroups. Thus far, no study has been able to definitively rule out the possibility of attenuation bias associated with student sorting.</p> <p>Nonetheless, disparities in estimated teaching effectiveness based on classroom composition may arise due to various factors unrelated to differences in student ability. First, differences in estimated teaching effectiveness based on classroom composition may be reflective of actual differences in teaching practice or teachers' demographic-based biases. Furthermore, the demographic similarity between a teacher and their students (i.e., "race match") may also influence the degree to which teachers alter teaching practice or exhibit bias against students based on their demographic backgrounds. Second, in the case of observation-based measures of teaching effectiveness, rater biases may artificially deviate teaching effectiveness based on the composition of students across classrooms. We expand upon these points in greater detail below.</p> <hd id="AN0155256492-5">Differentiation in Teaching Practice</hd> <p>Differentiation in teaching practice serves as a likely mechanism for observed differences in a teacher's ability to effectively instruct students based on their demographic backgrounds. For example, in an attempt to meet the needs of a culturally and linguistically diverse student demographic, both across and within classrooms, teachers are often encouraged and expected to implement differentiated instruction (Mayer et al. [<reflink idref="bib72" id="ref65">72</reflink>]). Differentiation, often taking the form of varied classroom management procedures, is designed to promote student engagement and motivation, grouping practices, and assessments, and it provides yet another channel by which teaching effectiveness can vary across classroom contexts (Waitoller and Artiles [<reflink idref="bib99" id="ref66">99</reflink>]). Furthermore, movements to enact culturally responsive pedagogy within classrooms to better affirm student's cultural identities in support of their thinking process continues to gain traction in the field (Gay [<reflink idref="bib38" id="ref67">38</reflink>]). The extent to which a teacher is equipped to work with students from diverse sets of cultural and linguistic backgrounds and the ability of that teacher to effectively deliver quality instruction may vary across classroom contexts composed of students from different demographic backgrounds—at times, to the detriment to students of color.</p> <hd id="AN0155256492-6">Teacher Bias</hd> <p>In addition, teaching effectiveness may vary across classroom contexts in ways that are unrelated to deliberate differentiation in instruction, but rather due to bias—conscious or unconscious—against students of color that ultimately minimizes their opportunity to learn. Prior research reveals that teachers often have lower academic perceptions of Black and Latinx students compared with their White peers (Ferguson [<reflink idref="bib31" id="ref68">31</reflink>]; Riegle-Crumb and Humphries [<reflink idref="bib83" id="ref69">83</reflink>]), are less likely to recommend students of color to advanced classes and programs (Borman and Dowling [<reflink idref="bib6" id="ref70">6</reflink>]; Haller [<reflink idref="bib48" id="ref71">48</reflink>]; Tenenbaum and Ruck [<reflink idref="bib95" id="ref72">95</reflink>]), and make more severe discipline referrals of Black, Latinx, and Indigenous students (Gregory et al. [<reflink idref="bib46" id="ref73">46</reflink>]; Skiba et al. [<reflink idref="bib88" id="ref74">88</reflink>]).</p> <hd id="AN0155256492-7">Teacher-Student Race Match</hd> <p>Within the body of research that examines teacher-student racial/ethnic dynamics, a number of studies focus on the notion of "race matching" and how a teacher's racial/ethnic background interplays with the race/ethnicity of their students to affect the learning process such that students perform better academically from working with teachers of the same race or ethnicity (Grissom et al. [<reflink idref="bib47" id="ref75">47</reflink>]; Simpson and Erickson [<reflink idref="bib87" id="ref76">87</reflink>]). Scholars examining student-teacher race matching continue to explore the underlying mechanisms for this pattern and have found that race matching is associated with improved teacher perceptions of student academic ability (Fox [<reflink idref="bib33" id="ref77">33</reflink>]; Gershenson et al. [<reflink idref="bib41" id="ref78">41</reflink>], [<reflink idref="bib40" id="ref79">40</reflink>]), teacher perceptions of their relationships with students (Saft and Pianta [<reflink idref="bib85" id="ref80">85</reflink>]), and student achievement (Dee [<reflink idref="bib22" id="ref81">22</reflink>], [<reflink idref="bib23" id="ref82">23</reflink>]; Egalite et al. [<reflink idref="bib29" id="ref83">29</reflink>]; Ehrenberg et al. [<reflink idref="bib30" id="ref84">30</reflink>]). In light of the preponderance of evidence that a teacher's racial background may shape their perceptions of and engagement with students of marginalized racial/ethnic backgrounds, it therefore may be the case that teaching may be particularly effective in classrooms where the race/ethnicity of teachers and students are the same.</p> <hd id="AN0155256492-8">Observation Measures of Teaching Effectiveness and Rater Bias</hd> <p>The incorporation of classroom observations for high-stakes accountability of teaching effectiveness has stimulated increased debate about the extent to which observers (raters) provide a valid rating of the observed (teachers in the classroom setting), and whether ratings are susceptible to trends and bias in rater behavior. Variation in rater's understanding and implementation of rubric scoring can become more severe over time and, thus, provide one potential source of error in a single teacher's observation ratings. Commonly referred to as "rater drift," prior research affirms dynamic changes in rater severity when assessing instructional quality and suggests that drift is particularly large during a rater's initial days of observation and can persist up through 2 years of scoring (Casabianca et al. [<reflink idref="bib12" id="ref85">12</reflink>]; Congdon and McQueen [<reflink idref="bib19" id="ref86">19</reflink>]).</p> <p>Beyond examining the presence of rater drift, a number of studies have investigated how rater bias may influence their final assessment of instructional quality. More specifically, evidence across multiple studies indicates that the incoming achievement level as well as the racial/ethnic composition of a teacher's class tends to significantly and substantively influence their classroom observation scores (Campbell and Ronfeldt [<reflink idref="bib11" id="ref87">11</reflink>]; Steinberg and Garrett [<reflink idref="bib93" id="ref88">93</reflink>]; Whitehurst et al. [<reflink idref="bib101" id="ref89">101</reflink>]). Three possible explanations arise in relation to teachers systematically receiving lower observation ratings when teaching particular kinds of students based on prior achievement. The first explanation is that less effective teachers may be nonrandomly assigned to certain students, particularly those who are initially low performing. The second explanation entails low-performing students being more challenging to teach, thus making it difficult for a single teacher to perform at the same level when working with higher-performing children—a point to which we return to below. The third explanation involves rater bias, particularly bias that is a function of the composition of students within a teacher's classroom.</p> <p>Recent work sheds light on the extent to which systematic differences in classroom observations related to student characteristics are driven by nonrandom sorting to teachers of students versus possible rater bias. A secondary data analysis of the MET data revealed that the subsample of teachers randomly assigned to work with higher-achieving students received higher observation ratings, above and beyond what might be attributable to time-invariant aspects of teacher quality (Steinberg and Garrett [<reflink idref="bib93" id="ref90">93</reflink>]). Campbell and Ronfeldt similarly leverage the randomization of teachers under the MET study and have found that teachers' observation scores are related to the sociodemographic makeup of their students. Specifically, teachers assigned to teach high concentrations of Black and Hispanic students received significantly lower observation ratings. The authors suggest that rater biases likely explain this pattern, as "differences are unlikely due to actual differences in teaching quality" ([<reflink idref="bib11" id="ref91">11</reflink>], 1233). Taken together, due to the random assignment of teachers to students in the second year of the MET study, these studies affirm that differences in observation ratings associated with student characteristics are unrelated to nonrandom sorting of less effective teachers to low-performing students and may be possibly driven by observation rater bias.</p> <hd id="AN0155256492-9">Research Questions and Scope of the Present Study</hd> <p>Prior work that examines the inequitable distribution of teacher human resources across student racial/ethnic groups relies on measures of teacher quality that are static (i.e., are assumed not to vary within teachers) and that are often tenuously linked with student outcomes. As a result, we may be missing a large portion of the picture of how students of color are differentially exposed to quality teaching practices. This omission is particularly glaring given that prior work finds that teachers often perceive and interact with students of color in ways that may compound their disadvantage.</p> <p>We reviewed literature on possible sources of within-teacher variation in teaching effectiveness and seek to test those that are particularly relevant for student racial/ethnic inequities. We ask two research questions that determine whether classroom variations in teaching effectiveness are due to (<reflink idref="bib1" id="ref92">1</reflink>) random noise or (<reflink idref="bib2" id="ref93">2</reflink>) classroom demographic composition. Furthermore, we test whether associations between teaching effectiveness and classroom demographic composition are related to student prior performance due to academic tracking, teacher-student race matching, and rater bias.</p> <p></p> <ulist> <item> 1. What proportion of the overall classroom-to-classroom variation in teaching effectiveness is due to differences among teachers, as opposed to differences among classrooms taught by the same teacher?</item> <p></p> <item> 2. Is this classroom-to-classroom and within-teacher variations associated with classroom racial/ethnic composition?</item> <p></p> </ulist> <p>• a.</p> <p></p> <ulist> <item> Does this association persist after controlling for other plausible sources of classroom-to-classroom variation, notably prior academic performance (i.e., tracking)?</item> <p></p> </ulist> <p>• b.</p> <p></p> <ulist> <item> Is this association linked to teacher-student race matches?</item> <p></p> </ulist> <p>• c.</p> <p></p> <ulist> <item> Does this association depend on the type of outcome measure used, and, in particular, is it observed when the outcome is based on VAM rather than observational measures that may be subject to rater effects?</item> </ulist> <p>We address these research questions utilizing data from the MET project, which we describe in more detail in the following section. Prior work also generated from this data source has revealed how rated teaching effectiveness is associated with the demographic composition of students taught by a teacher; however, these empirical studies do not adequately account for differences associated with teacher characteristics (Osei-Twumasi and Pinetta [<reflink idref="bib79" id="ref94">79</reflink>]), nor do they explore whether similar patterns arise with test-based measures of teaching effectiveness (Campbell and Ronfeldt [<reflink idref="bib11" id="ref95">11</reflink>]). In this regard, the secondary analysis of the MET project data presented herein marks an attempt to enhance prior work methodologically and analytically by employing different analytic methods to address these two limitations.</p> <hd id="AN0155256492-10">Data and Methods</hd> <p>The MET study was conducted during the academic years of 2009–10 and 2010–11 and collected data on 2,756 teachers in 317 schools and 6 US school districts. The focus of the present analysis is English language arts (ELA) and mathematics teachers in grades 4 through 9. We excluded 245 ninth-grade biology teachers because data were not available on most (79.32%) of their students' previous achievement in science. Of the remaining 2,515 teachers, 1,000 (39.88%) taught ELA only, 895 (35.58%) taught math only, and 616 (24.49%) were subject matter generalists who taught both subjects. Nearly all generalists (98.37%) taught grades 4 and 5.[<reflink idref="bib2" id="ref96">2</reflink>]</p> <p>Table A1 summarizes the number of classroom sections that were available for analysis with type of teacher. It should be noted that the MET study randomly assigned students to teachers in the second year of the study; however, the randomization design only allowed comparison between teachers within randomization blocks and not within. Therefore, we were unable to utilize this randomization in a more causal design. Our final sample excluded teachers where we only observe one of their classrooms as we are unable to examine within teachers across their classroom sections. We also conducted our analyses using a pooled sample across time for two reasons. First, although there was a passage of time across the 2-year study, it is unlikely that growth in teaching over time is associated with classroom demographics, which is one of the focal points of our study. Second, using a pooled sample allows us to preserve 20% of our classrooms that have variations in classroom but in the same year.[<reflink idref="bib3" id="ref97">3</reflink>]</p> <hd id="AN0155256492-11">Measures</hd> <p>Our focal outcome, teaching quality, was measured in three ways. Two of these were in-classroom observational rating rubrics for evaluating teaching practices, the Framework for Teaching (FFT; Danielson [<reflink idref="bib20" id="ref98">20</reflink>]) and CLASS (Pianta and Hamre [<reflink idref="bib81" id="ref99">81</reflink>]). Table A2 summarizes the versions of the rubrics used in the MET study, and table A3 describes the rating protocol and summarizes the number of ratings per class section for the teachers considered in this study.</p> <p>In the present analysis, all rating data were first aggregated (i.e., averaged) to the classroom level. We obtained a total score for each classroom section using an unweighted average over the items (8 for FFT, 12 for CLASS). Using Cronbach's alpha, the reliability of the classroom-level total scores was estimated to be 0.92 and 0.94 for FFT and CLASS, respectively, and in both subjects. To facilitate interpretability, the total scores were standardized to have a mean of zero and standard deviation of one in the sample. Higher values reflect better teaching practices.</p> <p>Our third outcome variable was students' gain scores on standardized state exams, in math and ELA, which were reestimated without student race/ethnicity as a covariate. Many approaches to estimating VAMs have been discussed in the literature (Lockwood and McCaffrey [<reflink idref="bib70" id="ref100">70</reflink>]; McCaffrey et al. [<reflink idref="bib73" id="ref101">73</reflink>]; Morganstein and Wasserstein [<reflink idref="bib77" id="ref102">77</reflink>]). In the present study, we estimated teacher value added using the following linear regression model using ordinary least squares (OLS):</p> <p> <ephtml> <math display="block" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable displaystyle="true"><mlabeledtr><mtd><mtext>(1)</mtext></mtd><mtd><mrow><msub><mi>Y</mi><mrow><mi>i</mi><mi>j</mi><mi>k</mi></mrow></msub><mo>=</mo><msup><mi>β</mi><mo>′</mo></msup><msub><mi>X</mi><mrow><mi>i</mi><mi>j</mi><mi>k</mi></mrow></msub><mo>+</mo><msup><mi>γ</mi><mo>′</mo></msup><msub><mi>Z</mi><mrow><mi>i</mi><mi>j</mi><mi>k</mi></mrow></msub><mo>+</mo><msub><mi>δ</mi><mrow><mi>j</mi><mi>k</mi></mrow></msub><mo>+</mo><msub><mi>ϵ</mi><mrow><mi>i</mi><mi>j</mi><mi>k</mi><mtext>,</mtext></mrow></msub></mrow></mtd></mlabeledtr></mtable></mrow></math> </ephtml> </p> <p>where</p> <p> <emph>Y<subs>ijk</subs></emph> is the standardized test score for student <emph>i</emph> in classroom <emph>j</emph> of teacher <emph>k</emph>,</p> <p> <emph>X<subs>ijk</subs></emph> are the student-level district administrative covariates,</p> <p> <emph>Z<subs>ijk</subs></emph> are the student-level prior year math and ELA state exams scores,</p> <p>δ<subs><emph>jk</emph></subs> is the fixed effect of classroom section <emph>j</emph> of teacher <emph>k</emph>, and</p> <p>ε<subs><emph>ijk</emph></subs> is a residual term.</p> <p>Equation (<reflink idref="bib1" id="ref103">1</reflink>) was estimated separately for each combination of academic year (2009–10, 2010–11) by subject (math and ELA), by grade (4 through 9), and by district. In addition, within year-subject-grade-district combination, the outcome variable was scaled to have a mean of zero and variance of one. The district administrative covariates used in our model included students' age, gender, English-language learner (ELL) status, special education status, gifted status, free and reduced lunch status, and the students' previous year performance in both subjects. Note that we did not control for students' race when computing gain scores due to our interest in whether student gains were associated with the racial composition of classrooms.</p> <hd id="AN0155256492-12">Key covariates</hd> <p>We include independent variables that correspond to our second research question. First, to consider that tracking may occur due to nonrandom sorting based on observed student achievement, we include prior achievement in math and ELA, as measured by district-administered math and ELA tests in the previous school year. Second, to examine associations between teaching effectiveness and student demographics, we include covariates that describe classroom student racial/ethnic composition: percentage Asian American, percentage Black, percentage Latinx, and percentage Other. Third, to investigate the importance of teacher-student race matching, we include variables representing the teacher's race (self-reported as "White," "Black," "Hispanic," or "Other"; Other, which we exclude due to limitations in sample size, includes teachers identifying as non-White, Black, or Hispanic).</p> <hd id="AN0155256492-13">Control variables</hd> <p>We also included several additional classroom-level covariates as statistical controls that may be linked to teaching effectiveness: percentage male, percentage free/reduced lunch (as a proxy measure for classroom socioeconomic status), percentage special education, percentage emergent bilingual (ELLs), percentage in gifted program, average age, and the number of students in classroom. We also include controls for grade level, coded as "primary" (grades 4 and 5), "middle" (grades 6–8), and "high" (grade 9). For analyses that involved either FFT or CLASS, we additionally included a teacher-reported dummy variable that reflects the focal topic covered during the video-recorded sessions used for rating. For the sake of parsimony, we do not show coefficients for grade level and focal topic, as results do not vary significantly by either of these variables. Table 1 summarizes the outcome variables and covariates used in the study.</p> <p>Table 1. Descriptive Statistics of All Variables Used in Analyses</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom">Mean/Proportion</td><td valign="bottom">SD</td></tr></thead><tbody><tr><td valign="bottom">Outcome variables:</td></tr><tr><td valign="bottom"> FFT scale</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">.81</td></tr><tr><td valign="bottom"> CLASS scale</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">.81</td></tr><tr><td valign="bottom"> Math value added</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">.27</td></tr><tr><td valign="bottom"> English language arts (ELA) value added</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">.21</td></tr><tr><td valign="bottom">Covariates and controls:</td></tr><tr><td valign="bottom"> Math classroom characteristics:</td></tr><tr><td valign="bottom"> Percentage Asian</td><td valign="bottom" char=".">.06</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Black</td><td valign="bottom" char=".">.32</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Latino</td><td valign="bottom" char=".">.33</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Other</td><td valign="bottom" char=".">.03</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage White</td><td valign="bottom" char=".">.26</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage male</td><td valign="bottom" char=".">.50</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage special education</td><td valign="bottom" char=".">.08</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage free/reduced lunch</td><td valign="bottom" char=".">.56</td><td valign="bottom" /></tr><tr><td valign="bottom"> Number of students in classroom</td><td valign="bottom" char=".">24.59</td><td valign="bottom" char=".">6.81</td></tr><tr><td valign="bottom"> Prior math achievement (2009)</td><td valign="bottom" char=".">.03</td><td valign="bottom" char=".">.60</td></tr><tr><td valign="bottom"> Grade level: 4</td><td valign="bottom" char=".">.11</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 5</td><td valign="bottom" char=".">.12</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 6</td><td valign="bottom" char=".">.21</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 7</td><td valign="bottom" char=".">.19</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 8</td><td valign="bottom" char=".">.17</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 9</td><td valign="bottom" char=".">.20</td><td valign="bottom" /></tr><tr><td valign="bottom"> ELA classroom characteristics:</td></tr><tr><td valign="bottom"> Percentage Asian</td><td valign="bottom" char=".">.06</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Black</td><td valign="bottom" char=".">.33</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Latino</td><td valign="bottom" char=".">.33</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage Other</td><td valign="bottom" char=".">.03</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage White</td><td valign="bottom" char=".">.26</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage male</td><td valign="bottom" char=".">.50</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage special education</td><td valign="bottom" char=".">.08</td><td valign="bottom" /></tr><tr><td valign="bottom"> Percentage free/reduced lunch</td><td valign="bottom" char=".">.56</td><td valign="bottom" /></tr><tr><td valign="bottom"> Number of students in classroom</td><td valign="bottom" char=".">24.55</td><td valign="bottom" char=".">6.93</td></tr><tr><td valign="bottom"> Prior ELA achievement (2009)</td><td valign="bottom" char=".">.02</td><td valign="bottom" char=".">.60</td></tr><tr><td valign="bottom"> Grade level: 4</td><td valign="bottom" char=".">.12</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 5</td><td valign="bottom" char=".">.14</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 6</td><td valign="bottom" char=".">.20</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 7</td><td valign="bottom" char=".">.17</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 8</td><td valign="bottom" char=".">.17</td><td valign="bottom" /></tr><tr><td valign="bottom"> Grade level: 9</td><td valign="bottom" char=".">.20</td><td valign="bottom" /></tr><tr><td valign="bottom"> Teacher race/ethnicity:</td></tr><tr><td valign="bottom"> White</td><td valign="bottom" char=".">.63</td><td valign="bottom" /></tr><tr><td valign="bottom"> Black</td><td valign="bottom" char=".">.31</td><td valign="bottom" /></tr><tr><td valign="bottom"> Latinx</td><td valign="bottom" char=".">.06</td><td valign="bottom" /></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>1 Note. CLASS = Classroom Assessment Scoring System; FFT = Framework for Teaching.</p> <hd id="AN0155256492-14">Analytic Strategy</hd> <p></p> <hd id="AN0155256492-15">Research question 1</hd> <p>We begin with two purely descriptive analyses. Because most research works to date have focused exclusively on between-teacher variation, we first report the proportions of variance in each outcome variable that are between versus within teachers. Variance components are estimated using a simple random-effects model without predictors. This analysis provides an initial indication of the extent to which within-teacher variation may further contribute to our understanding of inequities in teaching quality.</p> <hd id="AN0155256492-16">Research question 2</hd> <p>Second, we consider the extent to which the within-teacher variation in the outcome measures is associated with within-teacher variation in the racial/ethnic composition of classrooms. For this analysis, we group-mean center all outcome and classroom composition variables at the teacher level, and then report Pearson correlations between each outcome and the composition variables. This provides an initial indication of whether teaching quality and classroom composition are indeed associated, after removing variation due to fixed characteristics of teachers.</p> <hd id="AN0155256492-17">Research question 2a</hd> <p>We report a number of models that are intended to disentangle the relative contributions of classroom racial composition from other classroom composition factors, and in particular prior academic achievement. We use a standard fixed-effects approach to control for static characteristics of the teacher, such as teacher credentials, and experience, which can help address issues of nonrandom sorting of teachers.</p> <hd id="AN0155256492-18">Research question 2b</hd> <p>To examine potential "race-matching effects," we modified our fixed-effects equation by interacting teacher race/ethnicity variables with classroom racial composition variables to obtain fixed-effects estimates (β) specific to each subgroup. We report <emph>F</emph>-tests for each subgroup by focal variable interaction term, as well as the point estimates of the fixed effects in each subgroup.</p> <hd id="AN0155256492-19">Research question 2c</hd> <p>We estimate VAM to address issues of rater bias, as VAM are predicted by our measures of teaching effectiveness (FFT and CLASS) but are not assigned by raters. To facilitate the interpretation of the fixed-effects models, we also standardize the fixed effects for each focal variable <emph>m</emph> = 1, ... , <emph>M</emph>. We compute the within-teacher standardized fixed effect as</p> <p> <ephtml> <math display="block" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable displaystyle="true"><mlabeledtr><mtd><mtext>(2)</mtext></mtd><mtd><mrow><msub><mi>d</mi><mi>m</mi></msub><mo>=</mo><mfrac><mrow><msub><mover accent="true"><mi>β</mi><mo>^</mo></mover><mi>m</mi></msub><mo>×</mo><mi>SD</mi><mrow><mo>(</mo><mrow><msub><mi>X</mi><mrow><mi>j</mi><mi>k</mi><mi>m</mi></mrow></msub><mo>−</mo><msub><mover accent="true"><mi>X</mi><mo>¯</mo></mover><mrow><mi>k</mi><mi>m</mi></mrow></msub></mrow><mo>)</mo></mrow></mrow><mrow><mi>SD</mi><mrow><mo>(</mo><mrow><msub><mi>Y</mi><mrow><mi>j</mi><mi>k</mi></mrow></msub><mo>−</mo><msub><mover accent="true"><mi>Y</mi><mo>¯</mo></mover><mi>k</mi></msub></mrow><mo>)</mo></mrow></mrow></mfrac><mtext>,</mtext></mrow></mtd></mlabeledtr></mtable></mrow></math> </ephtml> </p> <p>where SD denotes the sample standard deviation, which uses <emph>N</emph><subs>classrooms</subs> – <emph>N</emph><subs>teachers</subs> in the denominator for group-mean-centered variables. Equation (<reflink idref="bib2" id="ref104">2</reflink>) corresponds closely to Hedges and Hedberg's ([<reflink idref="bib49" id="ref105">49</reflink>]) within-level effect size for cluster-randomized trials. In the present context, it may be interpreted as the expected difference on the outcome variable, when a teacher's classroom composition on the focal variable is increased by one standard deviation unit. The difference is standardized in terms of within-teacher variation on the outcome. We also report coefficients standardized using the total variance of the outcome, and we compare the standardized coefficients of our focal variables with those of other variables that have been addressed in past research.</p> <hd id="AN0155256492-20">Results</hd> <p></p> <hd id="AN0155256492-21">Between- and Within-Teacher Variations in Teaching Effectiveness (Research Question 1)</hd> <p>Our first research question asks what proportion of the overall classroom-to-classroom variation in teaching effectiveness is due to differences among teachers, as opposed to differences among classrooms taught by the same teacher. We begin our descriptive analysis by presenting table 2, which shows the interclass correlations (ICC) for random-effects models estimating FFT and CLASS and value added for math and ELA. The ICC indicate the proportions of variance in each outcome variable that are between versus within teachers. Here, we see that a large proportion of variance—in some cases the majority—exists within teachers. For example, 50.4% of the variance in FFT scale scores in math is within teachers (with the rest existing between different teachers). Stated another way, the variation in the quality of teaching received by two classes of students taught by the same teacher is the same, if not slightly larger, as the variation in quality received by two classes taught by completely different teachers.</p> <p>Table 2. Interclass Correlations (ICC) for Fixed-Effects Regression Models in Table 3</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom">Between</td><td valign="bottom">Within</td></tr></thead><tbody><tr><td valign="bottom">FFT—math</td><td valign="bottom" char=".">.496</td><td valign="bottom" char=".">.504</td></tr><tr><td valign="bottom">FFT—ELA</td><td valign="bottom" char=".">.553</td><td valign="bottom" char=".">.447</td></tr><tr><td valign="bottom">CLASS—math</td><td valign="bottom" char=".">.505</td><td valign="bottom" char=".">.495</td></tr><tr><td valign="bottom">CLASS—ELA</td><td valign="bottom" char=".">.623</td><td valign="bottom" char=".">.377</td></tr><tr><td valign="bottom">Value added—math</td><td valign="bottom" char=".">.519</td><td valign="bottom" char=".">.481</td></tr><tr><td valign="bottom">Value added—ELA</td><td valign="bottom" char=".">.349</td><td valign="bottom" char=".">.651</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>2 Note. ELA = English language arts; FFT = Framework for Teaching.</p> <hd id="AN0155256492-22">Teaching Effectiveness and Classroom Composition (Research Question 2)</hd> <p>Next, we turn to our second research question, which focuses on whether classroom-to-classroom and within-teacher variations are linked with classroom compositional measures. We examine correlations between student racial/ethnic demographic composition of classrooms and observation-based ratings of teaching effectiveness and value added, which are shown in figure 1. From this figure, we see that the percentage of students who are White within a classroom is positively correlated with all outcomes (FFT math: 0.15, FFT ELA: 0.12, CLASS math: 0.09, CLASS ELA: 0.06, math value added: 0.15, and ELA value added: 0.16). In contrast, the percentage for every other racial/ethnic group is less positively correlated with stronger teaching effectiveness, in particular percentage Black and percentage Latinx, both of which are negatively correlated with teaching effectiveness and value-added scores.</p> <p>Graph: Fig. 1. Correlations among measures of teaching effectiveness and classroom racial demographic measures, within teacher. All variables were aggregated to the classroom level and centered on teacher means. CLASS = Classroom Assessment Scoring System; ELA = English language arts; FFT = Framework for Teaching.</p> <p>Continuing our examination of correlates of classroom-to-classroom and within-teacher variations in teaching effectiveness, and in particular how prior academic performance may shape results (research question 2a), we turn to table 3, which shows coefficients from teacher fixed-effects OLS regression models estimating FFT and CLASS scale scores. Two specifications estimate separately FFT math teaching scores and ELA teaching scores and two other specifications estimate separately CLASS math teaching scores and ELA teaching scores. All models include student racial/ethnic demographic composition: percentage Asian American, percentage Black, percentage Latinx, and percentage Other, prior achievement of the classroom, percentage male, percentage special education, percentage free/reduced lunch, number of students in classroom, and focal topics covered in the class (not shown). Overall, results show that the percentage Black is negatively and persistently associated with teaching effectiveness on both the math and ELA FFT and math CLASS measures (coefficients are −0.008, −0.009, and −0.008 in models 1a, 2b, and 1b, respectively), and percentage Latinx and percentage Other are negatively associated with the math FFT measure (−0.008 for percentage Latinx and −0.013 for percentage Other).[<reflink idref="bib4" id="ref106">4</reflink>]</p> <p>Table 3. Coefficients from Teacher Fixed-Effects Linear Regression Models Estimating FFT and CLASS Scale Scores</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom"><sc>FFT Scale</sc></td><td valign="bottom"><sc>CLASS Scale</sc></td></tr><tr><td valign="bottom">Math</td><td valign="bottom">ELA</td><td valign="bottom">Math</td><td valign="bottom">ELA</td></tr><tr><td valign="bottom">(1a)</td><td valign="bottom">(2a)</td><td valign="bottom">(1b)</td><td valign="bottom">(2b)</td></tr></thead><tbody><tr><td valign="bottom">Percentage Asian</td><td valign="bottom" char=".">−.004</td><td valign="bottom" char=".">−.006</td><td valign="bottom" char=".">−.003</td><td valign="bottom" char=".">−.007</td></tr><tr><td valign="bottom">Percentage Black</td><td valign="bottom" char=".">−.008<xref ref-type="table-fn" rid="tfn5">*</xref></td><td valign="bottom" char=".">−.009<xref ref-type="table-fn" rid="tfn6">**</xref></td><td valign="bottom" char=".">−.008<xref ref-type="table-fn" rid="tfn5">*</xref></td><td valign="bottom" char=".">−.004</td></tr><tr><td valign="bottom">Percentage Latinx</td><td valign="bottom" char=".">−.008<xref ref-type="table-fn" rid="tfn6">**</xref></td><td valign="bottom" char=".">−.004</td><td valign="bottom" char=".">−.004</td><td valign="bottom" char=".">0</td></tr><tr><td valign="bottom">Percentage Other</td><td valign="bottom" char=".">−.013<xref ref-type="table-fn" rid="tfn5">*</xref></td><td valign="bottom" char=".">−.011<xref ref-type="table-fn" rid="tfn4">+</xref></td><td valign="bottom" char=".">−.008</td><td valign="bottom" char=".">−.003</td></tr><tr><td valign="bottom">Percentage male</td><td valign="bottom" char=".">−.002</td><td valign="bottom" char=".">−.005<xref ref-type="table-fn" rid="tfn5">*</xref></td><td valign="bottom" char=".">−.002</td><td valign="bottom" char=".">−.003</td></tr><tr><td valign="bottom">Percentage free/reduced lunch</td><td valign="bottom" char=".">−.005<xref ref-type="table-fn" rid="tfn4">+</xref></td><td valign="bottom" char=".">−.001</td><td valign="bottom" char=".">−0</td><td valign="bottom" char=".">−.002</td></tr><tr><td valign="bottom">Percentage special education</td><td valign="bottom" char=".">−.002</td><td valign="bottom" char=".">.002</td><td valign="bottom" char=".">.001</td><td valign="bottom" char=".">.001</td></tr><tr><td valign="bottom">Percentage emergent bilingual</td><td valign="bottom" char=".">.002</td><td valign="bottom" char=".">−.002</td><td valign="bottom" char=".">.001</td><td valign="bottom" char=".">−.002</td></tr><tr><td valign="bottom">Percentage in gifted program</td><td valign="bottom" char=".">−.001</td><td valign="bottom" char=".">−.002</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">−.001</td></tr><tr><td valign="bottom">Average age</td><td valign="bottom" char=".">.028</td><td valign="bottom" char=".">.013</td><td valign="bottom" char=".">−.101<xref ref-type="table-fn" rid="tfn6">**</xref></td><td valign="bottom" char=".">−.032</td></tr><tr><td valign="bottom">Number of students in classroom</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">−.014<xref ref-type="table-fn" rid="tfn6">**</xref></td><td valign="bottom" char=".">−.004</td><td valign="bottom" char=".">−.006</td></tr><tr><td valign="bottom">Prior achievement: math</td><td valign="bottom" char=".">.179</td><td valign="bottom" char=".">−.115</td><td valign="bottom" char=".">.089</td><td valign="bottom" char=".">.056</td></tr><tr><td valign="bottom">Prior achievement: ELA</td><td valign="bottom" char=".">−.080</td><td valign="bottom" char=".">.300<xref ref-type="table-fn" rid="tfn5">*</xref></td><td valign="bottom" char=".">−.014</td><td valign="bottom" char=".">.075</td></tr><tr><td valign="top">Constant</td><td valign="bottom" char=".">.326</td><td valign="bottom" char=".">.854</td><td valign="bottom" char=".">1.816<xref ref-type="table-fn" rid="tfn6">**</xref></td><td valign="bottom" char=".">.857</td></tr><tr><td valign="bottom"> Observations</td><td valign="bottom">1,892</td><td valign="bottom">2,054</td><td valign="bottom">1,844</td><td valign="bottom">1,999</td></tr><tr><td valign="bottom"> Number of teachers</td><td valign="bottom">786</td><td valign="bottom">849</td><td valign="bottom">763</td><td valign="bottom">822</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <ulist> <item>3 Note. All models include controls for focal topic of class lesson and grade level (coefficients not shown). CLASS = Classroom Assessment Scoring System; ELA = English language arts; FFT = Framework for Teaching.</item> <item>4 + <emph>p</emph> <.10.</item> <item>5 * <emph>p</emph> <.05.</item> <item>6 ** <emph>p</emph> <.01.</item> </ulist> <p>Next, we turn to research question 2b, which asks whether the race/ethnicity of the teacher is linked with this finding—in other words, are there race-matching findings. Table 4 shows coefficients from fixed-effects linear regression models that estimate FFT and CLASS scale scores with interaction terms between the percentage racial demographic and teacher race/ethnicity. Model specification 1a estimates FFT math scale scores, specification 2a estimates FFT ELA scale scores, specification 1b estimates CLASS math scale scores, and 2b estimates CLASS ELA scale scores. Each model includes the same covariates as the full models in table 3. Overall, we find that of the 32 student-teacher race dyads, only 2 are statistically significant at the <emph>p</emph> <.05 level: they are a positive interaction term between the percentage Asian in a classroom and having a Latinx teacher for ELA FFT scores and a negative interaction term between percentage Latinx in a classroom and having a Latinx teacher for ELA CLASS scores.</p> <p>Table 4. Coefficients from Fixed-Effects Linear Regression Models Estimating FFT and CLASS Scale Scores with Teacher Race/Ethnicity Interactions</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom"><sc>FFT Scale</sc></td><td valign="bottom"><sc>CLASS Scale</sc></td></tr><tr><td valign="bottom">Math</td><td valign="bottom">ELA</td><td valign="bottom">Math</td><td valign="bottom">ELA</td></tr><tr><td valign="bottom">(1a)</td><td valign="bottom">(2a)</td><td valign="bottom">(1b)</td><td valign="bottom">(2b)</td></tr></thead><tbody><tr><td valign="bottom">Percentage Asian:</td></tr><tr><td valign="bottom"> × Black</td><td valign="bottom" char=".">−.012</td><td valign="bottom" char=".">.001</td><td valign="bottom" char=".">.007</td><td valign="bottom" char=".">−.018</td></tr><tr><td valign="bottom"> × Latinx</td><td valign="bottom" char=".">−.018</td><td valign="bottom" char=".">.039<xref ref-type="table-fn" rid="tfn9">*</xref></td><td valign="bottom" char=".">.004</td><td valign="bottom" char=".">.004</td></tr><tr><td valign="bottom">Percentage Black:</td></tr><tr><td valign="bottom"> × Black</td><td valign="bottom" char=".">−.006</td><td valign="bottom" char=".">−.010</td><td valign="bottom" char=".">.001</td><td valign="bottom" char=".">−.003</td></tr><tr><td valign="bottom"> × Latinx</td><td valign="bottom" char=".">−.017</td><td valign="bottom" char=".">−.018</td><td valign="bottom" char=".">−.001</td><td valign="bottom" char=".">−.028<xref ref-type="table-fn" rid="tfn8">+</xref></td></tr><tr><td valign="bottom">Percentage Latinx:</td></tr><tr><td valign="bottom"> × Black</td><td valign="bottom" char=".">−.009</td><td valign="bottom" char=".">−.007</td><td valign="bottom" char=".">.002</td><td valign="bottom" char=".">.001</td></tr><tr><td valign="bottom"> × Latinx</td><td valign="bottom" char=".">−.013</td><td valign="bottom" char=".">−.020<xref ref-type="table-fn" rid="tfn8">+</xref></td><td valign="bottom" char=".">.009</td><td valign="bottom" char=".">−.029<xref ref-type="table-fn" rid="tfn10">**</xref></td></tr><tr><td valign="bottom">Percentage Other:</td></tr><tr><td valign="bottom"> × Black</td><td valign="bottom" char=".">−.012</td><td valign="bottom" char=".">−.016</td><td valign="bottom" char=".">−.006</td><td valign="bottom" char=".">−.018</td></tr><tr><td valign="top"> × Latinx</td><td valign="bottom" char=".">−.038<xref ref-type="table-fn" rid="tfn8">+</xref></td><td valign="bottom" char=".">−.003</td><td valign="bottom" char=".">−.022</td><td valign="bottom" char=".">.013</td></tr><tr><td valign="bottom"> Observations</td><td valign="bottom">1,827</td><td valign="bottom">1,972</td><td valign="bottom">1,779</td><td valign="bottom">1,917</td></tr><tr><td valign="bottom"> Number of teachers</td><td valign="bottom">757</td><td valign="bottom">812</td><td valign="bottom">734</td><td valign="bottom">785</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <ulist> <item>7 Note. All models include controls found in table 3. All models include controls for focal topic of class lesson and grade level (coefficients not shown). CLASS = Classroom Assessment Scoring System; ELA = English language arts; FFT = Framework for Teaching.</item> <item>8 + <emph>p</emph> <.10.</item> <item>9 * <emph>p</emph> <.05.</item> <item>10 ** <emph>p</emph> <.01.</item> </ulist> <p>In an additional set of analyses not shown, we also find that patterns in table 4 are consistent across grade level.</p> <p>Turning to our last research question (2c), we seek to address whether these patterns may reflect biases of raters who assigned FFT and CLASS scores. We present table 5, which shows coefficients from analogous (to the prior table) fixed-effects regression models that estimate math (model 1) and ELA (model 2) value-added scores. Findings from table 5, which mirror findings from table 3, suggest that teachers and not raters may be exhibiting biases in teaching that, on average, disadvantage students of color. For example, the coefficients for percentage Black are negative and statistically significant in models estimating math (model 1: −0.002) and ELA (model 2: −0.001).</p> <p>Table 5. Coefficients from Fixed-Effects Linear Regression Models Estimating Math and English Language Arts Value Added</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom">Math</td><td valign="bottom">ELA</td></tr><tr><td valign="bottom">(1)</td><td valign="bottom">(2)</td></tr></thead><tbody><tr><td valign="bottom">Percentage Asian</td><td valign="bottom" char=".">.003<xref ref-type="table-fn" rid="tfn14">**</xref></td><td valign="bottom" char=".">0</td></tr><tr><td valign="bottom">Percentage Black</td><td valign="bottom" char=".">−.002<xref ref-type="table-fn" rid="tfn13">*</xref></td><td valign="bottom" char=".">−.001<xref ref-type="table-fn" rid="tfn12">+</xref></td></tr><tr><td valign="bottom">Percentage Latinx</td><td valign="bottom" char=".">−.001<xref ref-type="table-fn" rid="tfn12">+</xref></td><td valign="bottom" char=".">−0</td></tr><tr><td valign="bottom">Percentage Other</td><td valign="bottom" char=".">−.003<xref ref-type="table-fn" rid="tfn14">**</xref></td><td valign="bottom" char=".">.002<xref ref-type="table-fn" rid="tfn12">+</xref></td></tr><tr><td valign="bottom">Percentage male</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">−.001<xref ref-type="table-fn" rid="tfn13">*</xref></td></tr><tr><td valign="bottom">Percentage free/reduced lunch</td><td valign="bottom" char=".">.001</td><td valign="bottom" char=".">0</td></tr><tr><td valign="bottom">Percentage special education</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">−.001<xref ref-type="table-fn" rid="tfn13">*</xref></td></tr><tr><td valign="bottom">Percentage emergent bilingual</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">0</td></tr><tr><td valign="bottom">Percentage in gifted program</td><td valign="bottom" char=".">0</td><td valign="bottom" char=".">.001</td></tr><tr><td valign="bottom">Average age</td><td valign="bottom" char=".">.035<xref ref-type="table-fn" rid="tfn14">**</xref></td><td valign="bottom" char=".">.027<xref ref-type="table-fn" rid="tfn14">**</xref></td></tr><tr><td valign="bottom">Number of students in classroom</td><td valign="bottom" char=".">−.003<xref ref-type="table-fn" rid="tfn14">**</xref></td><td valign="bottom" char=".">−.002<xref ref-type="table-fn" rid="tfn13">*</xref></td></tr><tr><td valign="bottom">Prior achievement: math</td><td valign="bottom" char=".">.082<xref ref-type="table-fn" rid="tfn14">**</xref></td><td valign="bottom" char=".">.061<xref ref-type="table-fn" rid="tfn13">*</xref></td></tr><tr><td valign="bottom">Prior achievement: ELA</td><td valign="bottom" char=".">.008</td><td valign="bottom" char=".">.025</td></tr><tr><td valign="top">Constant</td><td valign="bottom" char=".">−.240<xref ref-type="table-fn" rid="tfn13">*</xref></td><td valign="bottom" char=".">−.189<xref ref-type="table-fn" rid="tfn12">+</xref></td></tr><tr><td valign="bottom"> Observations</td><td valign="bottom">2,565</td><td valign="bottom">2,886</td></tr><tr><td valign="bottom"> Number of teachers</td><td valign="bottom">1,059</td><td valign="bottom">1,183</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <ulist> <item>11 Note. All models include controls for focal topic of class lesson and grade level (coefficients not shown). ELA = English language arts.</item> <item>12 <sups>+</sups><emph>p</emph> <.10.</item> <item>13 * <emph>p</emph> <.05.</item> <item>14 ** <emph>p</emph> <.01.</item> </ulist> <p>To complement our findings, it is important to interpret the magnitude of the coefficients with respect to the within-teacher variability on both the focal predictors and outcomes. For this purpose, we refer to table 6, which provides the standardized coefficients described in equation (<reflink idref="bib2" id="ref107">2</reflink>). To get a sense of the relative magnitude of these coefficients for our focal variables, we additionally include standardized coefficients for the prior year on-subject performance, which Steinberg and Garrett ([<reflink idref="bib93" id="ref108">93</reflink>]) found to predict a substantial portion of between-teacher variation on FFT. Overall, we find that the magnitude of the relationships with percentage Black is comparable to, or in some cases larger than, the standardized coefficients for prior achievement. For example, the standardized coefficient for the percentage Black and FFT in Math is −0.112, which is about the same magnitude, but in the opposite direction, as the coefficient for prior year math achievement in the same model (0.099). As another example, the standardized coefficient for the percentage Black and FFT in ELA is −0.12, which is about the same magnitude, but in the opposite direction, as the coefficient for prior achievement in ELA in the same model (0.17). These findings from table 6 again suggest that the racial composition of classrooms is an important factor when examining teaching quality.</p> <p>Table 6. Within-Teacher and Total Effect Sizes, Standardized within Teachers</p> <p> <ephtml> <table><thead><tr><td valign="bottom" /><td valign="bottom" /><td valign="bottom">Within Teacher</td><td valign="bottom">Total</td></tr></thead><tbody><tr><td valign="bottom">Percentage Black:</td></tr><tr><td valign="top"> Math</td><td valign="bottom">FFT</td><td valign="bottom" char=".">−.112</td><td valign="bottom" char=".">−.059</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">−.113</td><td valign="bottom" char=".">−.053</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">−.089</td><td valign="bottom" char=".">−.044</td></tr><tr><td valign="top"> English</td><td valign="bottom">FFT</td><td valign="bottom" char=".">−.122</td><td valign="bottom" char=".">−.065</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">−.063</td><td valign="bottom" char=".">−.030</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">−.053</td><td valign="bottom" char=".">−.032</td></tr><tr><td valign="bottom">Percentage Latinx:</td></tr><tr><td valign="top"> Math</td><td valign="bottom">FFT</td><td valign="bottom" char=".">−.111</td><td valign="bottom" char=".">−.059</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">−.064</td><td valign="bottom" char=".">−.030</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">−.057</td><td valign="bottom" char=".">−.028</td></tr><tr><td valign="top"> English</td><td valign="bottom">FFT</td><td valign="bottom" char=".">−.047</td><td valign="bottom" char=".">−.025</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">.007</td><td valign="bottom" char=".">.004</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">−.016</td><td valign="bottom" char=".">−.010</td></tr><tr><td valign="bottom">Reference: math/ELA:</td></tr><tr><td valign="top"> Math</td><td valign="bottom">FFT</td><td valign="bottom" char=".">.099</td><td valign="bottom" char=".">.052</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">.055</td><td valign="bottom" char=".">.022</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">.018</td><td valign="bottom" char=".">.009</td></tr><tr><td valign="top"> English</td><td valign="bottom">FFT</td><td valign="bottom" char=".">.165</td><td valign="bottom" char=".">.088</td></tr><tr><td valign="bottom">CLASS</td><td valign="bottom" char=".">.044</td><td valign="bottom" char=".">.022</td></tr><tr><td valign="bottom">Value added</td><td valign="bottom" char=".">.147</td><td valign="bottom" char=".">.090</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>15 Note. CLASS = Classroom Assessment Scoring System; ELA = English language arts; FFT = Framework for Teaching.</p> <hd id="AN0155256492-23">Discussion</hd> <p>Prior research on between-teacher variations has revealed a stark racial/ethnic inequality in teacher human resource distribution: youth of color are taught by less experienced and credentialed teachers than their White counterparts. Much of this work has informed current educational policies, and one recommendation that has emerged is assigning high-quality teachers to classrooms with students most in need of "good" teachers. These efforts seek to address, in part, the structural inequalities of teacher sorting. To complement scholarship and policies that seek to address between-teacher variations on certain quality measures, our findings revealed that within-teacher variation in how a teacher instructs their classrooms is a site for more attention and intervention.</p> <p>We posed two overarching questions. We first asked what proportion of the overall classroom-to-classroom variation in teaching effectiveness was due to differences among teachers, as opposed to differences among classrooms taught by the same teacher. Our findings reveal that slightly less than half of the variability in teacher VAM is within a teacher rather than between teachers. Stated another way: the variation in a teacher's VAM between different teachers is roughly the same as the variation among the classes that a single teacher instructs.</p> <p>Our second research question asked whether this classroom-to-classroom and within-teacher variations were associated with classroom racial/ethnic composition, and whether this association was linked with prior academic performance. We found that this variation between classrooms is linked with the racial composition in such a way that perpetuates racial inequalities. Teachers had lower teaching effectiveness scores when they instructed classrooms with more Black, and to a lesser extent Latinx, students. The within-teacher standardized coefficients of percentage Black were in some cases larger than the same coefficients computed on student's incoming test scores. Given that more than a century of scholarly work has described the deep-rooted racial stereotypes that are pervasive in US society (Du Bois [<reflink idref="bib24" id="ref109">24</reflink>]; Ladson-Billings [<reflink idref="bib67" id="ref110">67</reflink>]), it is perhaps unsurprising that teaching efficacy would also work in ways that would disadvantage youth of color, and in particular, Black youth.</p> <p>Adding nuance to the finding that patterns of teaching effectiveness disadvantage youth of color, we asked whether classroom-to-classroom and within-teacher variations associated with classroom racial/ethnic composition were linked with teacher-student race matches. We also found that teachers across racial/ethnic groups show the same patterns in teaching that disadvantage Black youth, which suggests that all teachers, not just White teachers, can benefit from better training and development. But it is worthwhile to ask: Why would White, Latinx, and Black teachers, on average, teach in ways that disadvantage their Black students? Recent empirical work has found that teacher biases often do not vary by teacher race (Cherng and Halpin [<reflink idref="bib15" id="ref111">15</reflink>]; Hibel et al. [<reflink idref="bib50" id="ref112">50</reflink>]). Scholars have long argued that teachers who occupy "different existential worlds" than their students, such as White and non-Black Latinx teachers, can struggle in forging strong relationships (Goodwin [<reflink idref="bib44" id="ref113">44</reflink>]; Villegas and Lucas [<reflink idref="bib98" id="ref114">98</reflink>]). It is logical then that these teachers, who often lack adequate multicultural training, have few tools to understand difference without relying on pervasive racial stereotypes (Dumas [<reflink idref="bib25" id="ref115">25</reflink>], [<reflink idref="bib26" id="ref116">26</reflink>]). In turn, this process perpetuates rather than challenges historical racial inequalities that disadvantage Black youth. For Black teachers, one explanation of why they may also teach in ways that disadvantage Black youth is that they, like teachers belonging to other racial/ethnic groups, are inadequately trained to work with students of color. Although Black teachers share meaningful experiences with their Black youth, particularly around notions of racial identity, they may also subscribe to anti-Black biases, with which they are intimately familiar (Ladson-Billings [<reflink idref="bib68" id="ref117">68</reflink>]). Research also finds that Black teachers and other teachers of color struggle to navigate the White cultural norms embedded in the system of US education (Banks [<reflink idref="bib4" id="ref118">4</reflink>]; Gay [<reflink idref="bib38" id="ref119">38</reflink>]), and that this marginalization may divert energy spent on classroom practices. Therefore, Black teachers may not be adequately trained to use their rich experiences to help Black youth navigate their schooling systems. Moreover, common measures of teacher quality may not capture practices such as culturally responsive teaching and antiracism, which leave important practices that are often shared between Black teachers and their students unrecognized and unvalued. Future research should examine these dynamics to better support teachers in their efforts with an increasingly diverse student population.</p> <p>Finally, we asked whether the association with classroom racial/ethnic composition and classroom-to-classroom and within-teacher variations in teaching efficacy was similar to VAM, which is less subject to potential rater biases. We found that patterns between FFT and CLASS scores and classroom composition were similar to links between VAM and classroom demographics, which suggests that biases lie within teachers and not the raters of teaching videos. The nature of these biases may reflect multiple components. Common measures of teaching efficacy, including FFT and CLASS as well as teacher VAMs, often do not capture vital practices such as culturally responsive and antiracism teaching and pedagogy. It is important that future research continues to examine what factors shape teachers' use of these practices, as well as their links to both student academic and socioemotional outcomes. And although FFT and CLASS do not measure culturally responsive and antiracism teaching, they do measure beneficial practices for all students, such as the degree to which the teacher fosters safe environments or uses open-ended questions. And our findings reveal that teachers use these practices in ways that are consistent with broader racial stereotypes that exist in US society.</p> <p>Moreover, we found evidence that the subject matter of the course matters: teaching effectiveness scores in math classes were more tightly linked to classroom demographic composition than in ELA classes. Why might this be the case? Researchers have argued that math ability, more so than other subjects, may be perceived as an in-born talent, which may in turn shape how teachers interact with students. Ladson-Billings states in regard to math achievement that "a notion prevails in American culture that academic excellence is a result of genetic good fortune. This concept that some students 'have it' whereas others do not is particularly pernicious when directed toward African American students" ([<reflink idref="bib66" id="ref120">66</reflink>], 702). Work examining teacher perceptions of student academic ability has also found that math teachers manifest more racial biases than their English-teaching counterparts (Cherng [<reflink idref="bib14" id="ref121">14</reflink>]).</p> <p>As key actors in shaping long-standing racial achievement gaps in education, teachers have long been identified as key players in rectifying these inequalities. Prior work has established that the sorting of teachers across school districts, schools, and classrooms is inequitable, which reflects forms of structural racism that disadvantage youth of color, and in particular Black youth. Focusing on the dynamic nature of teaching, we find inequalities that suggest that racial bias, on the interactional level, is a major mechanism underlying large and persistent racial inequalities on academic outcomes.</p> <hd id="AN0155256492-24">Study Limitations</hd> <p>Our study has a number of limitations that should be noted. First, our measures of race/ethnicity were monolithic and do not represent diversity within each category. This is particularly an issue for Latinx students, as Latinidad reflects racial and geographic diversity, among other factors. The meaning of this category may also be linked with our findings that Latinx teachers were particularly effective at teaching classrooms with more Asian American students, whereas the opposite was true of classrooms with more Latinx students.</p> <p>Second, our data source, the MET study (see Data and Methods), did not allow us to identify a causal effect of classroom racial composition on within-teacher differences in teaching effectiveness. The MET study used block random assignment of teachers to classrooms in its second year, which served to identify a causal effect defined between teachers and within randomization blocks (Kane et al. [<reflink idref="bib60" id="ref122">60</reflink>]). Our interest was variation within teachers, and thus the MET randomization does not directly play into our analytic strategy. An appropriate design for identifying the effect of classroom racial composition on a teacher's effectiveness would be to assign the same teacher to multiple classrooms that differed systematically only in their racial composition. We are not aware of any study that has implemented such a design.</p> <p>Third, we did not attempt to disentangle how cross-sectional and longitudinal sources of within-teacher variation were associated with classroom racial composition. Cross-sectional variation is confounded by tracking and other sources of nonrandom assignment of students to classrooms. Our analytic strategy statistically controls for tracking based on observed characteristics of students using regression methodology, and we further control for nonrandom assignment of teachers to classrooms based on persistent characteristics of teachers using teacher fixed effects. This is not sufficient to eliminate all potential cross-sectional sources of confounding, but it does provide stronger evidence than zero-order correlations. Longitudinal variation is also confounded by time-varying but systematic changes in teachers' practices (e.g., experience) as well as nonrandom assignment of classrooms to teachers based on nonpersistent teacher characteristics (e.g., previous year's performance).</p> <p>Fourth, a notable limitation of value-added estimates of teaching effectiveness hinges on their nature as a measure derived from student performance on standardized tests. Standardized test scores are relatively narrow measures of student learning and necessarily do not capture classroom interaction quality or student growth in other areas. In fact, recent research affirms the multidimensionality of teaching effectiveness, as performance measures based on test-based VAMs are weakly related to effects on student growth mindset, grit, and effort in class (Kraft [<reflink idref="bib64" id="ref123">64</reflink>]). Despite these limitations, the predictive evidence on value-added estimates of teaching effectiveness is clear in that value-added scores are strong predictors of student long-term outcomes, including college enrollment and job earnings (Chetty et al. [<reflink idref="bib16" id="ref124">16</reflink>]). And finally, the present study has limited generalizability outside the US context, especially considering idiosyncratic concerns regarding sorting of highly qualified and effective teachers occurring between districts that is, in part, due to differences in funding mechanisms and salary levels (Adamson and Darling-Hammond [<reflink idref="bib2" id="ref125">2</reflink>]). Scholars have posited that such concerns surrounding teacher sorting are less salient in other national settings with more rigorous and uniform entry and credential standards for teachers, such as Finland (Stewart [<reflink idref="bib94" id="ref126">94</reflink>]).</p> <hd id="AN0155256492-25">Implications for Education Policy and Practice</hd> <p>Despite these limitations, we have shown that the quality of teachers' in-classroom practices tends to vary across classrooms in a way that systematically disadvantages students of color, and in particular, Black youth. These findings bring to light the importance of teachers' in-classroom practices, not simply their credentials, in educational research and policy targeted at reducing racial inequality in US public schools.</p> <p>As current educational policies continue to focus on teacher quality and racial inequalities, one recommendation that has emerged is assigning high-quality teachers to classrooms with students most in need of "good" teachers. These efforts seek to address, in part, the structural inequalities of teacher sorting. Our finding that teachers instruct quite differently across their classrooms suggests that assigning "good" teachers to classrooms is unlikely to have a consistent impact across the many classrooms with which a teacher works.</p> <p>A more fruitful course of action would be to focus on how to train teachers to work better with different groups of students. This approach also most closely aligns with the main findings of this article: on average, teachers instruct in ways that disadvantage Black youth. One branch of research and reform focuses on the ability of teachers to work with a diverse student population. For more than half a century, efforts have been made to help teachers work with an increasingly non-White student population (Goodwin [<reflink idref="bib45" id="ref127">45</reflink>], [<reflink idref="bib44" id="ref128">44</reflink>]). As our article reveals that teachers enact variable teaching across their own classrooms in ways that disadvantage Black and Latinx youth, it is important to ask how this variation speaks to differentiation: the notion that teachers should vary their teaching to meet the different needs of their students. Although differentiation emphasizes the importance of culturally responsive pedagogical practice, it may be the case that teachers are attempting to adapt their practices to better support their students but are doing so in ways that hinder classroom learning. A body of work that emphasizes the importance of pedagogy that meets the needs of diverse classrooms is culturally responsive and multicultural pedagogy (Gay [<reflink idref="bib37" id="ref129">37</reflink>], [<reflink idref="bib38" id="ref130">38</reflink>]; Gay and Howard [<reflink idref="bib39" id="ref131">39</reflink>]; Irvine [<reflink idref="bib53" id="ref132">53</reflink>]; Irvine and Armento [<reflink idref="bib54" id="ref133">54</reflink>]). This scholarship gives rise to studies on dismantling anti-Black racism within the classroom (Baker-Bell [<reflink idref="bib3" id="ref134">3</reflink>]; McKinney de Royston [<reflink idref="bib75" id="ref135">75</reflink>]) and research that uplifts teachers who "see education as a vehicle for racial liberation and justice and their role as educators as inherently political because of the racialized discourses and practices that stretch into and beyond schools" (McKinney de Royston et al. [<reflink idref="bib76" id="ref136">76</reflink>], 97). Our results, therefore, suggest that there need to be greater pushes to train teachers to be more multiculturally literate, particularly in contexts like those of the MET study: large urban school districts that serve large populations of youth of color.</p> <ref id="AN0155256492-26"> <title> Notes </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> Hua-Yu Sebastian Cherng is an associate professor of international education at New York University. Peter F. Halpin is an associate professor of quantitative methods at the University of North Carolina–Chapel Hill School of Education. Luis A. Rodriguez is an assistant professor of education leadership and policy studies at New York University.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref2" type="bt">2</bibl> <bibtext> The video-recording procedure is described in detail by Whitehurst and colleagues ([101]), which we briefly summarize here. Subject matter generalist teachers (mostly grades 4–5) were recorded on four separate days in each year that they participated in the study. Each day, teachers were recorded in both math and ELA, producing a total of eight videos per teacher per year. Subject matter specialist teachers (mostly grades 6–8) were recorded on two separate days in the first year of the study. On each day, two different class sections were recorded, producing a total of four videos per teacher in year one. In year two, subject matter specialists were observed on four different days but only in the one class section that was randomized as part of the MET study. This yielded another four videos per teacher in year two. Each video recording was approximately 30–35 minutes long, but the different observation protocols were scored using different time intervals. For the CLASS, scoring was done in 15-minute increments. Thus, each video was divided into 15-minute segments, and the segments were scored separately. For CLASS, each rater scored only one segment per video. For FFT, each video recording was scored only once, with the first 15 minutes and last 10 minutes used for observations. In the MET study, the term "video segment" denotes the unit of observation for each protocol, and the number of segments per video is specific to a given observational protocol. For both protocols, 5% of the video segments were double-coded to evaluate interrater reliability. Generalizability studies reported by the MET researchers (Kane and Staiger [62]) showed that less than 10% of variation in ratings was due to rater main effects (this analysis did not address rater-by-teacher interactions).</bibtext> </blist> <blist> <bibl id="bib3" idref="ref3" type="bt">3</bibl> <bibtext> Analyses using the larger, unrestricted sample mirror in whole the findings of this study.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref106" type="bt">4</bibl> <bibtext> Across all of the FFT subdomains, the coefficients for percentage Black are negative and statistically significant. The coefficients for percentage Latinx are negative and statistically significant for all of the subdomains except for "Respect" and "Management." There is more heterogeneity in terms of CLASS subdomains and classroom demographics. The coefficients for percentage Black are statistically significant for 9 of the 15 subdomains. No clear pattern emerges between what subdomains are (and are not) linked.</bibtext> </blist> </ref> <ref id="AN0155256492-27"> <title> References </title> <blist> <bibtext> Aaronson, Daniel, Lisa Barrow, and William Sander. 2007. "Teachers and Student Achievement in the Chicago Public High Schools." Journal of Labor Economics 25 (1): 95–135.</bibtext> </blist> <blist> <bibtext> Adamson, Frank, and Linda Darling-Hammond. 2012. "Funding Disparities and the Inequitable Distribution of Teachers: Evaluating Sources and Solutions." Education Policy Analysis Archives 20:37.</bibtext> </blist> <blist> <bibtext> Baker-Bell, April. 2020. "Dismantling Anti-Black Linguistic Racism in English Language Arts Classrooms: Toward an Anti-Racist Black Language Pedagogy." Theory Into Practice 59 (1): 8–21.</bibtext> </blist> <blist> <bibtext> Banks, James A. 1993. "Multicultural Education: Historical Development, Dimensions, and Practice." Review of Research in Education 19:3–49.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref16" type="bt">5</bibl> <bibtext> Barnett, W. Steven. 2003. "Better Teachers, Better Preschools: Student Achievement Linked to Teacher Qualifications." NIEER Preschool Policy Matters 2, <ulink href="http://eric.ed.gov/?id=ed480818">http://eric.ed.gov/?id=ed480818</ulink>.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref70" type="bt">6</bibl> <bibtext> Borman, Geoffrey, and Maritza Dowling. 2010. "Schools and Inequality: A Multilevel Analysis of Coleman's Equality of Educational Opportunity Data." Teachers College Record 112 (5): 1201–46.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref4" type="bt">7</bibl> <bibtext> Borman, Geoffrey D., and Steven M. Kimball. 2005. "Teacher Quality and Educational Equality: Do Teachers with Higher Standards-Based Evaluation Ratings Close Student Achievement Gaps?" Elementary School Journal 106 (1): 3–20.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref47" type="bt">8</bibl> <bibtext> Brimijoin, Kay. 2005. "Differentiation and High-Stakes Testing: An Oxymoron?" Theory Into Practice 44 (3): 254–61.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref52" type="bt">9</bibl> <bibtext> Brühwiler, Christian, and Peter Blatchford. 2011. "Effects of Class Size and Adaptive Teaching Competency on Classroom Processes and Academic Outcome." Learning and Instruction 21 (1): 95–108.</bibtext> </blist> <blist> <bibtext> Buddin, Richard, and Gema Zamarro. 2009. "Teacher Qualifications and Student Achievement in Urban Elementary Schools." Journal of Urban Economics 66 (2): 103–15.</bibtext> </blist> <blist> <bibtext> Campbell, Shanyce L., and Matthew Ronfeldt. 2018. "Observational Evaluation of Teachers: Measuring More Than We Bargained For?" American Educational Research Journal 55 (6): 1233–67.</bibtext> </blist> <blist> <bibtext> Casabianca, Jodi M., J. R. Lockwood, and Daniel F. McCaffrey. 2015. "Trends in Classroom Observation Scores." Educational and Psychological Measurement 75 (2): 311–37.</bibtext> </blist> <blist> <bibtext> Cherng, Hua-Yu Sebastian. 2017a. "If They Think I Can: Teacher Bias and Youth of Color Expectations and Achievement." Social Science Research 66:170–86.</bibtext> </blist> <blist> <bibtext> Cherng, Hua-Yu Sebastian. 2017b. "The Ties That Bind: Teacher Relationships, Academic Expectations, and Racial/Ethnic and Generational Inequality." American Journal of Education 124 (1): 67–100.</bibtext> </blist> <blist> <bibtext> Cherng, Hua-Yu Sebastian, and Peter F. Halpin. 2016. "The Importance of Minority Teachers: Student Perceptions of Minority versus White Teachers." Educational Researcher 45 (7): 407–20.</bibtext> </blist> <blist> <bibtext> Chetty, Raj, John N. Friedman, and Jonah E. Rockoff. 2014. "Measuring the Impacts of Teachers I: Evaluating Bias in Teacher Value-Added Estimates." American Economic Review 104 (9): 2593–632.</bibtext> </blist> <blist> <bibtext> Clotfelter, Charles T., Helen F. Ladd, and Jacob Vigdor. 2005. "Who Teaches Whom? Race and the Distribution of Novice Teachers." Economics of Education Review 24 (4): 377–92.</bibtext> </blist> <blist> <bibtext> Cohen, David K., and Heather C. Hill. 1998. "Instructional Policy and Classroom Performance: The Mathematics Reform in California." <ulink href="http://eric.ed.gov/?id=ED417942">http://eric.ed.gov/?id=ED417942</ulink>.</bibtext> </blist> <blist> <bibtext> Congdon, Peter J., and Joy McQueen. 2000. "Unmodeled Rater Discrimination Error." Objective Measurement: Theory Into Practice 5:165–180.</bibtext> </blist> <blist> <bibtext> Danielson, Charlotte. 2013. The Framework for Teaching Evaluation Instrument. 2013 Instructionally Focused Edition. Princeton, NJ: Danielson Group, <ulink href="http://www.danielsongroup.org/framework">http://www.danielsongroup.org/framework</ulink>.</bibtext> </blist> <blist> <bibtext> Darling-Hammond, Linda. 2009. "Recognizing and Enhancing Teaching Effectiveness." International Journal of Educational and Psychological Assessment, <ulink href="http://psycnet.apa.org/psycinfo/2010-11469-001">http://psycnet.apa.org/psycinfo/2010-11469-001</ulink>.</bibtext> </blist> <blist> <bibtext> Dee, Thomas S. 2004. "Teachers, Race, and Student Achievement in a Randomized Experiment." Review of Economics and Statistics 86 (1): 195–210.</bibtext> </blist> <blist> <bibtext> Dee, Thomas S. 2005. "A Teacher Like Me: Does Race, Ethnicity, or Gender Matter?" American Economic Review 95 (2): 158–65.</bibtext> </blist> <blist> <bibtext> Du Bois, William Edward Burghardt. (1903) 2008. The Souls of Black Folk. Oxford: Oxford University Press.</bibtext> </blist> <blist> <bibtext> Dumas, Michael J. 2014. "'Losing an Arm': Schooling as a Site of Black Suffering." Race, Ethnicity and Education 17 (1): 1–29.</bibtext> </blist> <blist> <bibtext> Dumas, Michael J. 2016. "Against the Dark: Antiblackness in Education Policy and Discourse." Theory Into Practice 55 (1): 11–19.</bibtext> </blist> <blist> <bibtext> Early, Diane M., Donna M. Bryant, Robert C. Pianta, Richard M. Clifford, Margaret R. Burchinal, Sharon Ritchie, and Oscar Barbarin. 2006. "Are Teachers' Education, Major, and Credentials Related to Classroom Quality and Children's Academic Gains in Pre-Kindergarten?" Early Childhood Research Quarterly 21 (2): 174–95.</bibtext> </blist> <blist> <bibtext> Early, Diane M., Kelly L. Maxwell, Margaret Burchinal, Soumya Alva, Randall H. Bender, Donna Bryant, Karen Cai, et al. 2007. "Teachers' Education, Classroom Quality, and Young Children's Academic Skills: Results from Seven Studies of Preschool Programs." Child Development 78 (2): 558–80.</bibtext> </blist> <blist> <bibtext> Egalite, Anna J., Brian Kisida, and Marcus A. Winters. 2015. "Representation in the Classroom: The Effect of Own-Race Teachers on Student Achievement." Economics of Education Review 45:44–52.</bibtext> </blist> <blist> <bibtext> Ehrenberg, Ronald G., Daniel D. Goldhaber, and Dominic J. Brewer. 1995. "Do Teachers' Race, Gender, and Ethnicity Matter? Evidence from NELS88." Working Paper no. 4669, National Bureau of Economic Research, Cambridge, MA.</bibtext> </blist> <blist> <bibtext> Ferguson, Ronald F. 2003. "Teachers' Perceptions and Expectations and the Black-White Test Score Gap." Urban Education 38 (4): 460–507.</bibtext> </blist> <blist> <bibtext> Fox, Lindsay. 2016a. "Playing to Teachers' Strengths: Using Multiple Measures of Teaching Effectiveness to Improve Teacher Assignments." Education Finance and Policy 11 (1): 70–96.</bibtext> </blist> <blist> <bibtext> Fox, Lindsay. 2016b. "Seeing Potential: The Effects of Student-Teacher Demographic Congruence on Teacher Expectations and Recommendations." AERA Open 2 (1): 1–17.</bibtext> </blist> <blist> <bibtext> Fraser, Barry J., Herbert J. Walberg, Wayne W. Welch, and John A. Hattie. 1987. "Syntheses of Educational Productivity Research." International Journal of Educational Research 11 (2): 147–252.</bibtext> </blist> <blist> <bibtext> Freeman, Catherine E., Benjamin Scafidi, and David L. Sjoquist. 2005. "Racial Segregation in Georgia Public Schools, 1994–2001." In School Resegregation: Must the South Turn Back?, ed. John Charles Boger and Gary Orfield. Chapel Hill: University of North Carolina Press.</bibtext> </blist> <blist> <bibtext> Garrett, Rachel, and Matthew P. Steinberg. 2015. "Examining Teaching Effectiveness Using Classroom Observation Scores: Evidence from the Randomization of Teachers to Students." Educational Evaluation and Policy Analysis 37 (2): 224–42.</bibtext> </blist> <blist> <bibtext> Gay, Geneva. 1997. "Multicultural Infusion in Teacher Education: Foundations and Applications." Peabody Journal of Education 72 (1): 150–77.</bibtext> </blist> <blist> <bibtext> Gay, Geneva. 2010. "Acting on Beliefs in Teacher Education for Cultural Diversity." Journal of Teacher Education 61 (1–2): 143–52.</bibtext> </blist> <blist> <bibtext> Gay, Geneva, and Tyrone C. Howard. 2000. "Multicultural Teacher Education for the 21st Century." Teacher Educator 36 (1): 1–16.</bibtext> </blist> <blist> <bibtext> Gershenson, Seth, Cassandra Hart, Joshua Hyman, Constance Lindsay, and Nicholas W. Papageorge. 2018. "The Long-Run Impacts of Same-Race Teachers." Working Paper no. 25254, National Bureau of Economic Research, Cambridge, MA.</bibtext> </blist> <blist> <bibtext> Gershenson, Seth, Stephen B. Holt, and Nicholas W. Papageorge. 2016. "Who Believes in Me? The Effect of Student-Teacher Demographic Match on Teacher Expectations." Economics of Education Review 52:209–24.</bibtext> </blist> <blist> <bibtext> Goldhaber, Dan, Lesley Lavery, and Roddy Theobald. 2015. "Uneven Playing Field? Assessing the Teacher Quality Gap between Advantaged and Disadvantaged Students." Educational Researcher 44 (5): 293–307.</bibtext> </blist> <blist> <bibtext> Goldhaber, Dan D., and Dominic J. Brewer. 2000. "Does Teacher Certification Matter? High School Teacher Certification Status and Student Achievement." Educational Evaluation and Policy Analysis 22 (2): 129–45.</bibtext> </blist> <blist> <bibtext> Goodwin, A. 2000. "Teachers as (Multi) Cultural Agents in Schools." In Addressing Cultural Issues in Organizations: Beyond the Corporate Context, ed. Robert T. Carter. Thousand Oaks, CA: Sage.</bibtext> </blist> <blist> <bibtext> Goodwin, A. Lin. 1997. "Historical and Contemporary Perspectives on Multicultural Teacher Education: Past Lessons, New Directions." In Preparing Teachers for Cultural Diversity, ed. J. E. King, E. R. Hollins, and W. C. Hayman. New York: Teachers College Press.</bibtext> </blist> <blist> <bibtext> Gregory, Anne, Russell J. Skiba, and Pedro A. Noguera. 2010. "The Achievement Gap and the Discipline Gap Two Sides of the Same Coin?" Educational Researcher 39 (1): 59–68.</bibtext> </blist> <blist> <bibtext> Grissom, Jason A., Emily C. Kern, and Luis A. Rodriguez. 2015. "The 'Representative Bureaucracy' in Education: Educator Workforce Diversity, Policy Outputs, and Outcomes for Disadvantaged Students." Educational Researcher 44 (3): 185–92.</bibtext> </blist> <blist> <bibtext> Haller, Emil J. 1985. "Pupil Race and Elementary School Ability Grouping: Are Teachers Biased against Black Children?" American Educational Research Journal 22 (4): 465–83.</bibtext> </blist> <blist> <bibtext> Hedges, Larry V., and Eric C. Hedberg. 2007. "Intraclass Correlation Values for Planning Group Randomized Trials in Education." Educational Evaluation and Policy Analysis 29:60–87.</bibtext> </blist> <blist> <bibtext> Hibel, Jacob, George Farkas, and Paul L. Morgan. 2010. "Who Is Placed into Special Education?" Sociology of Education 83 (4): 312–32.</bibtext> </blist> <blist> <bibtext> Hill, Heather C., Brian Rowan, and Deborah L. Ball. 2005. "Effects of Teachers' Mathematical Knowledge for Teaching on Student Achievement." American Educational Research Journal 42 (2): 371–406.</bibtext> </blist> <blist> <bibtext> Ho, Andrew D., and Thomas J. Kane. 2013. "The Reliability of Classroom Observations by School Personnel." MET Project Research Paper, Bill and Melinda Gates Foundation, Seattle.</bibtext> </blist> <blist> <bibtext> Irvine, Jacqueline J. 2003. Educating Teachers for Diversity: Seeing with a Cultural Eye. New York: Teachers College Press.</bibtext> </blist> <blist> <bibtext> Irvine, Jacqueline J., and Beverly J. Armento. 2001. Culturally Responsive Teaching. New York: McGraw-Hill.</bibtext> </blist> <blist> <bibtext> Isenberg, Eric, Jeffrey Max, Philip Gleason, Liz Potamites, Robert Santillano, Heinrich Hock, and Michael Hansen. 2013. "Access to Effective Teaching for Disadvantaged Students." NCEE 2014–4001. Institute of Education Sciences, <ulink href="http://eric.ed.gov/?id=ED544345">http://eric.ed.gov/?id=ED544345</ulink>.</bibtext> </blist> <blist> <bibtext> Jackson, C. Kirabo. 2009. "Student Demographics, Teacher Sorting, and Teacher Quality: Evidence from the End of School Desegregation." Journal of Labor Economics 27 (2): 213–56.</bibtext> </blist> <blist> <bibtext> Jerald, Craig D., and Richard Ingersoll. 2002. "All Talk No Action: Putting an End to Out-of-Field Teaching." <ulink href="http://repository.upenn.edu/gse%5fpubs/142/">http://repository.upenn.edu/gse%5fpubs/142/</ulink>.</bibtext> </blist> <blist> <bibtext> Kalogrides, Demetra, Susanna Loeb, and Tara Béteille. 2013. "Systematic Sorting: Teacher Characteristics and Class Assignments." Sociology of Education 86 (2): 103–23.</bibtext> </blist> <blist> <bibtext> Kane, Thomas J., and Steven Cantrell. 2010. "Learning about Teaching: Initial Findings from the Measures of Effective Teaching Project." MET Project Research Paper, Bill and Melinda Gates Foundation, Seattle.</bibtext> </blist> <blist> <bibtext> Kane, Thomas J., Daniel F. McCaffrey, Trey Miller, and Douglas O. Staiger. 2013. "Have We Identified Effective Teachers? Validating Measures of Effective Teaching Using Random Assignment." MET Project Research Paper, Bill and Melinda Gates Foundation, Seattle.</bibtext> </blist> <blist> <bibtext> Kane, Thomas J., Jonah E. Rockoff, and Douglas O. Staiger. 2008. "What Does Certification Tell Us about Teaching Effectiveness? Evidence from New York City." Economics of Education Review 27 (6): 615–31.</bibtext> </blist> <blist> <bibtext> Kane, Thomas J., and Douglas O. Staiger. 2012. "Gathering Feedback for Teaching: Combining High-Quality Observations with Student Surveys and Achievement Gains." MET Project Research Paper, Bill and Melinda Gates Foundation, Seattle.</bibtext> </blist> <blist> <bibtext> Koedel, Cory, and Julian R. Betts. 2007. "Re-Examining the Role of Teacher Quality in the Educational Production Function." Working Paper 2007-03, National Center on Performance Incentives, Vanderbilt University's Peabody College, Nashville, TN.</bibtext> </blist> <blist> <bibtext> Kraft, Matthew A. 2019. "Teacher Effects on Complex Cognitive Skills and Social-Emotional Competencies." Journal of Human Resources 54 (1): 1–36.</bibtext> </blist> <blist> <bibtext> Ladd, Helen F., and Lucy C. Sorensen. 2017. "Returns to Teacher Experience: Student Achievement and Motivation in Middle School." Education Finance and Policy 12 (2): 241–79.</bibtext> </blist> <blist> <bibtext> Ladson-Billings, Gloria. 1997. "It Doesn't Add Up: African American Students' Mathematics Achievement." Journal for Research in Mathematics education 28 (6): 697–708.</bibtext> </blist> <blist> <bibtext> Ladson-Billings, Gloria. 2006. "From the Achievement Gap to the Education Debt: Understanding Achievement in US Schools." Educational Researcher 35 (7): 3–12.</bibtext> </blist> <blist> <bibtext> Ladson-Billings, Gloria. 2009. The Dreamkeepers: Successful Teachers of African American Children. San Francisco: Jossey-Bass.</bibtext> </blist> <blist> <bibtext> Lankford, Hamilton, Susanna Loeb, and James Wyckoff. 2002. "Teacher Sorting and the Plight of Urban Schools: A Descriptive Analysis." Educational Evaluation and Policy Analysis 24 (1): 37–62.</bibtext> </blist> <blist> <bibtext> Lockwood, J. R., and Daniel F. McCaffrey. 2013. "Correcting for Test Score Measurement Error in ANCOVA Models for Estimating Treatment Effects." Journal of Educational and Behavioral Statistics 39 (1): 22–52.</bibtext> </blist> <blist> <bibtext> Loeb, Susanna, James Soland, and Lindsay Fox. 2014. "Is a Good Teacher a Good Teacher for All? Comparing Value-Added of Teachers with Their English Learners and Non-English Learners." Educational Evaluation and Policy Analysis 36 (4): 457–75.</bibtext> </blist> <blist> <bibtext> Mayer, Anysia, Kimberly LeChasseur, and Morgaen Donaldson. 2018. "The Structure of Tracking: Instructional Practices of Teachers Leading Low- and High-Track Classes." American Journal of Education 124 (4): 445–77.</bibtext> </blist> <blist> <bibtext> McCaffrey, Daniel F., J. R. Lockwood, Daniel M. Koretz, and Laura S. Hamilton. 2003. Evaluating Models for Teacher Accountability. Santa Monica, CA: RAND Corporation.</bibtext> </blist> <blist> <bibtext> McCaffrey, Daniel F., Tim R. Sass, J. R. Lockwood, and Kata Mihaly. 2009. "The Intertemporal Variability of Teacher Effect Estimates." Education Finance and Policy 4 (4): 572–606.</bibtext> </blist> <blist> <bibtext> McKinney de Royston, Maxine. 2020. "Black Womanist Teachers' Political Clarity in Theory and Practice." Theory Into Practice 59 (4): 379–88.</bibtext> </blist> <blist> <bibtext> McKinney de Royston, Maxine, Tia C. Madkins, Jarvis R. Givens, and Na'ilah Suad Nasir. 2020. "'I'm a Teacher, I'm Gonna Always Protect You': Understanding Black Educators' Protection of Black Children." American Educational Research Journal 58 (1): 68–106.</bibtext> </blist> <blist> <bibtext> Morganstein, David, and Ron Wasserstein. 2014. "ASA Statement on Value-Added Models." Statistics and Public Policy 1 (1): 108–110.</bibtext> </blist> <blist> <bibtext> National Council on Teacher Quality. 2015. "State of the States 2015: Evaluating Teaching, Leading, and Learning." https://<ulink href="http://www.nctq.org/dmsView/StateofStates2015">www.nctq.org/dmsView/StateofStates2015</ulink>.</bibtext> </blist> <blist> <bibtext> Osei-Twumasi, Olivia, and Bernardette J. Pinetta. 2019. "Quality of Classroom Interactions and the Demographic Divide: Evidence from the Measures of Effective Teaching Study." Urban Education, https://doi.org/10.1177/0042085919893744.</bibtext> </blist> <blist> <bibtext> Papay, John P., and Matthew A. Kraft. 2015. "Productivity Returns to Experience in the Teacher Labor Market: Methodological Challenges and New Evidence on Long-Term Career Improvement." Journal of Public Economics 130:105–19.</bibtext> </blist> <blist> <bibtext> Pianta, Robert C., and Bridget K. Hamre. 2009. "Conceptualization, Measurement, and Improvement of Classroom Processes: Standardized Observation Can Leverage Capacity." Educational Researcher 38 (2): 109–19.</bibtext> </blist> <blist> <bibtext> Preston, Courtney. 2017. "University-Based Teacher Preparation and Middle Grades Teaching Effectiveness." Journal of Teacher Education 68 (1): 102–16.</bibtext> </blist> <blist> <bibtext> Riegle-Crumb, Catherine, and Melissa Humphries. 2012. "Exploring Bias in Math Teachers' Perceptions of Students' Ability by Gender and Race/Ethnicity." Gender and Society 26 (2): 290–322.</bibtext> </blist> <blist> <bibtext> Rivkin, Steven G., Eric A. Hanushek, and John F. Kain. 2005. "Teachers, Schools, and Academic Achievement." Econometrica 73 (2): 417–58.</bibtext> </blist> <blist> <bibtext> Saft, Elizabeth W., and Robert C. Pianta. 2001. "Teachers' Perceptions of Their Relationships with Students: Effects of Child Age, Gender, and Ethnicity of Teachers and Children." School Psychology Quarterly 16 (2): 125–41.</bibtext> </blist> <blist> <bibtext> Scheerens, Jaap, and Roel J. Bosker. 1997. The Foundations of Educational Effectiveness. Oxford: Pergamon.</bibtext> </blist> <blist> <bibtext> Simpson, Adelaide W., and Marilyn T. Erickson. 1983. "Teachers' Verbal and Nonverbal Communication Patterns as a Function of Teacher Race, Student Gender, and Student Race." American Educational Research Journal 20 (2): 183–98.</bibtext> </blist> <blist> <bibtext> Skiba, Russell J., Robert H. Horner, Choong-Geun Chung, M. K. Rausch, Seth L. May, and Tary Tobin. 2011. "Race Is Not Neutral: A National Investigation of African American and Latino Disproportionality in School Discipline." School Psychology Review 40 (1): 85.</bibtext> </blist> <blist> <bibtext> Smith, Tom W. 2005. "Generation Gaps in Attitudes and Values from the 1970s to the 1990s." In On the Frontier of Adulthood: Theory, Research and Public Policy, ed. Richard A. Settersten Jr., Frank F. Furstenberg Jr., and Rubén G. Rumbaut. Chicago: University of Chicago Press.</bibtext> </blist> <blist> <bibtext> Snow, Catherine E., Peg Griffin, and M. Susan Burns. 2005. "A Model of Professional Growth in Reading Education." In Knowledge to Support the Teaching of Reading. San Francisco: Jossey-Bass.</bibtext> </blist> <blist> <bibtext> Spillane, James P. 1999. "External Reform Initiatives and Teachers' Efforts to Reconstruct Their Practice: The Mediating Role of Teachers' Zones of Enactment." Journal of Curriculum Studies 31 (2): 143–75.</bibtext> </blist> <blist> <bibtext> Steele, Jennifer L., Matthew J. Pepper, Matthew G. Springer, and J. R. Lockwood. 2015. "The Distribution and Mobility of Effective Teachers: Evidence from a Large, Urban School District." Economics of Education Review 48:86–101.</bibtext> </blist> <blist> <bibtext> Steinberg, Matthew P., and Rachel Garrett. 2016. "Classroom Composition and Measured Teacher Performance: What Do Teacher Observation Scores Really Measure?" Educational Evaluation and Policy Analysis 38 (2): 293–317.</bibtext> </blist> <blist> <bibtext> Stewart, Vivien. 2011. "Raising Teacher Quality around the World." Educational Leadership 68 (4): 16–20.</bibtext> </blist> <blist> <bibtext> Tenenbaum, Harriet R., and Martin D. Ruck. 2007. "Are Teachers' Expectations Different for Racial Minority Than for European American Students? A Meta-analysis." Journal of Educational Psychology 99 (2): 253–73.</bibtext> </blist> <blist> <bibtext> Tennessee Department of Education. 2007. Tennessee's Most Effective Teachers: Are They Assigned to the Schools That Need Them Most? Nashville: Tennessee Department of Education.</bibtext> </blist> <blist> <bibtext> US Department of Education Office for Civil Rights. 2014. "Civil Rights Data Collection, Data Snapshot: Teacher Equity." Issue Brief No. 4, US Department of Education Office for Civil Rights, Washington, DC.</bibtext> </blist> <blist> <bibtext> Villegas, Ana M., and Tamara Lucas. 2002. Educating Culturally Responsive Teachers: A Coherent Approach. SUNY Series, Teacher Preparation and Development. Albany: State University of New York Press.</bibtext> </blist> <blist> <bibtext> Waitoller, Federico R., and Alfredo J. Artiles. 2013. "A Decade of Professional Development Research for Inclusive Education: A Critical Review and Notes for a Research Program." Review of Educational Research 83 (3): 319–56.</bibtext> </blist> <blist> <bibtext> Wenglinsky, Harold. 2002. "The Link between Teacher Classroom Practices and Student Academic Performance." Education Policy Analysis Archives 10:12.</bibtext> </blist> <blist> <bibtext> Whitehurst, G., Matthew M. Chingos, and Katharine M. Lindquist. 2014. "Evaluating Teachers with Classroom Observations." Brown Center on Education Policy, Brookings Institution, Washington, DC.</bibtext> </blist> <blist> <bibtext> Zeichner, Ken. 2010. "Rethinking the Connections between Campus Courses and Field Experiences in College- and University-Based Teacher Education." Journal of Teacher Education 61 (1–2): 89–99.</bibtext> </blist> </ref> <aug> <p>By Hua-Yu Sebastian Cherng; Peter F. Halpin and Luis A. Rodriguez</p> <p>Reported by Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib17" firstref="ref5"></nolink> <nolink nlid="nl2" bibid="bib35" firstref="ref6"></nolink> <nolink nlid="nl3" bibid="bib42" firstref="ref7"></nolink> <nolink nlid="nl4" bibid="bib56" firstref="ref8"></nolink> <nolink nlid="nl5" bibid="bib57" firstref="ref9"></nolink> <nolink nlid="nl6" bibid="bib58" firstref="ref10"></nolink> <nolink nlid="nl7" bibid="bib69" firstref="ref11"></nolink> <nolink nlid="nl8" bibid="bib96" firstref="ref12"></nolink> <nolink nlid="nl9" bibid="bib97" firstref="ref13"></nolink> <nolink nlid="nl10" bibid="bib10" firstref="ref17"></nolink> <nolink nlid="nl11" bibid="bib27" firstref="ref18"></nolink> <nolink nlid="nl12" bibid="bib28" firstref="ref19"></nolink> <nolink nlid="nl13" bibid="bib43" firstref="ref20"></nolink> <nolink nlid="nl14" bibid="bib51" firstref="ref21"></nolink> <nolink nlid="nl15" bibid="bib61" firstref="ref22"></nolink> <nolink nlid="nl16" bibid="bib82" firstref="ref23"></nolink> <nolink nlid="nl17" bibid="bib84" firstref="ref24"></nolink> <nolink nlid="nl18" bibid="bib89" firstref="ref25"></nolink> <nolink nlid="nl19" bibid="bib102" firstref="ref26"></nolink> <nolink nlid="nl20" bibid="bib52" firstref="ref27"></nolink> <nolink nlid="nl21" bibid="bib62" firstref="ref28"></nolink> <nolink nlid="nl22" bibid="bib92" firstref="ref29"></nolink> <nolink nlid="nl23" bibid="bib55" firstref="ref31"></nolink> <nolink nlid="nl24" bibid="bib78" firstref="ref32"></nolink> <nolink nlid="nl25" bibid="bib13" firstref="ref33"></nolink> <nolink nlid="nl26" bibid="bib68" firstref="ref34"></nolink> <nolink nlid="nl27" bibid="bib83" firstref="ref35"></nolink> <nolink nlid="nl28" bibid="bib65" firstref="ref36"></nolink> <nolink nlid="nl29" bibid="bib80" firstref="ref37"></nolink> <nolink nlid="nl30" bibid="bib63" firstref="ref41"></nolink> <nolink nlid="nl31" bibid="bib36" firstref="ref43"></nolink> <nolink nlid="nl32" bibid="bib100" firstref="ref44"></nolink> <nolink nlid="nl33" bibid="bib20" firstref="ref45"></nolink> <nolink nlid="nl34" bibid="bib81" firstref="ref46"></nolink> <nolink nlid="nl35" bibid="bib18" firstref="ref48"></nolink> <nolink nlid="nl36" bibid="bib21" firstref="ref49"></nolink> <nolink nlid="nl37" bibid="bib90" firstref="ref50"></nolink> <nolink nlid="nl38" bibid="bib91" firstref="ref51"></nolink> <nolink nlid="nl39" bibid="bib34" firstref="ref53"></nolink> <nolink nlid="nl40" bibid="bib86" firstref="ref54"></nolink> <nolink nlid="nl41" bibid="bib59" firstref="ref55"></nolink> <nolink nlid="nl42" bibid="bib74" firstref="ref56"></nolink> <nolink nlid="nl43" bibid="bib101" firstref="ref57"></nolink> <nolink nlid="nl44" bibid="bib79" firstref="ref60"></nolink> <nolink nlid="nl45" bibid="bib32" firstref="ref61"></nolink> <nolink nlid="nl46" bibid="bib71" firstref="ref62"></nolink> <nolink nlid="nl47" bibid="bib72" firstref="ref65"></nolink> <nolink nlid="nl48" bibid="bib99" firstref="ref66"></nolink> <nolink nlid="nl49" bibid="bib38" firstref="ref67"></nolink> <nolink nlid="nl50" bibid="bib31" firstref="ref68"></nolink> <nolink nlid="nl51" bibid="bib48" firstref="ref71"></nolink> <nolink nlid="nl52" bibid="bib95" firstref="ref72"></nolink> <nolink nlid="nl53" bibid="bib46" firstref="ref73"></nolink> <nolink nlid="nl54" bibid="bib88" firstref="ref74"></nolink> <nolink nlid="nl55" bibid="bib47" firstref="ref75"></nolink> <nolink nlid="nl56" bibid="bib87" firstref="ref76"></nolink> <nolink nlid="nl57" bibid="bib33" firstref="ref77"></nolink> <nolink nlid="nl58" bibid="bib41" firstref="ref78"></nolink> <nolink nlid="nl59" bibid="bib40" firstref="ref79"></nolink> <nolink nlid="nl60" bibid="bib85" firstref="ref80"></nolink> <nolink nlid="nl61" bibid="bib22" firstref="ref81"></nolink> <nolink nlid="nl62" bibid="bib23" firstref="ref82"></nolink> <nolink nlid="nl63" bibid="bib29" firstref="ref83"></nolink> <nolink nlid="nl64" bibid="bib30" firstref="ref84"></nolink> <nolink nlid="nl65" bibid="bib12" firstref="ref85"></nolink> <nolink nlid="nl66" bibid="bib19" firstref="ref86"></nolink> <nolink nlid="nl67" bibid="bib11" firstref="ref87"></nolink> <nolink nlid="nl68" bibid="bib93" firstref="ref88"></nolink> <nolink nlid="nl69" bibid="bib70" firstref="ref100"></nolink> <nolink nlid="nl70" bibid="bib73" firstref="ref101"></nolink> <nolink nlid="nl71" bibid="bib77" firstref="ref102"></nolink> <nolink nlid="nl72" bibid="bib49" firstref="ref105"></nolink> <nolink nlid="nl73" bibid="bib24" firstref="ref109"></nolink> <nolink nlid="nl74" bibid="bib67" firstref="ref110"></nolink> <nolink nlid="nl75" bibid="bib15" firstref="ref111"></nolink> <nolink nlid="nl76" bibid="bib50" firstref="ref112"></nolink> <nolink nlid="nl77" bibid="bib44" firstref="ref113"></nolink> <nolink nlid="nl78" bibid="bib98" firstref="ref114"></nolink> <nolink nlid="nl79" bibid="bib25" firstref="ref115"></nolink> <nolink nlid="nl80" bibid="bib26" firstref="ref116"></nolink> <nolink nlid="nl81" bibid="bib66" firstref="ref120"></nolink> <nolink nlid="nl82" bibid="bib14" firstref="ref121"></nolink> <nolink nlid="nl83" bibid="bib60" firstref="ref122"></nolink> <nolink nlid="nl84" bibid="bib64" firstref="ref123"></nolink> <nolink nlid="nl85" bibid="bib16" firstref="ref124"></nolink> <nolink nlid="nl86" bibid="bib94" firstref="ref126"></nolink> <nolink nlid="nl87" bibid="bib45" firstref="ref127"></nolink> <nolink nlid="nl88" bibid="bib37" firstref="ref129"></nolink> <nolink nlid="nl89" bibid="bib39" firstref="ref131"></nolink> <nolink nlid="nl90" bibid="bib53" firstref="ref132"></nolink> <nolink nlid="nl91" bibid="bib54" firstref="ref133"></nolink> <nolink nlid="nl92" bibid="bib75" firstref="ref135"></nolink> <nolink nlid="nl93" bibid="bib76" firstref="ref136"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1325802
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Teaching Bias? Relations between Teaching Quality and Classroom Demographic Composition
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Cherng%2C+Hua-Yu+Sebastian%22">Cherng, Hua-Yu Sebastian</searchLink><br /><searchLink fieldCode="AR" term="%22Halpin%2C+Peter+F%2E%22">Halpin, Peter F.</searchLink><br /><searchLink fieldCode="AR" term="%22Rodriguez%2C+Luis+A%2E%22">Rodriguez, Luis A.</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22American+Journal+of+Education%22"><i>American Journal of Education</i></searchLink>. Feb 2022 128(2):171-201.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: University of Chicago Press. Journals Division, P.O. Box 37005, Chicago, IL 60637. Tel: 877-705-1878; Tel: 773-753-3347; Fax: 877-705-1879; Fax: 773-753-0811; e-mail: subscriptions@press.uchicago.edu; Web site: http://www.journals.uchicago.edu/journals/aje/about
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 31
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2022
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Teacher+Effectiveness%22">Teacher Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Characteristics%22">Teacher Characteristics</searchLink><br /><searchLink fieldCode="DE" term="%22Minority+Group+Students%22">Minority Group Students</searchLink><br /><searchLink fieldCode="DE" term="%22Racial+Composition%22">Racial Composition</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Quality%22">Educational Quality</searchLink><br /><searchLink fieldCode="DE" term="%22Racial+Differences%22">Racial Differences</searchLink><br /><searchLink fieldCode="DE" term="%22Ethnicity%22">Ethnicity</searchLink><br /><searchLink fieldCode="DE" term="%22African+American+Students%22">African American Students</searchLink><br /><searchLink fieldCode="DE" term="%22Hispanic+American+Students%22">Hispanic American Students</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1086/717676
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0195-6744
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Purpose: Prior work has drawn consistent conclusions about systematic racial disparities in the allocation of high-quality teachers in US public schools, such as classrooms with more students of color having teachers with fewer credentials and less experience. However, these fixed characteristics of teachers are only proxies for the quality of teaching, which may vary within teacher by the different classrooms they teach. Research Methods/Approach: Using data from the Measures of Effective Teaching (MET) project, we consider various sources of within-teacher, across-classroom variation of teaching effectiveness using a teacher fixed-effects modeling approach and highlight those that may, on average, disadvantage youth of color. Findings: We find that (1) about half of the variation in classroom teaching efficacy is within teachers, (2) classrooms taught by the same teacher with higher percentages of Black and Latinx students receive lower quality of teaching, and (3) these patterns are consistent across teacher racial/ethnic groups. A number of plausible explanations of this association are considered, including rater biases on observational measures and differences in teaching practices; we also consider how these variations differ by teacher race. Implications: Our findings highlight the importance of teachers' in-classroom practices, not simply their credentials, in educational research, policy, and practice targeted at reducing racial inequality.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2022
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1325802
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1325802
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1086/717676
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 31
        StartPage: 171
    Subjects:
      – SubjectFull: Teacher Effectiveness
        Type: general
      – SubjectFull: Teacher Characteristics
        Type: general
      – SubjectFull: Minority Group Students
        Type: general
      – SubjectFull: Racial Composition
        Type: general
      – SubjectFull: Educational Quality
        Type: general
      – SubjectFull: Racial Differences
        Type: general
      – SubjectFull: Ethnicity
        Type: general
      – SubjectFull: African American Students
        Type: general
      – SubjectFull: Hispanic American Students
        Type: general
    Titles:
      – TitleFull: Teaching Bias? Relations between Teaching Quality and Classroom Demographic Composition
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Cherng, Hua-Yu Sebastian
      – PersonEntity:
          Name:
            NameFull: Halpin, Peter F.
      – PersonEntity:
          Name:
            NameFull: Rodriguez, Luis A.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Type: published
              Y: 2022
          Identifiers:
            – Type: issn-print
              Value: 0195-6744
          Numbering:
            – Type: volume
              Value: 128
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: American Journal of Education
              Type: main
ResultId 1