View in EDS HTML Full Text PDF Full Text

Selective Attrition in Longitudinal Studies: Effective Processes for Facebook Tracing

Saved in:
Bibliographic Details
Title: Selective Attrition in Longitudinal Studies: Effective Processes for Facebook Tracing
Language: English
Authors: Bennetts, Shannon K., Love, Jasmine, Hackworth, Naomi J., Mensah, Fiona K., Westrupp, Elizabeth M., Berthelsen, Donna, Levickis, Penny, Bennett, Clair, Nicholson, Jan M.
Source: International Journal of Social Research Methodology. 2021 24(2):135-147.
Availability: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
Peer Reviewed: Y
Page Count: 13
Publication Date: 2021
Document Type: Journal Articles
Reports - Research
Education Level: Adult Education
Descriptors: Attrition (Research Studies), Longitudinal Studies, Social Media, Family (Sociological Unit), Randomized Controlled Trials, Child Rearing, Parent Education, Foreign Countries
Geographic Terms: Australia
DOI: 10.1080/13645579.2020.1765104
ISSN: 1364-5579
Abstract: Participant attrition is a significant challenge for longitudinal studies, particularly those involving disadvantaged populations. We attempted to re-engage n = 990 families in a five-year follow-up of a randomised controlled trial of an early childhood parenting intervention. Tracing was attempted for n = 90/990 (9.1%) uncontactable participants. Evidence for the identification of a matching Facebook profile was classified as strong, moderate or weak. A private message was sent to those with strong evidence (n = 63/90, 70.0%), of which 32 responded, 22 agreed to participate and 19 provided data. Compared to participants engaged using traditional methods, those re-engaged through Facebook were more likely at initial recruitment to have a younger child (p =0.02), to be a single parent (p =0.04), less educated (p <0.001), lower income (p =0.01) and in an unemployed household (p <0.001). Findings suggest that social media can be an effective engagement tool, helping to minimise overall and selective attrition in longitudinal studies, particularly for hard-to-reach populations.
Abstractor: As Provided
Entry Date: 2021
Accession Number: EJ1289842
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwFGHJilSWHnmPevxQ1nGLl3AAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDMgHpOTQ979-y1__rQIBEICBmwmE08qrcIR0NEKxoSWORUeBfwxvZix4tvojiDLK4x2Gpd1y2NdiE4ZZxf9TnXLxvzeIWBBqeTpdPTMSHh5E2o6r2D_J9AsS6PYnIb62MDN4mVB9LcWG0G6uWxf8_uIM3ot4UyQ0Ju7npMab2saXAi3tOJzOT9Tz5Zbde0twgVtzC1QOtOo94b4KfR2SfoB2Jtofkv5APVPvlFEA
Text:
  Availability: 1
  Value: &lt;anid&gt;AN0149012894;9eb01mar.21;2021Mar04.02:01;v2.2.500&lt;/anid&gt; &lt;title id=&quot;AN0149012894-1&quot;&gt;Selective attrition in longitudinal studies: effective processes for Facebook tracing&#160;&lt;/title&gt; &lt;p&gt;Participant attrition is a significant challenge for longitudinal studies, particularly those involving disadvantaged populations. We attempted to re-engage n = 990 families in a five-year follow-up of a randomised controlled trial of an early childhood parenting intervention. Tracing was attempted for n = 90/990 (9.1%) uncontactable participants. Evidence for the identification of a matching Facebook profile was classified as strong, moderate or weak. A private message was sent to those with strong evidence (n = 63/90, 70.0%), of which 32 responded, 22 agreed to participate and 19 provided data. Compared to participants engaged using traditional methods, those re-engaged through Facebook were more likely at initial recruitment to have a younger child (p =.02), to be a single parent (p =.04), less educated (p &amp;lt;.001), lower income (p =.01) and in an unemployed household (p &amp;lt;.001). Findings suggest that social media can be an effective engagement tool, helping to minimise overall and selective attrition in longitudinal studies, particularly for hard-to-reach populations.&lt;/p&gt; &lt;p&gt;Keywords: Facebook; social media; longitudinal; tracing; parents; attrition; selective attrition&lt;/p&gt; &lt;hd id=&quot;AN0149012894-2&quot;&gt;Introduction&lt;/hd&gt; &lt;p&gt;Loss of participants in longitudinal studies is a major challenge for researchers (Young et al., [&lt;reflink idref=&quot;bib33&quot; id=&quot;ref1&quot;&gt;33&lt;/reflink&gt;]) and often disproportionately affects disadvantaged groups, such as families experiencing multiple or complex social adversities (McDonald et al., [&lt;reflink idref=&quot;bib20&quot; id=&quot;ref2&quot;&gt;20&lt;/reflink&gt;]). Such attrition can introduce bias, hampering the identification of true intervention effects (Asendorpf et al., [&lt;reflink idref=&quot;bib1&quot; id=&quot;ref3&quot;&gt;1&lt;/reflink&gt;]). Participants who are difficult to reach or re-engage often represent more marginalised sub-groups, given that complex social circumstances can contribute to frequent changes in residential addresses and phone numbers (Bonevski et al., [&lt;reflink idref=&quot;bib7&quot; id=&quot;ref4&quot;&gt;7&lt;/reflink&gt;]). Offline methods (e.g., phone calls, mailouts) are increasingly viewed as costly, time-consuming, and of diminishing effectiveness (Harris et al., [&lt;reflink idref=&quot;bib11&quot; id=&quot;ref5&quot;&gt;11&lt;/reflink&gt;]). Social media platforms like Facebook are rapidly becoming a recognized research tool, not only for recruitment and retention, but for minimising attrition through tracing (e.g., Harris et al., [&lt;reflink idref=&quot;bib11&quot; id=&quot;ref6&quot;&gt;11&lt;/reflink&gt;]; Kayrouz et al., [&lt;reflink idref=&quot;bib17&quot; id=&quot;ref7&quot;&gt;17&lt;/reflink&gt;]). Evidence of tracing effectiveness remains limited, with procedures often reported without sufficient detail to permit replication (Hokke et al., [&lt;reflink idref=&quot;bib13&quot; id=&quot;ref8&quot;&gt;13&lt;/reflink&gt;]). This paper describes the process and feasibility of locating and re-engaging otherwise lost-to-follow up families five years after their participation in a randomised controlled trial (RCT) of an early childhood parenting program, and the impact on study retention and selective attrition.&lt;/p&gt; &lt;p&gt;Internet and social media use is widespread in Australia, and while inequities in access persist (i.e., &#39;the digital divide&#39;) (Thomas et al., [&lt;reflink idref=&quot;bib31&quot; id=&quot;ref9&quot;&gt;31&lt;/reflink&gt;]), these have greatly reduced (Baker et al., [&lt;reflink idref=&quot;bib3&quot; id=&quot;ref10&quot;&gt;3&lt;/reflink&gt;]). For example, census data reveal that between 2016–17, 97% of households with children under 15 years had internet access (Australian Bureau of Statistics, [&lt;reflink idref=&quot;bib2&quot; id=&quot;ref11&quot;&gt;2&lt;/reflink&gt;]). Facebook remains the most popular social media platform in Australia, with 15 million active users, and 50% of the population logging in daily (Social Media News, [&lt;reflink idref=&quot;bib28&quot; id=&quot;ref12&quot;&gt;28&lt;/reflink&gt;]). Users can control the privacy settings on their account; some may choose to share their profile publicly, whereas others are unsearchable and visible only to users they choose to &#39;friend&#39;. All users must agree to terms and conditions, including policies which grant Facebook &#39;permission to use content that you create and share&#39; and &#39;permission to use your name, profile picture and information about your actions with ads and sponsored content&#39; (Facebook, [&lt;reflink idref=&quot;bib9&quot; id=&quot;ref13&quot;&gt;9&lt;/reflink&gt;]).&lt;/p&gt; &lt;p&gt;Strong Facebook usage amongst Australian families provides new opportunities to locate and re-engage participants in longitudinal studies, where offline methods (e.g., telephone and mail) have been unsuccessful. Facebook profiles are often more stable over time than phone numbers and home addresses, which change more frequently among disadvantaged populations (Bolanos et al., [&lt;reflink idref=&quot;bib6&quot; id=&quot;ref14&quot;&gt;6&lt;/reflink&gt;]; Mychasiuk &amp;amp; Benzies, [&lt;reflink idref=&quot;bib22&quot; id=&quot;ref15&quot;&gt;22&lt;/reflink&gt;]). It also allows participants to access and respond to messages when it is convenient, using any device with an internet connnection (Mychasiuk &amp;amp; Benzies, [&lt;reflink idref=&quot;bib22&quot; id=&quot;ref16&quot;&gt;22&lt;/reflink&gt;]). Although Facebook is increasingly being used for recruitment (e.g., Harris et al., [&lt;reflink idref=&quot;bib11&quot; id=&quot;ref17&quot;&gt;11&lt;/reflink&gt;]; Kayrouz et al., [&lt;reflink idref=&quot;bib17&quot; id=&quot;ref18&quot;&gt;17&lt;/reflink&gt;]), tracing remains in its infancy. A recent scoping review of child and family research identified 65 articles published between 2006 and 2016 which used the internet to recruit, retain or trace, of which only six utilised tracing (Hokke et al., [&lt;reflink idref=&quot;bib13&quot; id=&quot;ref19&quot;&gt;13&lt;/reflink&gt;]).&lt;/p&gt; &lt;p&gt;Participant tracing typically involves searching on known information such as names, birthdate, last known location, and known family members or friends (e.g., Masson et al., [&lt;reflink idref=&quot;bib19&quot; id=&quot;ref20&quot;&gt;19&lt;/reflink&gt;]; Rhodes &amp;amp; Marks, [&lt;reflink idref=&quot;bib25&quot; id=&quot;ref21&quot;&gt;25&lt;/reflink&gt;]). This process can raise ethical concerns for researchers and ethics committee members who report challenges in applying existing ethical guidelines within the context of social media and the online sphere (Crawford et al., [&lt;reflink idref=&quot;bib8&quot; id=&quot;ref22&quot;&gt;8&lt;/reflink&gt;]; Hibbin et al., [&lt;reflink idref=&quot;bib12&quot; id=&quot;ref23&quot;&gt;12&lt;/reflink&gt;]). Tracing raises questions about participants&#39; right to privacy, boundaries around consent, and the potential intrusiveness of being contacted online (Bhatia-Lin et al., [&lt;reflink idref=&quot;bib5&quot; id=&quot;ref24&quot;&gt;5&lt;/reflink&gt;]; Bolanos et al., [&lt;reflink idref=&quot;bib6&quot; id=&quot;ref25&quot;&gt;6&lt;/reflink&gt;]; Schneider et al., [&lt;reflink idref=&quot;bib26&quot; id=&quot;ref26&quot;&gt;26&lt;/reflink&gt;]). Understandings about how research is traditionally conducted may conflict with preferences for online communication and information sharing. Although Facebook users agree to terms and conditions, and can control their profile privacy settings, users may not fully comprehend the implications of publicly sharing personal information online (Masson et al., [&lt;reflink idref=&quot;bib19&quot; id=&quot;ref27&quot;&gt;19&lt;/reflink&gt;]). Studies using the internet for tracing provide scant detail on the process of generating evidence and identifying the &#39;correct&#39; person (Hokke et al., [&lt;reflink idref=&quot;bib13&quot; id=&quot;ref28&quot;&gt;13&lt;/reflink&gt;]). For example, how do researchers locate and document evidence? How is the evidence reviewed and by whom? What level of evidence is required before contacting a potential participant identified online?&lt;/p&gt; &lt;p&gt;The few published studies suggest that Facebook tracing is moderately effective. It is generally implemented either (i) as part of a suite of engagement strategies throughout the study (e.g., Mohan et al., [&lt;reflink idref=&quot;bib21&quot; id=&quot;ref29&quot;&gt;21&lt;/reflink&gt;]) or (ii) as a &#39;last-resort&#39; when offline engagement strategies have been unsuccessful (e.g., Mychasiuk &amp;amp; Benzies, [&lt;reflink idref=&quot;bib22&quot; id=&quot;ref30&quot;&gt;22&lt;/reflink&gt;]). The proportion of participants successfully located on Facebook where existing contact means were unsuccessful is around 21–45% (Bolanos et al., [&lt;reflink idref=&quot;bib6&quot; id=&quot;ref31&quot;&gt;6&lt;/reflink&gt;]; Jones et al., [&lt;reflink idref=&quot;bib16&quot; id=&quot;ref32&quot;&gt;16&lt;/reflink&gt;]; Nwadiuko et al., [&lt;reflink idref=&quot;bib24&quot; id=&quot;ref33&quot;&gt;24&lt;/reflink&gt;]), with overall study participation rates enhanced by 1–16% (Bolanos et al., [&lt;reflink idref=&quot;bib6&quot; id=&quot;ref34&quot;&gt;6&lt;/reflink&gt;]; Hurwitz et al., [&lt;reflink idref=&quot;bib14&quot; id=&quot;ref35&quot;&gt;14&lt;/reflink&gt;]; Jones et al., [&lt;reflink idref=&quot;bib16&quot; id=&quot;ref36&quot;&gt;16&lt;/reflink&gt;]; Mychasiuk &amp;amp; Benzies, [&lt;reflink idref=&quot;bib22&quot; id=&quot;ref37&quot;&gt;22&lt;/reflink&gt;]; Nwadiuko et al., [&lt;reflink idref=&quot;bib24&quot; id=&quot;ref38&quot;&gt;24&lt;/reflink&gt;]). Such variation may be attributable to the popularity of Facebook within the target population, timing of tracing attempts (i.e. as an initial or &#39;last-resort&#39; strategy), and changes to Facebook functionality over time (e.g., messaging functionality) (Hutchinson, [&lt;reflink idref=&quot;bib15&quot; id=&quot;ref39&quot;&gt;15&lt;/reflink&gt;]), likely impacting the effectiveness of tracing and limiting between-study comparisons. Evidence regarding demographic differences between participants engaged through offline methods (e.g., mail, phone) versus Facebook tracing is mixed (e.g., Mychasiuk &amp;amp; Benzies, [&lt;reflink idref=&quot;bib22&quot; id=&quot;ref40&quot;&gt;22&lt;/reflink&gt;]; Rhodes &amp;amp; Marks, [&lt;reflink idref=&quot;bib25&quot; id=&quot;ref41&quot;&gt;25&lt;/reflink&gt;]).&lt;/p&gt; &lt;p&gt;In this paper, we report on the use of Facebook tracing to re-engage families who participated five years earlier in the RCT of an early childhood parenting program. This paper aims to:&lt;/p&gt; &lt;p&gt;&lt;/p&gt; &lt;ulist&gt; &lt;item&gt; Describe the process used to conduct Facebook tracing for &lt;emph&gt;n&lt;/emph&gt; = 90 uncontactable participants; and&lt;/item&gt; &lt;p&gt;&lt;/p&gt; &lt;item&gt; Evaluate the effectiveness of this process by examining: (a) participants located on Facebook; (b) participants re-engaged using Facebook, and (c) response bias between three groups of participants: those who did not require Facebook tracing, those successfully re-engaged through Facebook, and those who could not be re-engaged through Facebook (using factors previously found to be related to tracing outcomes: parent/child gender, parent/child age, single parent status, household unemployment, Non-English speaking background, household income, receipt of Government benefits, parent education, socio-economic status of residential location, psychological distress, and parenting self-efficacy).&lt;/item&gt; &lt;/ulist&gt; &lt;hd id=&quot;AN0149012894-3&quot;&gt;Method&lt;/hd&gt; &lt;p&gt;&lt;/p&gt; &lt;hd id=&quot;AN0149012894-4&quot;&gt;Early Home Learning Study &amp;amp; EHLS at School Study&lt;/hd&gt; &lt;p&gt;The &lt;emph&gt;Early Home Learning Study&lt;/emph&gt; (&lt;emph&gt;EHLS&lt;/emph&gt;, 2010–2013) was a cluster RCT evaluating the &lt;emph&gt;smalltalk&lt;/emph&gt; program, involving 2,228 parents of infants (6–12 months) and toddlers (12–36 months) in the state of Victoria, Australia (Hackworth et al., [&lt;reflink idref=&quot;bib10&quot; id=&quot;ref42&quot;&gt;10&lt;/reflink&gt;]; Nicholson et al., [&lt;reflink idref=&quot;bib23&quot; id=&quot;ref43&quot;&gt;23&lt;/reflink&gt;]). The &lt;emph&gt;smalltalk&lt;/emph&gt; program aimed to support parents to provide an enriched home learning environment and thereby, to promote their child&#39;s language and literacy development and school readiness. Parents were recruited based on risk factors for poor child development (young parents, single parents, low household income, recipient of Government benefits, socially isolated, or culturally and linguistically diverse backgrounds). During initial recruitment, parents completed a Consent Form, on which they indicated their willingness to be contacted about future research.&lt;/p&gt; &lt;p&gt;Commencing in 2016, &lt;emph&gt;EHLS at School&lt;/emph&gt; (Westrupp et al., [&lt;reflink idref=&quot;bib32&quot; id=&quot;ref44&quot;&gt;32&lt;/reflink&gt;]) aimed to investigate child outcomes at age 7.5 years for the toddler cohort only. Given that children were aged 12 to 36 months at recruitment, they reached 7.5 years of age between 2016 and 2018. As part of &lt;emph&gt;EHLS at School&lt;/emph&gt;, families were invited to complete one home visit, including a parent-child observation, parent/child questionnaires, and direct child assessments. In February 2016, all eligible families (&lt;emph&gt;n&lt;/emph&gt; = 990) were sent a study information pack to their last known address. When children approached 7.5 years of age, attempts were made to contact participants via phone to confirm their willingness to participate in &lt;emph&gt;EHLS at School&lt;/emph&gt;. Where phone or postal contact methods were unsuccessful, we attempted to reach participants via alternative contacts (e.g., family or friends) whose details were provided during the original study.&lt;/p&gt; &lt;p&gt;Ethical approval for &lt;emph&gt;EHLS at School&lt;/emph&gt; was granted by La Trobe University Human Research Ethics Committee (15–028). An ethics modification to conduct Facebook tracing was approved by the Committee in November 2017.&lt;/p&gt; &lt;p&gt;Participants were eligible for &lt;emph&gt;EHLS at School&lt;/emph&gt; if: they had participated in the original study without actively withdrawing; had not declined future research contact; had completed the baseline assessments at initial recruitment; had completed either the post-assessment or 5-month follow-up assessment; and where the focus child was below 8.5 years.&lt;/p&gt; &lt;p&gt;Facebook tracing was attempted for &lt;emph&gt;n&lt;/emph&gt; = 90 participants who were still eligible to participate at the time that ethics clearance for tracing was granted (i.e. parents whose focus child was aged under 8.5 years, where the parent could not be reached through existing contact methods). Tracing was not attempted for participants where there was evidence of passive refusal (e.g., hanging up the phone, or voicemail indicating we had the correct number). In order to compare participants for whom tracing was required or not required, we describe all participants using baseline demographics collected at initial recruitment for the original &lt;emph&gt;EHLS&lt;/emph&gt;.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-5&quot;&gt;Facebook account and page&lt;/hd&gt; &lt;p&gt;An &lt;emph&gt;EHLS at School&lt;/emph&gt; Facebook account and Facebook Page were established, with permission and support from the La Trobe University Social Media Team. The publicly available Page featured the University logo, study logo, and a stock image of a parent and child. The &#39;About&#39; section included a brief description of the study, ethical approval details, and contact details for the research team. A post was pinned to the Facebook page: &#39;If you participated in the &lt;emph&gt;Early Home Learning Study&lt;/emph&gt; (2010–2013) and your child turns 7.5 in 2017 or 2018 and you would be interested in participating in the follow up, please contact us.&#39; Page settings were adjusted to minimise moderation requirements; i.e., the profanity filter was set to &#39;strong&#39; and posts were disabled.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-6&quot;&gt;Locating participants on Facebook&lt;/hd&gt; &lt;p&gt;Facebook searches were conducted immediately following ethics clearance, from December 2017 until September 2018, to generate pieces of evidence (referred to here as &#39;indicators&#39;) from which to identify &#39;lost&#39; participants. As outlined in Table 1, searches aimed to identify our participants&#39; Facebook profiles using seven types of indicators: (i) parent name; (ii) child name; (iii) images that illustrate parenting status and/or age of child; (iv) town, suburb or city of residence; (v) connections with known family members or friends; (vi) birthdays, birthyear or age; and (vii) images matching our study-collected video footage.&lt;/p&gt; &lt;p&gt;Table 1. Indicators used to identify participants on Facebook&lt;/p&gt; &lt;p&gt; &lt;ephtml&gt; &amp;lt;table&amp;gt;&amp;lt;thead&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Matching Indicator&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Description&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Frequency for Participants Successfully Re-engaged, n (%)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;/thead&amp;gt;&amp;lt;tbody&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;1. Parent&#39;s first/last name&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Profile name and/or Facebook username (URL) matches the study database (note: this could include abbreviated names, e.g., &#39;Bec&#39; instead of &#39;Rebecca&#39;).&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;19 (100)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;2. Child&#39;s first/last name&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;A photo caption, post or comment states the child&#39;s name or nickname, which matches the study database.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;16 (84.2)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;3. Images showing evidence of parenting status/age of child&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;A profile photo or post demonstrating visual evidence of the user&#39;s parenting status, and/or the age of the child.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;18 (94.7)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;4. Location of residence&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Profile states &#39;Lives in &amp;lt;italic&amp;gt;[suburb/town]&amp;lt;/italic&amp;gt;&#39; or participant has &#39;liked&#39; several pages related to &amp;lt;italic&amp;gt;[suburb/town]&amp;lt;/italic&amp;gt;.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;15 (78.9)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;5. Connections with known friends/relatives&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Facebook user is &#39;friends&#39; with other users who are listed as alternative contacts in our database, or these names appear in posts or tags.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;6 (31.6)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;6. Birthdays/age&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Profile information or posts reveal a parent or child&#39;s birthday which matches the study database.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;8 (42.1)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;7. Images matching study-collected video data&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Parent&#39;s or child&#39;s appearance in profile or posted photos resemble their appearance in video-recorded parent-child interactions from earlier study data collection. Where available, photos posted on Facebook at the time of the original study can be used to more accurately compare to video footage.&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;8 (42.1)&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;/tbody&amp;gt;&amp;lt;/table&amp;gt; &lt;/ephtml&gt; &lt;/p&gt; &lt;p&gt;1 Note: Frequency is the number of participants for whom each indicator was found (for the 19 participants re-engaged using Facebook).&lt;/p&gt; &lt;p&gt;Searches were conducted at least once per week from the study Facebook account by several team members familiar with Facebook. Only publicly available information was viewed. We did not add any Facebook users as &#39;friends&#39;, to protect their privacy as research participants. Google searches were also conducted to identify any supplementary information for these seven types of indicators; as a means to verify an identified profile or to provide additional search strategies to try in Facebook. Evidence was recorded in a centrally located password-protected Excel spreadsheet.&lt;/p&gt; &lt;p&gt;Searching was prioritised for participants with the eldest children, before they became ineligible. Initial attempts to identify a matching Facebook Profile involved name searches, including other known surnames or possible nicknames. We also searched for alternative contacts provided during the original study; either featured in a user&#39;s friends list, or in posts or tags. To confirm potential matches, we examined a user&#39;s &#39;Liked Pages&#39; &#39;Groups&#39; or &#39;Check-ins&#39; to verify the location of the participant, compared to their last known residential address.&lt;/p&gt; &lt;p&gt;Given the dynamic nature of social media, regular searching was required. At any given time, users can post new content, interact with other users, or adjust their settings; each change provides a potential opportunity to obtain evidence for tracing. For example, birthday-related posts provide distinctive information that can be cross-checked with existing date-of-birth records.&lt;/p&gt; &lt;p&gt;The strength of evidence for a potential match was summarised as: weak (0 or 1 indicator); moderate (2 indicators); or strong (3+ indicators). Colour-coding was used to support classification (Red, Yellow and Green, respectively). Participants were categorised and re-categorised during searches, as more evidence was located (e.g., a Red participant was subsequently updated to Green if 3+ indicators were found). Evidence was reviewed at fortnightly research team meetings. Participants categorised as Red with a child over 8.5 years of age were considered &#39;uncontactable&#39; and no further tracing attempts were made. Participants categorised as Green or Yellow were individually reviewed. The team carefully considered available evidence for each participant, before subsequently: (i) sending a private Facebook message to the identified user; or (ii) continuing to search for evidence. On a small number of occasions, team discussion resulted in upgrading a Yellow to a Green if the parent and/or child names were particularly uncommon (for example, if a Facebook search only resulted in one profile with the correct name, and the profile suggested the user had a child of the same age/gender that was reflected in the study database).&lt;/p&gt; &lt;hd id=&quot;AN0149012894-7&quot;&gt;Sending Facebook messages&lt;/hd&gt; &lt;p&gt;Participants categorised as Green (i.e. strong evidence) were contacted through Facebook Messenger using the &lt;emph&gt;EHLS at School&lt;/emph&gt; account. A message template was prepared, based on the protocol described by Mychasiuk and Benzies ([&lt;reflink idref=&quot;bib22&quot; id=&quot;ref45&quot;&gt;22&lt;/reflink&gt;]). Messages could only be viewed by the Facebook user. The message included the following content: (i) asked if they participated in the &lt;emph&gt;Early Home Learning Study&lt;/emph&gt;; (ii) emphasised that we have exhausted all other means of contact; (iii) provided our telephone number and email address; (iv) requested that the parent contacts us or replies via Facebook Messenger regarding their willingness to participate in &lt;emph&gt;EHLS at School&lt;/emph&gt;; and (v) apologised in advance for possibly contacting the wrong individual and request that, if this was the case, to reply to the message to ensure that they are not contacted in the future. For any non-responses, a reminder message was sent one week later. For non-responses 14 days after the reminder message, participants were considered uncontactable and no further contact attempts were made. No more than three messages of either kind (initial or reminder) were sent from the &lt;emph&gt;EHLS at School&lt;/emph&gt; account on any given day, to prevent messages from being interpreted as &#39;spam&#39;, given that we were not Facebook &#39;friends&#39; with those we messaged.&lt;/p&gt; &lt;p&gt;Participants remained active on the tracing list until they either (i) became out of scope (child ≥8.5 years), or (ii) did not respond to the initial or reminder messages. Data collection concluded in September 2018; at which time all remaining Yellow and Red participants were considered uncontactable.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-8&quot;&gt;Statistical analyses&lt;/hd&gt; &lt;p&gt;Descriptive statistics and tests were conducted using StataSE Version 14 (StataCorp, [&lt;reflink idref=&quot;bib29&quot; id=&quot;ref46&quot;&gt;29&lt;/reflink&gt;]), based on three participant groups: &lt;emph&gt;Group 1&lt;/emph&gt;, participants re-engaged using existing contact details; &lt;emph&gt;Group 2&lt;/emph&gt;, participants for whom tracing was attempted but unsuccessful (either not located or did not respond to our message); and &lt;emph&gt;Group 3&lt;/emph&gt;, participants who were successfully re-engaged through Facebook. These groups were selected to align with and extend evidence regarding the characteristics of participants who become lost-to-follow-up. Group 1 enabled us to compare &#39;lost&#39; participants with participants who were relatively easy to re-engage using existing contact details. We also wanted to extend current evidence about tracing effectiveness by comparing characteristics for whom tracing was successful or unsuccessful. Baseline data were used to compare the groups (collected at initial recruitment). While it is possible that participants&#39; circumstances had changed since initial recruitment, baseline variables were required to conduct between-group comparisons at the same time-point. Groups were compared using independent-samples t-tests for continuous measures, and chi-squared tests for categorical variables. This approach was employed for clarity and ease of interpretation, and because ANOVA with post-hoc tests would not add further robustness.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-9&quot;&gt;Results&lt;/hd&gt; &lt;p&gt;&lt;/p&gt; &lt;hd id=&quot;AN0149012894-10&quot;&gt;Participants located on Facebook&lt;/hd&gt; &lt;p&gt;As shown in Figure 1, most participants were located on Facebook with a strong level of evidence (i.e. at least three indicators, &lt;emph&gt;n&lt;/emph&gt; = 63/90, 70.0%). We found moderate evidence for a further 6 participants (i.e. two indicators, 6.7%) and no or little evidence for 21 participants (i.e. no indicators or one indicator, 23.3%).&lt;/p&gt; &lt;p&gt;Graph: Figure 1. Participant flow for n = 90 participants for whom Facebook tracing was attempted&lt;/p&gt; &lt;hd id=&quot;AN0149012894-11&quot;&gt;Participants re-engaged using Facebook&lt;/hd&gt; &lt;p&gt;The participants located on Facebook with a strong level of evidence (&lt;emph&gt;n&lt;/emph&gt; = 63) were sent a private Facebook message. Of these messages, &lt;emph&gt;n&lt;/emph&gt; = 32/63 (50.8%) responses were received, of which 20 responded on the same day, eight responded within two weeks, and four responded after 35, 80, 141 and 207 days (mean response time: 15.7 days). Of those who responded, &lt;emph&gt;n&lt;/emph&gt; = 22/32 (68.8%) agreed to participate, &lt;emph&gt;n&lt;/emph&gt; = 5/32 (15.6%) declined, and &lt;emph&gt;n&lt;/emph&gt; = 5/32 (15.6%) did not respond further. Of those who declined: two no longer had custody of the focus child; two could not recall participating in the original study; and one accepted our message request but posted a &#39;thumbs down&#39; emoji against our message. Of those who agreed to participate, &lt;emph&gt;n&lt;/emph&gt; = 19/22 (86.4%) provided data (via a home visit or online parent questionnaire).&lt;/p&gt; &lt;p&gt;Participants who were successfully re-engaged in the study represented 21.1% (19/90) of the participants for whom tracing was attempted, and 1.9% (19/990) of the entire eligible study sample. For these 19 participants, Table 1 shows the frequency counts for the seven types of evidence. We found evidence for an average of 4.7 of the 7 indicators (ranging from 3 to 7). Although we did not keep detailed logs of the time required to conduct Facebook tracing, we reviewed our records to assess the relative ease or difficulty of locating these 19 participants. Of this group, &lt;emph&gt;n&lt;/emph&gt; = 14/19 (73.7%) were located during the first rigorous search attempt, &lt;emph&gt;n&lt;/emph&gt; = 2/19 (10.5%) were located after the second search attempt, and &lt;emph&gt;n&lt;/emph&gt; = 3/19 (15.8%) were located after three or more search attempts.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-12&quot;&gt;Response bias&lt;/hd&gt; &lt;p&gt;Table 2 presents the baseline demographic characteristics for the three groups. Compared to those re-engaged using existing contact details (&lt;emph&gt;Group 1&lt;/emph&gt;), participants on the tracing list (&lt;emph&gt;Groups 2&lt;/emph&gt; and &lt;emph&gt;3&lt;/emph&gt;) were more likely at baseline to have a younger child, to be single, living in a household without paid employment, with low household income, and without completion of Year-12-level (secondary) education. There were no differences in baseline demographics between the successful and unsucessful tracing groups (&lt;emph&gt;Groups 2&lt;/emph&gt; and &lt;emph&gt;3&lt;/emph&gt;). Finally, there were no clear differences between the three groups on psychological distress or parenting self-efficacy, however there was some evidence of greater psychological distress amongst participants on the Facebook tracing list (&lt;emph&gt;Groups 2&lt;/emph&gt; and &lt;emph&gt;3&lt;/emph&gt;), compared to participants engaged offline (&lt;emph&gt;Group 1&lt;/emph&gt;).&lt;/p&gt; &lt;p&gt;Table 2. Demographic characteristics at initial recruitment for participants re-engaged using existing contact details, and for participants for whom Facebook tracing was successful and unsuccessful&lt;/p&gt; &lt;p&gt; &lt;ephtml&gt; &amp;lt;table&amp;gt;&amp;lt;thead&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td /&amp;gt;&amp;lt;td&amp;gt;Offline&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;Facebook&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;p&amp;lt;/italic&amp;gt; values&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td /&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 1&amp;lt;/italic&amp;gt; Existing Contact Details &amp;lt;italic&amp;gt;n&amp;lt;/italic&amp;gt; = 651&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 2&amp;lt;/italic&amp;gt; Tracing Unsuccessful &amp;lt;italic&amp;gt;n&amp;lt;/italic&amp;gt; = 71&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 3&amp;lt;/italic&amp;gt; Tracing Successful &amp;lt;italic&amp;gt;n&amp;lt;/italic&amp;gt; = 19&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 1&amp;lt;/italic&amp;gt; Vs &amp;lt;italic&amp;gt;Group 2&amp;lt;/italic&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 1&amp;lt;/italic&amp;gt; Vs &amp;lt;italic&amp;gt;Group 3&amp;lt;/italic&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;italic&amp;gt;Group 2&amp;lt;/italic&amp;gt; Vs &amp;lt;italic&amp;gt;Group 3&amp;lt;/italic&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;/thead&amp;gt;&amp;lt;tbody&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Parent, male, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;24 (3.7)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;3(4.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;1 (5.3)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.82&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.72&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.85&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Child, male, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;323 (49.6)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;42 (59.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;8 (42.1)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.13&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.52&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.18&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Parent age, years, m(sd)*&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;34.1 (5.4)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;30.2 (5.7)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;32.1 (6.4)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.12&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.21&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Child age, months, m(sd)*&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;22.1 (7.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;17.9 (4.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;18.2 (5.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.&amp;lt;bold&amp;gt;02&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.82&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Single parent, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;51 (7.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;14 (19.7)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;4 (21.1)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.01&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.&amp;lt;bold&amp;gt;04&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.90&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Household unemployed, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;54 (8.3)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;19 (26.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;6 (31.6)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.68&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;LOTE, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;202 (31.0)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;25 (35.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;8 (42.1)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.47&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.31&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.58&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Low household income, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;97 (15.3)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;28 (40.6)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;7 (36.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.&amp;lt;bold&amp;gt;01&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.77&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Receiving Government benefits, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;77 (11.9)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;24 (33.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;5 (26.3)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.06&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.54&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Education &amp;amp;#60; Year 12, n(%)&amp;lt;sup&amp;gt;^&amp;lt;/sup&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;36 (5.5)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;19 (26.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;5 (26.3)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;&amp;lt;bold&amp;gt;&amp;amp;#60;.001&amp;lt;/bold&amp;gt;&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.97&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;SEIFA, m(sd)*&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;985.0 (45.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;984.0 (40.8)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;986.4 (45.1)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.86&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.89&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.82&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Kessler-6, m(sd)*&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;3.6 (3.6)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;4.4 (3.6)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;4.8 (5.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.05&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.13&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.69&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;tr&amp;gt;&amp;lt;td&amp;gt;Parenting self-efficacy, m(sd)*&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;16.9 (2.5)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;16.8 (2.1)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;17.0 (2.2)&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.88&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.82&amp;lt;/td&amp;gt;&amp;lt;td&amp;gt;.74&amp;lt;/td&amp;gt;&amp;lt;/tr&amp;gt;&amp;lt;/tbody&amp;gt;&amp;lt;/table&amp;gt; &lt;/ephtml&gt; &lt;/p&gt; &lt;ulist&gt; &lt;item&gt;2 Notes: Bold denotes p values &amp;lt;.05; ^chi-squared test; *independent samples t-test.&lt;/item&gt; &lt;item&gt;3 Household unemployment: Both parents not in paid employment or single parent not in paid employment; LOTE: Language Other Than English spoken at home; Low household income: AUD$36,400 per annum or less; Year 12: Final year of secondary education in Australia; SEIFA: Socio-Economic Index for Areas – Disadvantage, where the Australian mean is 1000, SD is 50, lower scores indicate greater disadvantage; Kessler-6: 6-item measure of psychological distress, with higher scores indicating greater distress, possible range 0–24 (Kessler et al., [&lt;reflink idref=&quot;bib18&quot; id=&quot;ref47&quot;&gt;18&lt;/reflink&gt;]); Parenting self-efficacy: 4-item measure on a 5-point scale, with higher scores indicating greater parenting self-efficacy, possible range 4–20 (Zubrick et al., [&lt;reflink idref=&quot;bib34&quot; id=&quot;ref48&quot;&gt;34&lt;/reflink&gt;]).&lt;/item&gt; &lt;/ulist&gt; &lt;hd id=&quot;AN0149012894-13&quot;&gt;Discussion&lt;/hd&gt; &lt;p&gt;This paper describes how Facebook was used to locate and re-engage participants who were uncontactable using previously collected phone and postal details, more than five years after participation in the RCT of an early childhood parenting intervention. Over two-thirds of uncontactable participants were located, resulting in the re-engagement of 19 participants. Findings suggest that Facebook shows promise as a tool to minimise attrition in longitudinal research, however, we acknowledge several limitations and challenges.&lt;/p&gt; &lt;p&gt;We received responses to half of our Facebook messages; it remains unclear whether the remaining messages were ignored or not seen by the intended recipients. One of the limitations of Facebook tracing (given its current functionality) is that messages from non-friends appear in a Message Requests folder which is often referred to as a &#39;secret folder&#39; because many users are either not aware of it, or do not know how to access it (Shead, [&lt;reflink idref=&quot;bib27&quot; id=&quot;ref49&quot;&gt;27&lt;/reflink&gt;]). The Message Requests folder is a relatively new strategy implemented by Facebook in 2015 to minimise &#39;spam&#39; from non-friends. Depending on profile settings, users may not receive a notification and thus remain unaware the message exists. Indeed, four participants responded after a lengthy delay (up to 7 months). Participants who &lt;emph&gt;did&lt;/emph&gt; respond to our message appeared comfortable with this method of communication and subsequently updated their contact details. Some participants reported positive experiences with the original study and were pleased that we had located them via Facebook.&lt;/p&gt; &lt;p&gt;Comparisons between our findings and the few published studies are hampered by inconsistent reporting and procedures. Tracing &#39;success&#39; can be described in several ways, such as the number of participants who: (i) are located on Facebook, (ii) respond to a message, (iii) agree to participate; and (iv) provide data. Further, Facebook tracing can be implemented from the outset as part of a suite of engagement strategies or as a &#39;last-resort&#39; when other contact options have been exhausted. Our approach was the latter; we sought ethical approval to use Facebook tracing after data collection was well-underway, therefore some children were already out of scope (≥8.5 years). As noted by Mychasiuk and Benzies ([&lt;reflink idref=&quot;bib22&quot; id=&quot;ref50&quot;&gt;22&lt;/reflink&gt;]), tracing conducted from the outset may have enhanced our success rate. Nonetheless, our findings compare favourably to other tracing studies (e.g., Jones et al., [&lt;reflink idref=&quot;bib16&quot; id=&quot;ref51&quot;&gt;16&lt;/reflink&gt;]; Nwadiuko et al., [&lt;reflink idref=&quot;bib24&quot; id=&quot;ref52&quot;&gt;24&lt;/reflink&gt;]). Between-study differences may reflect the growing popularity of social media over time, changes in Facebook functionality (especially private messaging), as well as the proportion of active Facebook users within the study&#39;s target population.&lt;/p&gt; &lt;p&gt;Tracing provided a modest (2%) increase in our overall participation rate, but did minimise &lt;emph&gt;selective&lt;/emph&gt; attrition. This would almost certainly be higher, had tracing been implemented sooner, or as part of initial re-engagement strategies. Participants on our tracing list represented a more disadvantaged sub-group of the full cohort. These participants were more likely at recruitment to be a single parent with a younger child, living in a household without paid employment, with low household income, and without Year-12-level education. Although having a younger child is not necessarily considered to be a social risk factor, it is possible that this finding reflects the longer amount of time since last contact and thus a greater likelihood of having changed address or phone numbers (families were recruited between child age 12–36 months but were all re-engaged at child age 7.5 years). Overall, tracing helped to reduce potential bias in our sample, re-engaging a sub-group who would otherwise not have been as well represented in this research. Our findings are consistent with evidence that tracing is more likely to be required for participants who are single or without tertiary education (Rhodes &amp;amp; Marks, [&lt;reflink idref=&quot;bib25&quot; id=&quot;ref53&quot;&gt;25&lt;/reflink&gt;]). Findings speak to the potential value of social media for retaining and tracing the most marginalised participants; it is not necessarily that these participants are more frequent Facebook users, but that changing phone numbers and addresses make these methods of contact less effective. Researchers are encouraged to consider how social media can serve to &#39;future proof&#39; longitudinal cohorts. There is evidence that Facebook use amongst some age groups is declining (Sweney, [&lt;reflink idref=&quot;bib30&quot; id=&quot;ref54&quot;&gt;30&lt;/reflink&gt;]), and as such, researchers should be responsive to new and emerging platforms.&lt;/p&gt; &lt;p&gt;As reported elsewhere (Masson et al., [&lt;reflink idref=&quot;bib19&quot; id=&quot;ref55&quot;&gt;19&lt;/reflink&gt;]), tracing required a significant investment of time, however this varied enormously between participants. Nearly three-quarters of those successfully re-engaged were located during the first, rigorous search attempt. This suggests that tracing attempts over an extended period of time may not be an efficient use of limited resources. However, given the significant investment of time and resources during initial recruitment and intervention delivery, it could be argued that extensive tracing attempts are still warranted. Such decisions must be made within the context of the study, balancing capacity for tracing against the value of enhanced participation.&lt;/p&gt; &lt;p&gt;Search strategies likely improved and became more efficient throughout the tracing process. Children&#39;s birthdays or holiday periods were common times for posting content, often providing the extra indicator required to update the participant status to Green and allowing a private Facebook message to be sent. Thus, searches were often tailored for individual participants or holiday periods. Our strong success rate for locating participants on Facebook (70%) is likely attributable to relatively high Facebook usage among parent populations (Bartholomew et al., [&lt;reflink idref=&quot;bib4&quot; id=&quot;ref56&quot;&gt;4&lt;/reflink&gt;]), together with the broad range of search strategies carefully implemented by staff members familiar with Facebook over a 10-month period.&lt;/p&gt; &lt;p&gt;One of the main concerns regarding Facebook tracing, particularly among ethics committees, is the potential for participants to view tracing as intrusive or a breach of privacy (Crawford et al., [&lt;reflink idref=&quot;bib8&quot; id=&quot;ref57&quot;&gt;8&lt;/reflink&gt;]). Although we cannot comment on participants who did not respond, those who did respond expressed no concerns and appeared comfortable being contacted this way (with the exception of one participant who accepted our message request, then posted a &#39;thumbs down&#39; emoji). Similarly, Mychasiuk and Benzies ([&lt;reflink idref=&quot;bib22&quot; id=&quot;ref58&quot;&gt;22&lt;/reflink&gt;]) concluded that Facebook was not viewed by participants as intrusive or coercive. Although we did not collect data on participants&#39; responses to being contacted by Facebook Messenger, the 19 re-engaged participants responded positively during subsequent interactions, including at the home visit. Participants noted that they had enjoyed participating in the original study, which appeared to facilitate their re-engagement in the follow-up. Further research could examine the perspectives of participants contacted via Facebook. We agree with Bolanos et al. ([&lt;reflink idref=&quot;bib6&quot; id=&quot;ref59&quot;&gt;6&lt;/reflink&gt;]) that participants may find Facebook a more informal, less intrusive form of contact compared to a phone call; they can choose to ignore or decline the message request (without the sender&#39;s knowledge), or respond at a convenient time. Anecdotally, we found that several participants who were re-engaged via Facebook remained difficult to reach via phone (despite having confirmed their current contact number), yet were remarkably prompt to respond through Facebook Messenger. Indeed, Masson et al. ([&lt;reflink idref=&quot;bib19&quot; id=&quot;ref60&quot;&gt;19&lt;/reflink&gt;]) reported that some participants located using Facebook found it to be &#39;more secure and private ... than by telephone or letter&#39; (p.32).&lt;/p&gt; &lt;p&gt;Given that social media is the &#39;new communication normal&#39;, Facebook and other social media platforms are likely to be increasingly valuable elements of the research &#39;toolbox&#39;. We recommend that researchers consider the use of social media platforms during initial recruitment. Specific consent to contact participants on social media could be requested, along with the participant&#39;s profile URL (Schneider et al., [&lt;reflink idref=&quot;bib26&quot; id=&quot;ref61&quot;&gt;26&lt;/reflink&gt;]) which is known to be relatively stable and may prove useful for longitudinal studies (Bolanos et al., [&lt;reflink idref=&quot;bib6&quot; id=&quot;ref62&quot;&gt;6&lt;/reflink&gt;]). Consistent with Masson et al. ([&lt;reflink idref=&quot;bib19&quot; id=&quot;ref63&quot;&gt;19&lt;/reflink&gt;]), we found the names of alternative contacts provided during initial recruitment to be useful in supporting accurate and timely identification of participants on Facebook. Drawing on all existing information was important to carefully cross-check and verify profiles, in order to conduct tracing in an ethical manner. Efforts to obtain this information at recruitment (and updating alternative contacts at regular intervals) is encouraged. Future studies may also consider using social media as a primary (rather than &#39;last resort&#39;) engagement method.&lt;/p&gt; &lt;p&gt;Overall, our findings suggest that Facebook is a promising research tool for minimising both overall and selective attrition within longitudinal studies. The majority of our participants appeared to be active Facebook users; however we cannot assume that this is the case for other populations. We recommend that researchers consider implementing Facebook tracing early and as part of a broader suite of engagement strategies. Lastly, consulting with professionals regarding the ethical and technological aspects of tracing protocols would assist researchers to safeguard participants&#39; privacy/confidentiality and ensure that the functionality of the platform is used for optimal tracing effectiveness.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-14&quot;&gt;Acknowledgments&lt;/hd&gt; &lt;p&gt;We thank all participating parents, children and teachers; the EHLS at School staff including M&#233;abh Cullinane, Andisheh Vahedi, Anna Fragkoudi, Lisa Pascoe, Stacey Hokke, Laura Biggs and Ellen Koutsodimitropoulos; and the many staff involved with the original EHLS, including the Maternal and Child Health nurses, site coordinators, group facilitators, and home coaches, and staff from the Parenting Research Centre and the Victorian Government. The Early Home Learning Study (2009-2013) was commissioned and funded by the Victorian Government Department of Education and Early Childhood (now the Department of Education and Training) and was designed and conducted by the Parenting Research Centre. The authors acknowledge the contributions of the original study investigators: Honorary Professor Warren Cann, Doctor Naomi Hackworth, Associate Professor Jan Matthews, Professor Donna Berthelsen, and Professor Jan Nicholson. The views reported in this paper are those of the authors and do not necessarily represent the views of the Victorian Government or NHMRC.&lt;/p&gt; &lt;hd id=&quot;AN0149012894-15&quot;&gt;Disclosure statement&lt;/hd&gt; &lt;p&gt;No potential conflict of interest was reported by the authors.&lt;/p&gt; &lt;ref id=&quot;AN0149012894-16&quot;&gt; &lt;title&gt; References &lt;/title&gt; &lt;blist&gt; &lt;bibl id=&quot;bib1&quot; idref=&quot;ref3&quot; type=&quot;bt&quot;&gt;1&lt;/bibl&gt; &lt;bibtext&gt; Asendorpf, J. B., van de Schoot, R., Denissen, J. J. A., &amp;amp; Hutteman, R. (2014). Reducing bias due to systematic attrition in longitudinal studies: The benefits of multiple imputation. International Journal of Behavioral Development, 38 (5), 453 – 460. https://doi.org/10.1177/0165025414542713&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib2&quot; idref=&quot;ref11&quot; type=&quot;bt&quot;&gt;2&lt;/bibl&gt; &lt;bibtext&gt; Australian Bureau of Statistics. (2018). 8146.0 - Household Use of Information Technology, Australia, 2016-17. Retrieved from &lt;ulink href=&quot;http://www.abs.gov.au/ausstats/abs@.nsf/mf/8146.0&quot;&gt;http://www.abs.gov.au/ausstats/abs@.nsf/mf/8146.0&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib3&quot; idref=&quot;ref10&quot; type=&quot;bt&quot;&gt;3&lt;/bibl&gt; &lt;bibtext&gt; Baker, S., Sanders, M. R., &amp;amp; Morawska, A. (2017). Who uses online parenting support? A cross-sectional survey exploring Australian parents&#39; internet use for parenting. Journal of Child and Family Studies, 26 (3), 916 – 927. https://doi.org/10.1007/s10826-016-0608-1&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib4&quot; idref=&quot;ref56&quot; type=&quot;bt&quot;&gt;4&lt;/bibl&gt; &lt;bibtext&gt; Bartholomew, M. K., Schoppe‐Sullivan, S. J., Glassman, M., Kamp Dush, C. M., &amp;amp; Sullivan, J. M. (2012). New parents&#39; Facebook use at the transition to parenthood. Family Relations, 61 (3), 455 – 469. https://doi.org/10.1111/j.1741-3729.2012.00708.x&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib5&quot; idref=&quot;ref24&quot; type=&quot;bt&quot;&gt;5&lt;/bibl&gt; &lt;bibtext&gt; Bhatia-Lin, A., Boon-Dooley, A., Roberts, M. K., Pronai, C., Fisher, D., Parker, L., Engstrom, A., Ingraham, L., &amp;amp; Darnell, D. (2019). Ethical and regulatory considerations for using social media platforms to locate and track research participants. The American Journal of Bioethics, 19 (6), 47 – 61. https://doi.org/10.1080/15265161.2019.1602176&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib6&quot; idref=&quot;ref14&quot; type=&quot;bt&quot;&gt;6&lt;/bibl&gt; &lt;bibtext&gt; Bolanos, F., Herbeck, D., Christou, D., Lovinger, K., Pham, A., Raihan, A., ... Brecht, M. L. (2012). Using Facebook to maximize follow-up response rates in a longitudinal study of adults who use methamphetamine. Substance Abuse: Research and Treatment, 6 (1), 1 – 11. https://doi.org/10.4137/SART.S8485&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib7&quot; idref=&quot;ref4&quot; type=&quot;bt&quot;&gt;7&lt;/bibl&gt; &lt;bibtext&gt; Bonevski, B., Randell, M., Paul, C., Chapman, K., Twyman, L., Bryant, J., Brozek, I., &amp;amp; Hughes, C. (2014). Reaching the hard-to-reach: A systematic review of strategies for improving health and medical research with socially disadvantaged groups. BMC Medical Research Methodology, 14 (1), 42. https://doi.org/10.1186/1471-2288-14-42&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib8&quot; idref=&quot;ref22&quot; type=&quot;bt&quot;&gt;8&lt;/bibl&gt; &lt;bibtext&gt; Crawford, S. B., Hokke, S., Nicholson, J. M., Zion, L., Lucke, J. C., Keyzer, P., &amp;amp; Hackworth, N. (2018). &quot;It&#39;s not black and white&quot; Public health researchers&#39; and ethics committees&#39; perceptions of engaging research participants online. Internet Research, 29 (1), 123 – 143. https://doi.org/10.1108/IntR-07-2017-0278&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibl id=&quot;bib9&quot; idref=&quot;ref13&quot; type=&quot;bt&quot;&gt;9&lt;/bibl&gt; &lt;bibtext&gt; Facebook. (2018). Terms of service. https://&lt;ulink href=&quot;http://www.facebook.com/terms.php&quot;&gt;www.facebook.com/terms.php&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Hackworth, N., Berthelsen, D., Matthews, J., Westrupp, E., Cann, W., Ukoumunne, O., ... Nicholson, J. (2017). Impact of a brief group intervention to enhance parenting and the home learning environment for children aged 6-36 months: A cluster randomised controlled trial. Prevention Science, 18 (3), 337 – 349. https://doi.org/10.1007/s11121-017-0753-9&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Harris, M. L., Loxton, D., Wigginton, B., &amp;amp; Lucke, J. C. (2015). Recruiting online: Lessons from a longitudinal survey of contraception and pregnancy intentions of young Australian women. American Journal of Epidemiology, 181 (10), 737 – 746. https://doi.org/10.1093/aje/kwv006&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Hibbin, R. A., Samuel, G., &amp;amp; Derrick, G. E. (2018). From &quot;a fair game&quot; to &quot;a form of covert research&quot;: Research ethics committee members&#39; differing notions of consent and potential risk to participants within social media research. Journal of Empirical Research on Human Research Ethics, 13 (2), 149 – 159. https://doi.org/10.1177/1556264617751510&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Hokke, S., Hackworth, N. J., Quin, N., Bennetts, S. K., Win, H. Y., Nicholson, J. M., ... Crawford, S. B. (2018). Ethical issues in using the internet to engage participants in family and child research: A scoping review. PloS One, 13 (9), e0204572. https://doi.org/10.1371/journal.pone.0204572&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Hurwitz, L. B., Schmitt, K. L., &amp;amp; Olsen, M. K. (2017). Facilitating development research: Suggestions for recruiting and re-recruiting children and families. Frontiers in Psychology, 8, 1525. https://doi.org/10.3389/fpsyg.2017.01525&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Hutchinson, A. (2018). Facebook outlines what&#39;s coming for messenger, and messenger business, in 2018. Social Media Today. https://&lt;ulink href=&quot;http://www.socialmediatoday.com/news/facebook-outlines-whats-coming-for-messenger-and-messenger-business-in-2/514865/&quot;&gt;www.socialmediatoday.com/news/facebook-outlines-whats-coming-for-messenger-and-messenger-business-in-2/514865/&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Jones, L., Saksvig, B. I., Grieser, M., &amp;amp; Young, D. R. (2012). Recruiting adolescent girls into a follow-up study: Benefits of using a social networking website. Contemporary Clinical Trials, 33 (2), 268 – 272. https://doi.org/10.1016/j.cct.2011.10.011&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Kayrouz, R., Dear, B. F., Karin, E., &amp;amp; Titov, N. (2016). Facebook as an effective recruitment strategy for mental health research of hard to reach populations. Internet Interventions, 4, 1 – 10. https://doi.org/10.1016/j.invent.2016.01.001&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Kessler, R. C., Andrews, G., Colpe, L. J., Hiripi, E., Mroczek, D. K., Normand, S. L., ... Zaslavsky, A. M. (2002). Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychological Medicine, 32 (6), 959 – 976. https://doi.org/10.1017/S0033291702006074&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Masson, H., Balfe, M., Hackett, S., &amp;amp; Phillips, J. (2013). Lost without a trace? Social networking and social research with a hard-to-reach population. The British Journal of Social Work, 43 (1), 24 – 40. https://doi.org/10.1093/bjsw/bcr168&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; McDonald, L., FitzRoy, S., Fuchs, I., Fooken, I., &amp;amp; Klasen, H. (2012). Strategies for high retention rates of low-income families in FAST (Families and Schools Together): An evidence-based parenting programme in the USA, UK, Holland and Germany. European Journal of Developmental Psychology, 9 (1), 75 – 88. https://doi.org/10.1080/17405629.2011.632134&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Mohan, Y., Cornejo, M., Sidell, M., Smith, J., &amp;amp; Young, D. R. (2017). Re-recruiting young adult women into a second follow-up study. Contemporary Clinical Trials Communications, 5, 160 – 167. https://doi.org/10.1016/j.conctc.2017.02.006&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Mychasiuk, R., &amp;amp; Benzies, K. (2012). Facebook: An effective tool for participant retention in longitudinal research. Child: Care, Health and Development, 38 (5), 753 – 756. https://doi.org/10.1111/j.1365-2214.2011.01326.x&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Nicholson, J. M., Cann, W., Matthews, J., Berthelsen, D., Ukoumunne, O. C., Trajanovska, M., ... Hackworth, N. J. (2016). Enhancing the early home learning environment through a brief group parenting intervention: Study protocol for a cluster randomised controlled trial. BMC Pediatrics, 16 (1), 73. https://doi.org/10.1186/s12887-016-0610-1&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Nwadiuko, J., Isbell, P., Zolotor, A. J., Hussey, J., &amp;amp; Kotch, J. B. (2011). Using social networking sites in subject tracing. Field Methods, 23 (1), 77 – 85. https://doi.org/10.1177/1525822x10384088&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Rhodes, B., &amp;amp; Marks, E. (2011). Using Faceboook to locate sample members. Survey Practice, 4 (5), 1–6. https://doi.org/10.29115/SP-2011-0023&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Schneider, S., Burke-Garcia, A., &amp;amp; Thomas, G. (2015). Facebook as a tool for respondent tracing. Survey Practice, 8 (1), 1 – 6. https://doi.org/10.29115/SP-2015-0003&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Shead, S. (2016). Facebook is hiding messages from you in a secret inbox – Here&#39;s how to find it. Business Insider Australia. https://&lt;ulink href=&quot;http://www.businessinsider.com.au/how-to-find-hidden-message-requests-on-facebook-messenger-2016-4?r=UK&amp;amp;IR=T#/#1-open-the-facebook-messenger-app-1&quot;&gt;www.businessinsider.com.au/how-to-find-hidden-message-requests-on-facebook-messenger-2016-4?r=UK&amp;amp;IR=T#/#1-open-the-facebook-messenger-app-1&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Social Media News. (2019). Social media statistics Australia – January 2019. https://&lt;ulink href=&quot;http://www.socialmedianews.com.au/social-media-statistics-australia-january-2019/&quot;&gt;www.socialmedianews.com.au/social-media-statistics-australia-january-2019/&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; StataCorp. (2015). Stata statistical software: Release 14.&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Sweney, M. (2018). Is Facebook for old people? Over-55s flock in as the young leave The Guardian https://&lt;ulink href=&quot;http://www.theguardian.com/technology/2018/feb/12/is-facebook-for-old-people-over-55s-flock-in-as-the-young-leave&quot;&gt;www.theguardian.com/technology/2018/feb/12/is-facebook-for-old-people-over-55s-flock-in-as-the-young-leave&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Thomas, J., Wilson, C. K., &amp;amp; Park, S. (2018). Australia&#39;s digital divide is not going away. The Conversation https://theconversation.com/australias-digital-divide-is-not-going-away-91834&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Westrupp, E. M., Bennett, C., Cullinane, M., Hackworth, N. J., Berthelsen, D., Reilly, S., ... Nicholson, J. M. (2018). EHLS at School: School-age follow-up of the Early Home Learning Study cluster randomized controlled trial. BMC Pediatrics, 18 (1), 148. https://doi.org/10.1186/s12887-018-1122-y&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Young, A. F., Powers, J. R., &amp;amp; Bell, S. L. (2006). Attrition in longitudinal studies: Who do you lose? Australian and New Zealand Journal of Public Health, 30 (4), 353 – 361. https://doi.org/10.1111/j.1467-842X.2006.tb00849.x&lt;/bibtext&gt; &lt;/blist&gt; &lt;blist&gt; &lt;bibtext&gt; Zubrick, S. R., Lucas, N., Westrupp, E. M., &amp;amp; Nicholson, J. M. (2014). Parenting measures in the Longitudinal Study of Australian Children: Construct validity and measurement quality, Waves 1 to 4. Retrieved from Canberra : &lt;ulink href=&quot;http://hdl.handle.net/10620/4041&quot;&gt;http://hdl.handle.net/10620/4041&lt;/ulink&gt;&lt;/bibtext&gt; &lt;/blist&gt; &lt;/ref&gt; &lt;aug&gt; &lt;p&gt;By Shannon K. Bennetts; Jasmine Love; Naomi J. Hackworth; Fiona K. Mensah; Elizabeth M. Westrupp; Donna Berthelsen; Penny Levickis; Clair Bennett and Jan M. Nicholson&lt;/p&gt; &lt;p&gt;Reported by Author; Author; Author; Author; Author; Author; Author; Author; Author&lt;/p&gt; &lt;p&gt;&lt;/p&gt; &lt;p&gt;Dr Shannon K. Bennetts is a Research Fellow at the Judith Lumley Centre, La Trobe University, and Coordinator of the EHLS at School Study.&lt;/p&gt; &lt;p&gt;Ms Jasmine Love is a Research Officer at the Judith Lumley Centre, La Trobe University.&lt;/p&gt; &lt;p&gt;Dr Naomi J. Hackworth is a Senior Project Manager at the Parenting Research Centre.&lt;/p&gt; &lt;p&gt;Dr Fiona K. Mensah is a Senior Research Fellow in Epidemiology and Biostatistics within the Intergenerational Health Group at the Murdoch Children&#39;s Research Institute.&lt;/p&gt; &lt;p&gt;Dr Elizabeth M. Westrupp is a Senior Lecturer and Clinical Psychologist at Deakin University.&lt;/p&gt; &lt;p&gt;Professor Donna Berthelsen is an Adjunct Professor within the Faculty of Education, School of Early Childhood and Inclusive Education at Queensland University of Technology.&lt;/p&gt; &lt;p&gt;Dr Penny Levickis is a Postdoctoral Research Fellow in the Melbourne Graduate School of Education at The University of Melbourne.&lt;/p&gt; &lt;p&gt;Dr Clair Bennett is a Research Scientist at Columbia University in the United States, and formerly a Research Assistant at the Judith Lumley Centre, La Trobe University.&lt;/p&gt; &lt;p&gt;Professor Jan M. Nicholson is Director of the Judith Lumley Centre at La Trobe University and Chief Investigator of the EHLS at School Study.&lt;/p&gt; &lt;/aug&gt; &lt;nolink nlid=&quot;nl1&quot; bibid=&quot;bib33&quot; firstref=&quot;ref1&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl2&quot; bibid=&quot;bib20&quot; firstref=&quot;ref2&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl3&quot; bibid=&quot;bib11&quot; firstref=&quot;ref5&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl4&quot; bibid=&quot;bib17&quot; firstref=&quot;ref7&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl5&quot; bibid=&quot;bib13&quot; firstref=&quot;ref8&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl6&quot; bibid=&quot;bib31&quot; firstref=&quot;ref9&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl7&quot; bibid=&quot;bib28&quot; firstref=&quot;ref12&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl8&quot; bibid=&quot;bib22&quot; firstref=&quot;ref15&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl9&quot; bibid=&quot;bib19&quot; firstref=&quot;ref20&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl10&quot; bibid=&quot;bib25&quot; firstref=&quot;ref21&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl11&quot; bibid=&quot;bib12&quot; firstref=&quot;ref23&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl12&quot; bibid=&quot;bib26&quot; firstref=&quot;ref26&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl13&quot; bibid=&quot;bib21&quot; firstref=&quot;ref29&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl14&quot; bibid=&quot;bib16&quot; firstref=&quot;ref32&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl15&quot; bibid=&quot;bib24&quot; firstref=&quot;ref33&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl16&quot; bibid=&quot;bib14&quot; firstref=&quot;ref35&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl17&quot; bibid=&quot;bib15&quot; firstref=&quot;ref39&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl18&quot; bibid=&quot;bib10&quot; firstref=&quot;ref42&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl19&quot; bibid=&quot;bib23&quot; firstref=&quot;ref43&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl20&quot; bibid=&quot;bib32&quot; firstref=&quot;ref44&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl21&quot; bibid=&quot;bib29&quot; firstref=&quot;ref46&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl22&quot; bibid=&quot;bib18&quot; firstref=&quot;ref47&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl23&quot; bibid=&quot;bib34&quot; firstref=&quot;ref48&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl24&quot; bibid=&quot;bib27&quot; firstref=&quot;ref49&quot;&gt;&lt;/nolink&gt; &lt;nolink nlid=&quot;nl25&quot; bibid=&quot;bib30&quot; firstref=&quot;ref54&quot;&gt;&lt;/nolink&gt;
Header DbId: eric
DbLabel: ERIC
An: EJ1289842
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Selective Attrition in Longitudinal Studies: Effective Processes for Facebook Tracing
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: &lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Bennetts%2C+Shannon+K%2E%22&quot;&gt;Bennetts, Shannon K.&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Love%2C+Jasmine%22&quot;&gt;Love, Jasmine&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Hackworth%2C+Naomi+J%2E%22&quot;&gt;Hackworth, Naomi J.&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Mensah%2C+Fiona+K%2E%22&quot;&gt;Mensah, Fiona K.&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Westrupp%2C+Elizabeth+M%2E%22&quot;&gt;Westrupp, Elizabeth M.&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Berthelsen%2C+Donna%22&quot;&gt;Berthelsen, Donna&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Levickis%2C+Penny%22&quot;&gt;Levickis, Penny&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Bennett%2C+Clair%22&quot;&gt;Bennett, Clair&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Nicholson%2C+Jan+M%2E%22&quot;&gt;Nicholson, Jan M.&lt;/searchLink&gt;
– Name: TitleSource
  Label: Source
  Group: Src
  Data: &lt;searchLink fieldCode=&quot;SO&quot; term=&quot;%22International+Journal+of+Social+Research+Methodology%22&quot;&gt;&lt;i&gt;International Journal of Social Research Methodology&lt;/i&gt;&lt;/searchLink&gt;. 2021 24(2):135-147.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Routledge. Available from: Taylor &amp; Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 13
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2021
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles&lt;br /&gt;Reports - Research
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: &lt;searchLink fieldCode=&quot;EL&quot; term=&quot;%22Adult+Education%22&quot;&gt;Adult Education&lt;/searchLink&gt;
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: &lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Attrition+%28Research+Studies%29%22&quot;&gt;Attrition (Research Studies)&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Longitudinal+Studies%22&quot;&gt;Longitudinal Studies&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Social+Media%22&quot;&gt;Social Media&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Family+%28Sociological+Unit%29%22&quot;&gt;Family (Sociological Unit)&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Randomized+Controlled+Trials%22&quot;&gt;Randomized Controlled Trials&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Child+Rearing%22&quot;&gt;Child Rearing&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Parent+Education%22&quot;&gt;Parent Education&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Foreign+Countries%22&quot;&gt;Foreign Countries&lt;/searchLink&gt;
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: &lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Australia%22&quot;&gt;Australia&lt;/searchLink&gt;
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1080/13645579.2020.1765104
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1364-5579
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Participant attrition is a significant challenge for longitudinal studies, particularly those involving disadvantaged populations. We attempted to re-engage n = 990 families in a five-year follow-up of a randomised controlled trial of an early childhood parenting intervention. Tracing was attempted for n = 90/990 (9.1%) uncontactable participants. Evidence for the identification of a matching Facebook profile was classified as strong, moderate or weak. A private message was sent to those with strong evidence (n = 63/90, 70.0%), of which 32 responded, 22 agreed to participate and 19 provided data. Compared to participants engaged using traditional methods, those re-engaged through Facebook were more likely at initial recruitment to have a younger child (p =0.02), to be a single parent (p =0.04), less educated (p &lt;0.001), lower income (p =0.01) and in an unemployed household (p &lt;0.001). Findings suggest that social media can be an effective engagement tool, helping to minimise overall and selective attrition in longitudinal studies, particularly for hard-to-reach populations.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2021
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1289842
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1289842
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1080/13645579.2020.1765104
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 13
        StartPage: 135
    Subjects:
      – SubjectFull: Attrition (Research Studies)
        Type: general
      – SubjectFull: Longitudinal Studies
        Type: general
      – SubjectFull: Social Media
        Type: general
      – SubjectFull: Family (Sociological Unit)
        Type: general
      – SubjectFull: Randomized Controlled Trials
        Type: general
      – SubjectFull: Child Rearing
        Type: general
      – SubjectFull: Parent Education
        Type: general
      – SubjectFull: Foreign Countries
        Type: general
      – SubjectFull: Australia
        Type: general
    Titles:
      – TitleFull: Selective Attrition in Longitudinal Studies: Effective Processes for Facebook Tracing
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Bennetts, Shannon K.
      – PersonEntity:
          Name:
            NameFull: Love, Jasmine
      – PersonEntity:
          Name:
            NameFull: Hackworth, Naomi J.
      – PersonEntity:
          Name:
            NameFull: Mensah, Fiona K.
      – PersonEntity:
          Name:
            NameFull: Westrupp, Elizabeth M.
      – PersonEntity:
          Name:
            NameFull: Berthelsen, Donna
      – PersonEntity:
          Name:
            NameFull: Levickis, Penny
      – PersonEntity:
          Name:
            NameFull: Bennett, Clair
      – PersonEntity:
          Name:
            NameFull: Nicholson, Jan M.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2021
          Identifiers:
            – Type: issn-print
              Value: 1364-5579
          Numbering:
            – Type: volume
              Value: 24
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: International Journal of Social Research Methodology
              Type: main
ResultId 1