Text this: The effect of the timing and spacing of observations in longitudinal studies of tobacco and other drug use