A prospective, longitudinal study of the correlates and consequences of early grade retention (purchase price at link = $31.50) (Shane Jimerson, et al. "Journal of School Psychology", Volume 35, Issue 1, Spring 1997, Pages 3-25.)
The characteristics of children retained in early elementary school and the effects of retention on achievement and adjustment were examined throughout the elementary years and again at age 16 years. When compared to a group of nonretained children who displayed similar levels of early achievement and were comparable on two measures of intelligence, the retained subjects were more likely to be males with significantly poorer adjustment. Parents of comparison children were higher on IQ and were more involved with the school than parents of retained children. Controlling for initial levels of achievement and adjustment, little evidence was found supporting retention as an intervention for improving educational outcomes. The retained group showed a temporary advantage in math achievement, but this disappeared as both groups faced new material. Moreover, the retained group exhibited significantly lower emotional health in the sixth grade. It is concluded that elementary grade retention was an ineffective intervention for both achievement and adjustment.
From page 4:
Thus, while retained children constitute a group segregated from their peers reportedly on the basis of poor achievement or "immaturity," in fact their retention may reflect a subjective decision-making process based on a variety of factors.
The possibility of children being retained due to their behavior problems is very important, in light of the concern generated by the outcome literature examining the effects of retention on pupil adjustment. If the practice of retention is found to produce subsequent difficulties in adjustment, then students who were retained initially due to adjustment difficulties may face an exacerbation of their problems as a result of the intervention.
From the "Discussion", page 20:
Retained students were characterized as being significantly less confident, less self-assured, and less engaging than their academically similar peers. Retained students were also reported by teachers to be more unpopular and less socially competent than their peers.
... Furthermore, research does not support retention as an intervention for children with behavior problems (Holmes, 1989; Smith & Shepard, 1987). Our results indicate that, in fact, their behavior problems worsened. Following retention, the retained students displayed exacerbated behavior problems by sixth grade, while their low-achieving, but promoted peers remained stable.
And page 21:
There were significant differences between the groups regarding social and personal adjustment in the sixth grade. After controlling for initial adjustment levels, the retained group exhibited significantly lower rankings on the emotional health/self-esteem ranking in comparison to both the low-achieving promoted students and control group. The retained group's low-achieving but promoted peers displayed more gains in emotional adjustment.
An "ouch" from page 22:
In the current study, the retained subjects continued to display comparable achievement to their low-achieving peers at age 16. Ultimately, it appears that the intervention of retaining these students did not benefit them academically and was ineffective, if not harmful, in terms of their adjustment.
From the final paragraph:
Based on our findings, the practice of retention (nonpromotion of students) appears to be ill-advised. This research suggests that retention may appear to facilitate early academic performance that disappears over time and may prove potentially harmful regarding adjustment.