Misidentified or Misunderstood? A Reply to Wildeman, Wakefield, and Turney (2013)

Abstract

Our original review was intended to underscore both the methodological challenges of disentangling the effects of parental incarceration from other adversities that often co-occur with parental incarceration and the need for conceptual models that can explain how and why parental incarceration may have unique effects on child well-being (Johnson & Easterling, 2012). More generally, the review was designed to invigorate discussion and inquiry related to understanding the unique effects of parental incarceration on child well-being. The important and insightful commentary provided by Wildeman, Wakefield, and Turney (2013) suggests that the article served this purpose with regard to methodological issues and communicates the utility of embracing multiple methods for addressing the problem of selection bias, not only for individuals interested in understanding the effects of parental incarceration on wellbeing but also for any researcher dealing with endogeneity issues. At the same time, the commentary also underscores the need for greater attention to the conceptual issues we discussed in our original article. In this brief response, we elaborate on the scope of our review and consider the implications of some additional studies for the conclusions we made. We also reemphasize the need for conceptual models that can explain the unique effects of parental incarceration on child well-being and discuss directions for future research that we think are indicated by this exchange.The primary critique articulated by Wildeman et al. (2013) is that our review omitted several relevant studies and therefore mischaracterizes the literature on parental incarceration and child well-being. They conceded that several of the articles they identified may not have been published prior to the publication of our article but suggested that the remaining articles should have been included in our review. As we stated on pages 344 through 345 of the original article, our review was limited to articles that had been published in peer-reviewed journals and that utilized what we termed a comparative approach. This included studies that compared outcomes among children whose parents were incarcerated with groups of children who were similar on one or more theoretically relevant variables but did not experience parental incarceration (e.g., children who had been separated from parents because of divorce; children whose parents were incarcerated prior to their birth). Although three of the studies they identified (Craigie, 201 1; Foster & Hagan, 2007; Roettger, Swisher, Kuhl, & Chavez, 2010) are excellent examples of studies that rigorously controlled for multiple background characteristics and compared outcomes among subgroups of children whose parents were incarcerated (e.g., girls vs. boys, Black vs. White vs. Hispanic youth), we did not perceive them to be comparative studies in the sense that we used the phrase in our original article and therefore did not include them in our review. Although we originally did not perceive three of the remaining articles to meet our selection criteria, for a similar reason (Foster & Hagan, 2009; Wakefield & Wildeman, 201 1; Wildeman, 2010), we agree with Wildeman et al. that we likely defined matching too narrowly and that these studies, all of which used propensity score matching methods, do meet the spirit of our inclusion criteria and would have been appropriate to include.Wildeman et al. (2013) argued that the inclusion of these additional articles in our review would have led us to a different set of conclusions regarding both the methodological quality of research on parental incarceration and the substance of findings regarding parental incarceration and well-being. They perceived one of our conclusions to be that few studies on parental incarceration and well-being are rigorous, and they argued that the additional studies they identified are significantly more rigorous than the studies we included in our review and would therefore change our assessment of the rigor of the field (p. …