Abstract
Parental incarceration has been found to negatively impact families and communities. This study examined characteristics that impact receiving in-person visits among a sample of incarcerated mothers of minor children in a large Southern prison system. Factors impacting post release plans to live with children were also examined. Contact with their children, race/ethnicity, offense type, previous incarceration, and sentence length predicted whether incarcerated mothers received in-person visits from their children. Receiving visits from their children, race/ethnicity, having custody prior to arrest, and offense type predicted whether incarcerated mothers planned to live with their children after release from prison. Implications for policy and future research are presented.
Acknowledgment
The research contained in this document was coordinated in part by the Texas Department of Criminal Justice (Research Agreement #716-AR14). The contents of this report reflect the views of the author and do not necessarily reflect the views or policies of the Texas Department of Criminal Justice.
Notes
1 Within the current study there was variation by custody level between and within facilities. In order to ensure the confidentiality of women in smaller institutions, the administered survey did not collect information on the specific facility women were incarcerated within. Because of this, it is unclear how facility-level characteristics influenced visitation, however, the supervising correctional agency lists identical visitation policies and procedures for all facilities within the observed region.
2 Analyses were conducted using both dichotomous and categorical measures of visitation. While visitation was significant in both models, diagnostics suggested that the model using the categorical measure of visitation was of poorer fit as indicated by a reduction in pseudo R2 (0.265 vs. 0.245) as well an average increase of 0.20 across standard errors. Further, the categorical variant of visitation was potentially multicollinear with a variety of variables as indicated by the variable inflation factor for the categorical variant of visitation (VIF = 1.22).
3 Bivariate correlations between all variables were conducted to check for multicollinearity. Specific attention was paid to contact modalities (mail, telephone, and visitation contact). These forms of contact were not found to be significantly correlated with one another in the current sample. The strongest relationship was between visitation and telephone contact which were not significant and only moderately correlated with one another, although it is worth noting that this relationship might be interpreted as approaching significance (r = 0.40, p = 0.09).
4 The differences in the odds of experiencing the observed outcome levels were determined to be proportional. This was done by testing if the proportional odds assumption was violated. Post testing was conducted using the model command (Wolfe & Gould, 1997) which conducts a likelihood ratio test of whether coefficients are equal across outcome categories in order to measure compliance with the proportional odds assumption. Our findings indicate that the current model does not violate the proportional odds assumption (x2=85.34, p = 0.89).