Publication Type: Symposium Paper
Abstract: NICHD protocol-guided interviews of alleged victims were compared to matched non-protocol interviews with respect to the identification of investigative leads, defined as reports of information which suggest new directions for investigation. In Protocol-guided interviews, more leads were elicited using recall prompts, which are more likely to elicit accurate information. The leads provided in Protocol-guided interviews were themselves more likely to involve central aspects of the alleged incidents, and they pointed investigators towards details that were more likely to cast light on the accuracy of the children’s accounts As a result, it was easier to independently corroborate or cast doubt on the allegations made in Protocol interviews.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: In 1992, the Canadian Corrections and Conditional Release Act established the Correctional Investigator’s role as an independent Ombudsman for federal prisoners. In providing an accessible avenue for redress for offender complaints as well as for offender inquiries concerning the legitimacy of their treatment in prison, the Correctional Investigator is said to be a public watchdog, and to uphold federal prison law by ensuring that correctional policies and procedures associated with individual complaints are in keeping with the law. Thus, Canadian legislation arms the Correctional Investigator with operational tools as well as significant authority to carry out investigations, and to bring most prisoner complaints to a timely closure. A study of 123 complaint files related to segregation (solitary confinement), and received by the Correctional Investigator between 1995 and 2003 indicates that, contrary to what would be expected in the case of such disguised and secretive forms of confinement, investigations tend to be cursory. In this light, the paper will discuss the mechanisms by which positive attempts to render conditions of segregation visible tend to be short-circuited, hereby contributing inadvertently to the ongoing invisibilization of segregation. The efficacy of the Correctional Investigator as a watchdog of legal rights, and as a governmental self-surveillance mechanism will also be addressed.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: A growing body of literature exists which examines factors influencing case clearances for violent crimes, especially homicide. The large majority of this research investigates the impact of victim, incident, and investigative characteristics on the likelihood of solving the case. An issue that has received much less attention is the role of the neighborhood context in not only the successful identification of suspects, but also the successful prosecution of suspects in court. This study takes advantage of detailed data on homicide cases for Cleveland, Ohio from 1998 through 2002 to examine these issues. Hierarchical multinomial logistic regression is used to investigate the impact of case- and neighborhood-level characteristics on homicide clearances and convictions. The paper concludes with a discussion of obstacles to arresting and prosecuting homicide offenders in urban environments.

Publication Type: Symposium Paper
Abstract: Because the NICHD protocol elicits more extensive free narrative accounts of allegedly experienced events, we hypothesized that raters would be able to judge the credibility of allegations made in protocol interviews more accurately than those made in non-protocol interviews. In this report we present two Israeli studies designed to explore the effects of the NICHD investigative interview protocol on the assessment of credibility in statements by alleged victims of abuse. The first study describes child investigators' ability to assess credibility while the second focuses on the discriminating power of CBCA assessments in protocol interviews and non-protocol interviews. Overall, the NICHD protocol enhanced credibility assessment.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: Nagler (1994) discusses the scobit (skewed-logit) model as a possible alternative to logit and probit. This model is intuitively appealing as it relaxes the assumption that the largest impact of a change in an independent variable occurs where the initial probability of success is equal to 0.5. This is an especially important consideration if one is interested, for example, in which types of people are most sensitive to changes in stimuli. Following Nagler (1994), I examine the scobit model using Monte Carlo simulations and data from the Bureau of the Census Current Population Survey: Voter Supplement File. Through Monte Carlo simulations I find that scobit may not be reliable in small samples; and in an analysis of voter turnout using Census data backed by Monte Carlo simulations, I find that scobit is especially sensitive to specification changes. Furthermore, these alternative specifications influence the substantive interpretations one derives from the analysis.