Just a Normal Conversation: Investigative Interviews in a County Jail

Abstract

The majority of research on investigative interviewing has been on police attempting to solve a crime by obtaining a confession or gathering information, and comparatively fewer studies have examined interviewing at points “downstream” in the process, such as in the courts or correctional system. Furthermore, the focus of the research has been to measure the variable techniques or questioning strategies that produce confessions or information at the expense of analyzing factors related to the interview itself. Thus, we analyzed a sample of 50 corrections-based interviews for “dynamic” interviewing methods and interviewee responses that were measured at three points throughout the interview, and we measured 10 “static” interview factors. In the final multilevel model, we found that productive questioning methods increased a component score that combined interviewee cooperation, engagement, and forthcomingness, the several measures of accusatorial interrogation methods decreased the outcome measure, and the case-level variable of interviewee-initiated interviews increased it.

The interrogator‐suspect dynamic in custodial interrogations for high‐stakes crimes in Sweden: An application of the interrogation taxonomy framework

“…For example, Kelly et al . (2021) applied the taxonomy domains to a sample of investigative interviews of incarcerated men, yielding findings in line with past research: Kelly et al . (2016, 2021) both found consistent and strong negative effects on interviewee cooperation of the confrontation/competition and presentation of evidence domains.…”

“…A sizable number (73%) of our sequences were scored as having only a question or statement, not a domain, but we did not investigate this particular finding further. However, future research can and should continue to examine all elements of the interaction (see Hagsand, Zajac, et al ., 2022; Kelly et al ., 2021; Kelly & Valencia, 2021).…”

“…Such factors include whether an attorney was present and the strength of the evidence against the suspect, among other variables that are more appropriately considered properties of the interrogation itself, as opposed to the sequence or time block. There is no conclusive support, however, that these factors supersede the dynamic variables analyzed above, and only one study to date (Kelly et al ., 2021) has even attempted to simultaneously model both the dynamic and static, or contextual, factors. Indeed, this study found that in the final model, the contextual variables were largely rendered non‐significant while many of the interrogation methods measured at the time block level remained significant predictors of interviewee responses.…”

“…In the present study, we analyzed actual interrogations, allowing us to examine observed behavior by interrogators, instead of self‐reported behavior. While examinations of observed behavior have been recently addressed in studies focusing on high‐stakes crimes in US samples (e.g., Kelly et al ., 2016, 2021; Kelly et al ., 2016; Kelly, Redlich & Miller, 2015; Kelly & Valencia, 2021) and a UK sample (Izotovas, Kelly & Walsh, 2021), the extent to which these findings apply to other countries and investigative settings remains unclear (but see Hagsand, Zajac, et al ., 2022 for a recent study of Swedish interrogations in low‐stakes crimes related to alcohol and drugs). Therefore, the present study examined police interrogations 1 in high‐stakes crimes (i.e., murder and child sexual abuse) in a Swedish police sample.…”