Latent Class Analysis (LCA): Caregivers’ Stories, Perceptions, and Concerns

Abstract Information: Evaluation studies often have to assess the effect of interventions on specific groups. Evaluation questions that focus on identifying participant profiles or subgroups within a sample call for latent class analysis (LCA), a statistical technique that uses observed indicators to classify individuals with similar characteristics into subgroups. Using a person-centered approach, LCA utilizes people’s responses to obtain clusters of similar perceptions, concerns, and outcomes. It allows for examination of stories on how people feel and what they think about a particular issue. In this study, participants’ responses to items on depression, trauma, stress, and substance misuse scales were analyzed to identify subgroups in relation to aggression, and the use of sedatives and stimulants. This presentation will demonstrate the use of LCA in identifying profiles of child welfare involved caregivers based on the similarities of their stories emerged as a result of their responses to questions included in the structural assessments.

Relevance Statement: Developing targeted interventions for particular groups of individuals who would benefit from these interventions has become an essential part of evaluation studies. Although the average impact of a program or intervention is important, the examination of differential impacts on meaningful subgroups provides a better understanding of program effects (McGroder et al., 2012). Latent class analysis is an analytic technique developed to address research or evaluation questions for which identifying a cluster of characteristics or subgroups of people with similar characteristics is necessary. For example, LCA can be used to identify the profiles of children served in out-of-home care based on their responses regarding their school, living environment, attachment to peers or to assess whether child welfare involved caregivers can be classified into groups based on their responses to items included in the standardized measures developed to assess various aspects of a person’s behavioral health. LCA may be particularly valuable for evaluation studies because it provides an opportunity to obtain not only more accurate findings, but also new and often more interesting information that can be used in practice. Much of the current interest in evaluation studies can be directly linked to the eliciting additional information obtained from standardized measures. This presentation will demonstrate the use of LCA in identifying profiles of child welfare involved caregivers based on the similarities of their stories, which emerged as a result of their responses to questions included in structural assessments. The proposed presentation will also focus on the patterns of child welfare involved caregivers’ feelings, perceptions, and outcomes across several specific subsamples, and the similarities and differences among them. Using an example from the current study, this presentation will also illustrate the process of empirically deriving groups and examining their association with different outcomes. In addition, this presentation will provide examples of how LCA can be utilized in addressing various evaluation questions as well as practical guidance for preparing data for LCA, interpretation of various class-solutions, and comparing classes in relation to the examined outcomes. Thus, the proposed presentation will contribute to the evaluation field by demonstrating a sophisticated way of analyzing data, thereby creating a key resource for a diverse range of evaluators.

McGroder, S., P. Caronongan, A. Mraz Esposito, S. Asheer, and J. Kwon-Min. (2012). Inside
the Black Box of Interactions Between Programs and Participants: Re-conceptualizing
Subgroups for Fatherhood Program Evaluation. Washington, DC: Office of the Assistant
Secretary for Planning and Evaluation, U.S. Department of Health and Human Services.
Retrieved from
https://aspe.hhs.gov/report/inside-black-box-interactions-between-programs-and-participants-re-conceptualizing-subgroups-fatherhood-program-evaluation/subgroups-program-evaluation-research