A Latent Class Analysis of Perceived Employment Barriers Among Low-Income Job Seekers

Background/Purpose: Poverty is a major social problem that is difficult to address and revolve in the United States. Individuals living in poverty are likely to have multiple individual/structural barriers and some of these barriers may interact with each other. When these barriers are highly correlated with each other, it is difficult to distinguish the single and compounded effects on on low-income individuals, especially when they enter the job market (Danziger et al., 2000). Despite the complexity of employment-related barriers, research on poverty tends to focus on the effects of certain targeted risk factors on labor market outcomes. The purpose of this study is to better characterize the patterns of co-occurring employment barriers among welfare-to-work program participants.

Methods: Using a secondary data of 372 low-income individuals in a federally funded career pathways program, this study uses a latent class analysis (LCA) to identify unique patterns of multiple employment barriers. Further, it examines whether the identified barrier patterns have changed during program participation using a latent transition analysis (LTA). Perceived employment barriers scale (PEBS; Hong et al., 2014) was used to capture program participants’ employment barriers. PEBS (i.e., physical & mental health, labor market exclusion, child care, human capital and soft skills) is a Likert scale ranging from 1 to 5 on 27 items. For the LCA, each item was recoded into “not a barrier (1)” and “barrier (2-5)” to indicate presence or absence of employment barriers.

Results: Using LCA, meaningful subgroups have been found that share similar characteristics based on participant responses. At each time point, a 4-class LCA solution is determined to have the optimal balance of fit and parsimonious explanation of the data based on model fit indexes. Class 1 comprised 11% of program participants and was defined as “all high barriers”, Class 2 (54%) included participants with “high work-related barriers”, Class 3 (12%) included individuals with “high work-related and community-related barriers” and Class 4 (22%) included participants with “all low employment barriers”. One notable pattern from the LTA result was that there were no class transition to other sub-groups between the time points for participants in Classes 1, 3, and 4. Only Class 2 (“work-related barriers”) at Time 1 transitioned to Class 4 (“all low employment barriers”) at Time 2 (18%).

Conclusions and Implications: Findings of this study showed that each group of participants may have differing priority needs and/or issues when they enter the job market. In addition, LCA can offer a more nuanced understanding of workforce development program participants and enable social work practitioners to be more informed about participant needs when providing employment services. By examining possible distinct patterns of employment barriers, the results may help policymakers and practitioners to provide more specific and targeted interventions designed to meet the poverty population’s individual needs and circumstances associated with the different patterns of subgroups.