Absence of conflict of interest.
Citation
Kanengiser, H., & Schaberg, K. (2022). Employment and earnings effects of the WorkAdvance demonstration after seven years. New York: MDRC. [Tulsa Site – Madison Strategies]
Highlights
- The study's objective was to examine the long-term impact of WorkAdvance on employment and earnings for adults who were unemployed or earning low wages.
The study was a randomized controlled trial conducted across multiple sites and states that used administrative earnings data to compare outcomes for individuals who participated in WorkAdvance to a control group who could not enroll in WorkAdvance but could enroll in other community services. This discussion focuses on the results for the WorkAdvance model implemented by Madison Strategies Group in Tulsa, Oklahoma.
- Among
participants in the Tulsa study site, the study did not detect any statistically significant association between assignment to the WorkAdvance intervention and year-seven earnings or employment.
- The quality of causal evidence presented in this report is high because it was based on a well-implemented randomized controlled trial. This means we are confident that the estimated effects are attributable to WorkAdvance, and not to other factors. However, the study did not find statistically significant effects in its analysis of participants at the Tulsa site.
Intervention Examined
WorkAdvance
Features of the Intervention
The structure of the WorkAdvance training program was informed by prior research on sectoral strategies, which aim to train individuals for high-quality jobs in specific industries or occupational clusters with robust local demand and opportunities for career progression. It also incorporated insights from earlier studies on job retention and career advancement strategies. The WorkAdvance study aimed to expand on these existing findings by investigating whether sector programs emphasizing career advancement could serve as a pathway to upward mobility for individuals with low incomes.
The WorkAdvance program model has five main components: 1) intensive screening, 2) career-readiness services, 3) occupational skills training leading to a credential, 4) job development and placement, 5) retention and advancement services.
Features of the Study
The study was a randomized controlled trial that compared outcomes for individuals offered access to the WorkAdvance program to outcomes for those who could not access WorkAdvance but could receiver other services in the community. The WorkAdvance intervention was delivered by distinct local organizations at four study sites: Per Scholas in the Bronx, New York; St. Nicks Alliance in Brooklyn, New York; Madison Strategies Group in Tulsa, Oklahoma; and Towards Employment in northeast Ohio. The study examined the impact of WorkAdvance programs separately for each site, and in a pooled analysis that combined participants from all sites. This study profile summarizes findings from the Tulsa site. Findings from other sites and from the pooled analysis are reported in separate study profiles.
The study sample for the Tulsa site included 353 participants who were randomly assigned to participate in the WorkAdvance program delivered by Madison Strategies Group, and 344 participants who were randomly assigned to the control group. Madison Strategies Group is a nonprofit organization that provides workforce development services in Tulsa, Oklahoma, with an emphasis on the transportation and manufacturing sectors.
WorkAdvance targeted adults who were either unemployed or earning low wages (less than $15 per hour), with family incomes below 200 percent of the federal poverty level. At the Tulsa site, study participants were, on average, 35 years old and were predominantly male (84 percent). At baseline, 10 percent of these study participants had already obtained a postsecondary degree, 27 percent were employed, and 40 percent reported that they had a prior criminal conviction. Among this study sample, 6 percent identified as Hispanic/Latino, 28 percent identified as Black, non-Hispanic, 39 percent identified as white, non-Hispanic, and 21 percent identified as American Indian, non-Hispanic.
The primary data source used was administrative employer-reported quarterly earnings data collected by the National Directory of New Hires (NDNH). This data provides insights into the long-term economic outcomes of study participants, collected seven years after random assignment. This extended follow-up period aimed to assess whether the WorkAdvance programs led to sustained career advancements beyond initial job placements. The study had two pre-specified confirmatory outcomes: 1) total annual earnings in year seven following random assignment; and 2) the share of individuals with earnings of $40,000 or more in year seven following random assignment.
The authors utilized a statistical model to compare the long-term outcomes of individuals assigned to the treatment and control groups.
Findings
Earnings and Wages
- For the Tulsa site, the study did not detect any statistically significant association between assignment to the WorkAdvance intervention and earnings or wages seven years after random assignment.
Employment
- For the Tulsa site, the study did not detect any statistically significant association between assignment to the WorkAdvance intervention and an individual’s likelihood of being employed in year seven after random assignment.
Considerations for Interpreting the Findings
The impact of the WorkAdvance program on year-seven average earnings varied across study sites. Although the study did not detect statistically significant impacts in the Tulsa site, the study did detect favorable, statistically significant impacts on year-seven earnings at the study’s Bronx site, and for the pooled sample that included individuals from all four sites.
Causal Evidence Rating
The quality of causal evidence presented in this report is high because it was based on a well-implemented randomized controlled trial. This means we are confident that the estimated effects are attributable to WorkAdvance, and not to other factors. However, the study did not find statistically significant effects in its analysis of the Tulsa study sample.