Coleridge Initiative

Democratizing our Data:
A Challenge to Invest in Data and Evidence-based Policy

For questions, email: [email protected]

EDUCATION TO WORKFORCE TRANSITIONS

Research Question:

Looking into Teacher Brain Drain among Kentucky Graduates based on county and state of origin, geographic region of university, educational field of study and employment outcomes.

Data Used:

Kentucky Department of Education, KY CPE, wage data from KY, Ohio, Indiana, Illinois, Missouri, Tennessee

Preliminary Findings:

Retention of students varies by origin WIB. During 2014-2017, most graduates in Elementary Education had stayed in Kentucky – Northern Kentucky had the highest rate of those who did leave.

Research Question:

What are the employment and earnings outcomes for health care and STEM graduates from Associate’s and Bachelor’s degree programs at Kentucky postsecondary institutions, including those who work in Kentucky, Ohio, Indiana, Illinois, Missouri, and Tennessee?

Data Used:

KYStats, post-secondary and unemployment insurance data for Ohio, Indiana, Illinois, Missouri, Tennessee

Preliminary Findings:

Findings from literature are validated in terms of STEM graduates being mor likely to leave Kentucky than Health/other majors. Demographic differences between the groups helps explain differences in migration. Out-of-state and younger graduates are more likely to leave (also account for higher proportions of STEM graduates)

Research Question:

Are 2013 graduates with STEM degrees leaving Kentucky for employment (by origin and degree level)? To what extent are employers in advanced industries hiring KY STEM grads (by state and super-sector)?  

Data Used:

Kentucky Postsecondary Education Data System, OH wage records

Preliminary Findings:

Wages are higher for individuals when working outside of Kentucky (STEM wages > non-STEM wages) – advanced industries provide a wage premium regardless of level (and field) of degree. STEM graduates however are more likely to find employment in an advanced industry in nearby states than in Kentucky

Research Question:

Is it good advice to urge a trade student to work towards higher credentials?

Data Used:

KCTS (2014-2017), wage data from KY, Ohio, Indiana, Illinois, Missouri, Tennessee

Preliminary Findings:

Associate Degree earners have higher post-graduation wages. Mobility wise, we note how approximately 90% of students stay in state but those who do end up leaving have higher credentials. Industry wise, Manufacturing has the higher post-graduation wage.

Research Question:

What are the earnings distribution and employment characteristics for individuals who graduate from Kentucky universities in Bachelor degree programs, particularly for those who have high-paying, stable employment patterns? Are there key trends in the wages and destination of students who originated from in-state and out-of-state?

Data Used:

KPEDS, wage data from KY, Ohio, Indiana, Illinois, Missouri, Tennessee

Preliminary Findings:

High-income earners tend to have consistent and stable high incomes each year after graduation. Those in the highest earning clusters have the lowest rates of re-enrolment in the first 3 years after receiving their Bachelor’s degree. Kentucky sends more Bachelor’s degree graduates out of state than it brings in (more pronounced for high earners)

Research Question:

Which programs of study at Ohio community colleges result in a better labor market outcome?

Data Used:

Ohio Technical Centers, Unemployment Insurance Wage records

Preliminary Findings:

Graduates of non-traditional college age earn 29% more than those of traditional age. Health Professions graduates are most financially successful when working in the Public Administration industry. Park, Recreation, Leisure and Fitness Studies program graduates have the lowest earnings.

Research Question:

Assessing the value of in-demand majors in addition to looking at which institutions were doing the best of providing in-demand graduates

Data Used:

Ohio Technical Centers, Unemployment Insurance Wage records

Preliminary Findings:

In-demand majors included Health-related professions (in particular, nursing). Even though Miami University produced the highest proportion of in-demand graduates in the southwest region and third highest for the state, it produced a lower proportion of in-demand graduates.

Research Question:

Do students who complete welding courses have better employment outcomes than those who enroll but do not complete these courses?

Data Used:

Ohio Technical Centers, Unemployment Insurance Wage records.

Preliminary Findings:

Findings indicate a clear association of non-completion with lower wages and higher rates in patterns representing little to no employment. Analysis does not support hypothesis of employers taking students before completion due to unmet high demand for welders.

Research Question:

In terms of demographic characteristics, who takes on debt while in graduate school? How is federal funding related to debt and the demographic characteristics of students at UMETRICS institutions? Are there differences in career outcomes three to five years after receipt of a doctoral degree for those who assumed debt during graduate school?

Data Used:

SED (2012-2013), UMETRICS (2012-2015), IPEDS (2012-2015), HERD (2013), SDR (2015 and 2017)

Preliminary Findings:

Doctoral recipients in Social, Behavioral, Economics (SBE) fields are less likely to acquire debt than those not in SBE fields. Findings suggest increased support to doctoral students belonging to racial/ethnic groups with highest debt risk.

Research Question:

Are federal funds disproportionately allocated within certain groups such as under-represented minorities, women, and people with disabilities?

Data Used:

SED, UMETRICS

Preliminary Findings:

People of color as well as females are less likely to be federally funded. Funding rates are also inconsistent for these rates – those who do not receive federal funding are less likely to stay in academia. 

Research Question:

What is the relationship between the sources of funding a doctoral student draws on and the duration (time to completion) of doctoral studies?

Data Used:

SED, UMETRICS

Preliminary Findings:

Debt level has an impact on PhD time completion – PhD students with high debt levels take longer to complete PhD (regardless of funding). Cases for standardizing graduate level fellowship and introducing caps for student debt are made.

Research Question:

What is the funding profile of Black students that earn life science PhDs?

What differences are there, if any, between Black students that earn their bachelor’s degrees at HBCUs vs. Non-HBCUs?

Data Used:

SED, UMETRICS, SDR

Preliminary Findings:

The study finds that HBCU graduates surpass non-HBCU graduates in terms of salary, federal government support and post-doctoral position. Furthermore, students who earned a baccalaureate from an HBCU take the same number of years to complete their doctoral degree when compared to black students who earned their bachelors from non-HBCUs. Focusing more on funding, it is to be noted HBCU students accumulate more graduate debt than students attending non-HBCU. 

Research Question:

What factors predict that an individual will fail to complete community college with a credential on-time?

Data Used:

Indiana Commission on Higher Education, Indiana Department of Workforce Development, Illinois Department of Employment Services, Ohio Department of Job and Family Services, Ohio Department of Higher Education

Preliminary Findings:

Demographics are predictive of graduating with a credential within five years of initial matriculation. Out-of-state attendees were also less likely to graduate within 5 years. In contrast, individuals who were employed in their first quarter of enrollment were more likely to obtain a credential within five years.

Research Question:

What are the characteristics of bachelor’s degree earners at Indiana public college who gain employment within one year of graduation in Indiana versus Ohio, Illinois, and Missouri?

Data Used:

Indiana Commission on Higher Education, Indiana Department of Workforce Development, Illinois Department of Employment Services, Ohio Department of Job and Family Services, Missouri Division of Employment Security

Preliminary Findings:

In general, demographics and residency were predictive of in-state employment. Field of study also played a role, with graduates earning credentials in engineering, business, computer science, communications/journalism, and transportation/materials moving more likely to be employed out of state the year after graduation.

Research Question:

How do the characteristics of one’s first, stable job after graduation differ during and post-Great Recession?

Data Used:

Indiana Commission for Higher Education, Indiana Department of Workforce Development

Preliminary Findings:

Wage growth differed noticeably by credential in the recovery period, with bachelor degree recipients receiving significantly higher wages compared to the recession while recipients of both associates degrees and certificates seeing no changes in earnings. In general, the recession negatively impacted the likelihood of finding stable employment post-graduation.

Research Question:

What are the characteristics of Missouri’s Bachelor Degree graduates that seek employment outside of the state?

Data Used:

Missouri Department of Higher Education, Missouri Department of Employment Security, American Community Survey, QWI Explorer, Bureau of Labor Statistics, National Science Foundation, National Center for Education Statistics

Preliminary Findings:

Demographics, residency, and STEM degree receipt predictive of post-graduation out-of-state employment.

Research Question:

What are risk factors for Missouri’s public university graduates to be gainfully employed in-state within 1,3, and 5 years?

Data Used:

Missouri Department of Higher Education, Missouri Department of Employment Security, United States Department of Agriculture

Preliminary Findings:

Demographics, and gender specifically, were predictive of in-state employment post-graduation. Graduates majoring in STEM fields and communications or journalism were more likely to be employed out-of-state.

Research Question:

Which community college enrollees are not likely to graduate within four years?

Data Used:

Missouri Department of Higher Education

Preliminary Findings:

Demographics, including age and race, are predictive of community college degree attainment.

Research Question:

How do vocational rehabilitation (VR) program completers compare to participants who do not complete programming? Is VR completion associated with higher earnings? Do earnings vary with the level of VR training received?

Data Used:

Ohio Technical Centers, Ohio Department of Higher Education, Ohio Department of Job and Family Services, Vocational Rehabilitation data courtesy of Opportunities for Ohioans with Disabilities

Preliminary Findings:

Many VR participants do not receive post-secondary credentials associated with the VR training. Demographics–specifically gender and race–were predictive of educational attainment and future employment. Individuals receiving higher level VR training received higher post-training earnings.

Research Question:

What are common enrollment patterns for community college students? How do these patterns affect their future employment outcomes?

Data Used:

Ohio Department of Higher Education, Ohio Department of Job and Family Services

Preliminary Findings:

Demographics, especially race, were related to timely degree completion; however, those taking longer to complete their degree received higher median wages upon graduation.

Research Question:

To what extent are Ohio’s college graduates in high-wage and high-demand occupations migrating to Illinois and Missouri?

What are the earnings and employment outcomes of people who have Ohio universities’ bachelor’s degrees in high wage/high demand occupations?

Data Used:

Ohio Department of Higher Education, Ohio Department of Job and Family Services, Illinois Department of Employment Security, Missouri Division of Employment Security

Preliminary Findings:

Over 60 percent of high-wage and high-demand degree recipients were employed in Ohio up to 3 years after graduation. Variation was observed in both earnings and likelihood of in-state employment across degree type.

Research Question:

What are the differences in post-graduation earnings for those awarded similar credentials from different post-secondary pathways?

Data Used:

Ohio Department of Higher Education, Ohio Department of Job and Family Services

Preliminary Findings:

Earnings increased on average for graduates of training programs; however, apprenticeship program graduates experienced the largest earnings gains.

Research Question:

Is there a difference in wages between apprenticeship completers and dropouts? Is the wage trend different for apprentices with different characteristics, such as different ages, races, education levels, industries, and union status? What factors can predict apprenticeship dropouts?

Data Used:

Ohio Department of Job and Family Services

Preliminary Findings:

Individuals who completed their apprenticeship have higher earnings and earnings growth than non-completers. Results varied, however, by age, education level, race, union status, and occupation industry.

Research Question:

How many sources of financial support are used by doctorate recipients from UMETRICS universities? Does the pattern differ for those who are “IRIS employees”? What are the primary and secondary sources of financial support used by doctorate recipients? How many doctorate recipients use loans as a source of support during graduate school? Does this differ based on IRIS employment? How does using a loan during graduate school relate to resulting debt? How do graduate school loan or debt levels relate to field of study and/or post-graduation plans?

Data Used:

National Center for Science and Engineering, Institute for Research on Innovation and Science

Preliminary Findings:

Doctorate recipients used multiple forms of financial support during graduate school. Debt rates and post-graduate employment commitments varied strongly across doctoral students and fields of study.

Research Question:

Does receiving federal funding affect the employment status immediately after

graduation?

Data Used:

National Center for Science and Engineering, Institute for Research on Innovation and Science

Preliminary Findings:

The majority of PhD recipients were federally funded as graduate students; however, receipt of federal funding was not correlated with likelihood of employment, salary, or employment sector upon graduation.

Research Question:

Can the NCSES’s Survey of Earned Doctorates be used to build a model to predict the likelihood that a PhD recipient will stay in the same region of the country for their post-doctoral work? Using a machine learning model, can one predict the region of residence for doctoral degree holders?

Data Used:

National Center for Science and Engineering, Institute for Research on Innovation and Science

Preliminary Findings:

Individual demographics were not strongly predictive of the likelihood of staying in the same region for post-doctoral work. A subset of states appear to be STEM exporters, with recipients moving to different states upon graduation.

Research Question:

Does the likelihood of working outside one’s degree field vary by field of study? In what sector are doctoral recipients in STEM working, and are they working in fields close to their own? – Does the likelihood of working outside one’s degree field vary by demographic characteristics? Does the likelihood of working outside one’s degree field vary by educational debt levels? Does the likelihood of working one’s degree field vary by institutional characteristics? Why are doctoral recipients working outside their field of study two years after earning their doctoral degree? Are doctoral recipients working outside their field less satisfied with their jobs? Are we able to predict whether a STEM doctorate is working in a field that is not closely related to his/her field of study two years after PhD graduation using demographic and institutional characteristics? What are the key predictors?

Data Used:

National Center for Science and Engineering

Preliminary Findings:

Around 30% of individuals earning doctoral degrees in STEM fields report working outside their field of study, with additional variation across fields of study observed. Individual demographics, especially gender and age at dissertation receipt, are predictive of working outside one’s field of study.

Research Question:

What is the relationship between personal and institutional characteristics and the choice of doctorate recipients going into academia upon receiving their PhD? Do minority institutions use different words in their National Institutes of Health (NIH) grant awards compared to very high research institutions?

Data Used:

National Center for Science and Engineering, Institute for Research on Innovation and Science, National Institutes of Health

Preliminary Findings:

Gender and race was correlated with NIH award

Research Question:

Is institutional research and development funding and patenting related to Science and Engineering doctorate-holders’ choice of academic or non-academic employment 2 years after graduation?

Data Used:

National Center for Science and Engineering Statistics, United States Patenting and Trademark Office

Preliminary Findings:

The majority of Science and Engineering doctoral degree recipients were employed in academia after graduation. This result did not vary with institutional research and development funding level; but did appear to vary with institution patent rates.

Research Question:

What are educational and labor market outcomes of students entering New Jersey as first-time freshmen? How do these outcomes change over the 10-year period following the students’ initial enrollment in community colleges? How do these outcomes differ by college completion status? What are the natural groupings of the sample based on the long-term labor market outcomes?

Data Used:

New Jersey Office of the Secretary of Higher Education, New Jersey Department of Labor & Workforce Development

Preliminary Findings:

Community college completer wages initially lagged behind non-completers, but eventually caught up and surpassed wages of non-completers. Individuals who went on to pursue a bachelor’s degree did not earn more than associates degree recipients 10 years after initial matriculation.

Research Question:

Do the earnings of 2-year and 4-year STEM degree earners differ?

Data Used:

New Jersey Office of the Secretary of Higher Education, New Jersey Department of Labor & Workforce Development

Preliminary Findings:

Individuals earning 4-year degrees in STEM exhibited larger cumulative earnings compared to their counterparts who earned 2-year degrees.

Research Question:

How do associate degree holders compare to bachelor degree holders financially over a 10-year period?

Data Used:

New Jersey Office of the Secretary of Higher Education, New Jersey Department of Labor & Workforce Development

Preliminary Findings:

Bachelor’s degree recipients initially made less than associates degree recipients; however, their earnings outweigh associate degree recipients in the long run.

Research Question:

Do policies promoting on-time completion benefit New Jersey students and employers?

Data Used:

New Jersey Office of the Secretary of Higher Education, New Jersey Department of Labor & Workforce Development

Preliminary Findings:

In-state residents were more likely to take longer than 4 years to graduate with a bachelor’s degree and the majority of individuals taking longer to graduate were employed while in college.

Research Question:

What are the employment outcomes and earnings for graduates of

short-term occupational degrees, certificate and associate programs, in

Manufacturing in New Jersey?

Data Used:

New Jersey Office of the Secretary of Higher Education, New Jersey Department of Labor & Workforce Development

Preliminary Findings:

The majority of graduates of manufacturing programs earned sub-baccalaureate degrees. Employment rates did not differ substantially between individuals earning certificates in manufacturing and graduates of baccalaureate programs.

UNEMPLOYMENT TO REEMPLOYMENT PATHWAYS

Research Question:

Can we identify trends showing how the lockdown impacted attrition of UI claims within different industries?

Can we identify education-related trends within industries that could help workforce boards target resources?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Education attainment is not a clear corollary of return-to-work rates in every industry. Upon analysis of all seven industries in our cohort within both Illinois’ Central and North Central regions, two outlying industries were identified. Administrative and Support Services Industry in the North Central Region, individuals with “some college” were the slowest to return to work.

In the Manufacturing Industry in Illinois’ Central Region, individuals with “some college” were also the slowest to return to work.

Research Question:

How many Manufacturing UI Claimants were there during the peak week of May 9th? Does it differ between rural and urban areas of Illinois?

For those Manufacturing UI Claimants that are a part of the COVID Cohort (entered UI late March/Early April), how long did it take them to re-enter the workforce?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

About one in five who lost jobs in the region were in Manufacturing. As compared with the state of Illinois, one out of 10 jobs lost were in Manufacturing. Manufacturing exit rates were higher than Accommodation and Food Services for all three areas. With the average percent exited being 42.3 percent at week 9, all manufacturing exit rates were a lot higher than the average. Manufacturing drops significantly in week 27, which is where all areas of both industries seem to converge.

Research Question:

How can we best inform workforce boards for timely and effective allocation of limited resources in a dynamic environment?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Developed exhaustion rates and identification of those most at risk of exhaustion across time to better deploy resources at the proper time. Developed a dashboard to show these rates over time by various subgroups.

Research Question:

Given the high level of claims in the Accommodation and Food Service

industry and after accounting for occupation, are there differences in exit

rates between men and women?

How can these findings help inform response by local workforce boards

to unemployment faced by men and women in this particularly hard hit

industry?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Workforce boards may want to address return-to-work assistance tactics and

policies differently depending on the individual’s industry, occupation, and

gender. Within the Accommodation and Food Service industry, individuals in certain occupations, like Food Preparation and Serving, may be able to return to work quickly through traditional job matching activities. However, individuals in occupations like Management may require additional supportive services beyond just referral to open jobs.

Research Question:

How big are the difference in return to work for essential and nonessential Illinois industries during COVID-19? Do they differ across industry, geography, and education?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

When comparing duration of UI benefits within the same industry category and education levels, workforce boards are more similar than different. Workforce regions seem to blend together when looking at high school diplomas or lower whether we look at essential or nonessential industries.

Any bit of education seems to reduce the long term stay rates. Bachelor degrees typically had a faster return to work in the same parts of the state. Associate’s degrees or some college also had somewhat faster returns.

Industries that were able to stay open because they were essential had faster return to work but not for the lower education levels that were more susceptible to layoffs.

Research Question:

Who Experienced Multiple Spells of Unemployment During the Pandemic? A Study of Additional Claimants

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Most claimants who returned to benefits had filed for 26 weeks, followed by an additional claim. This number doubles between weeks 26 and 37, adding context to the importance of tracking claimants for a longer period. Additionally, this has implications for state’s Unemployment Trust Funds, in the sense of accounting for claimants who have the potential to return to benefits. We found that workers experiencing stuttering employment in the early weeks following the initial cohort were impacted similarly across various demographics, including age group, education, and gender. This was an important finding because while claimants differed in terms of who was initially impacted, those who had returned to UI benefits early on were impacted at similar rates.

Research Question:

Which groups within the healthcare industry were most affected, in terms of length of unemployment and wage loss, by the pandemic-induced recession?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Those who identified as Black and African American had slower exit rates than those who identified as white or another racial identity across both cohorts. However, this difference was most pronounced in the May cohort, especially for Blacks and African Americans, where the survival rate hovered consistently at approximately 10 percent or more than the other two racial groups. Those who identified as white were quicker to exit by week 8 in the May cohort, than they were in the August cohort: approximately 50 percent of whites in the August were still receiving benefits at week 8, compared to compared to 35-40 percent of whites in the May cohort.

Research Question:

Did lifting the stay-at-home orders (May 29th /June 26th) have a different effect on the likelihood of continuing to qualify for UI for males vs. females? Do gender differences persist across age-groups?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

More women stayed on unemployment longer than men in the 25-44 age group. This was true of all age groups with the exception of the 25 and under age group. Women in the under 25 age group were hit harder initially, and men stayed on unemployment longer than women. Comparing the weekly remaining share by gender between the 25-44 age group and the over 55 group as a control shows greater difficulties regaining employment for women in the over 55 group. It is possible that this shows that childcare responsibilities are not the reason why women have remained on unemployment longer. It is more likely that the choice of control group was problematic.  It was assumed that the over 55 group would be a good control group because of assumed lesser childcare responsibilities for females in the group. The childcare responsibilities might also be prevalent (for grandchildren etc.) or it could be differences in occupations or choices to retire.

Research Question:

What were the characteristics of claimants who stayed on benefits after the Executive Orders (EO) was lifted? How can Workforce Boards use this information to determine groups that would benefit the most from intervention or additional services?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Within healthcare, and before the EO, less than HS claimants were the quickest to return to employment. After the EO, they were the slowest. The disparities in education widened as the year progressed.

Research Question:

How does the generosity of unemployment insurance benefits affect claimant behavior? Does industry and other demographics have an impact? Does context matter?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Claimants with high wage replacement rates have longer durations than those with and medium and low wage replacement rates. More than three-fourths of the cohort’s wage replacement rate was greater than 100% and almost a third had wage replacement rates greater than 200%

Finance, Insurance, & Real Estate: number of initial claimants was relatively low, and they continued to claim benefits through Week 26. Accommodations & Food Services: number of initial claimants was high and remains high due to slow reopening. Healthcare & Social Services along with Manufacturing, Retail Trade and Construction: number of claimants was initially high then reached a relatively low rate. Suggests a demand side effect for labor rather than solely a supply side decision by workers based on wage replacement.

Research Question:

Considering the pandemic disproportionately affected the number of new, certified female claimants, is the stay rate higher among females?

If so, are differences in exit patterns driven by supply-side factors (wage replacement, childcare needs) or demand-side factors (lockdowns and unavailability of jobs)?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Women overrepresented in claims, and slower to exit benefits. When wage replacement rate was less than 1, women exited faster than men. When wage replacement rate was greater than 1, women exited slower than men. Could be that women are more represented in customer-facing jobs where demand was lower, and risk was higher. Could also be that women are more economically responsive than men.

 

Research Question:

How did this recession different than the last? More specifically, are different industries impacted? Have the demographics of the long-term unemployed changed?  Are there trends by race that continue to show in long-term unemployed persons?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Black individuals have a far lower share of the higher wage jobs and are also not returning to work as quickly. The sorting is based off the 2019 wages from the PROMIS file and the OES numbers for statewide 25th, 50th, and 75th percentile wages. People-Moving was harder hit than Product-Moving, especially in Northeast; recall this region surrounds the major travel hub of Chicago Blacks and Hispanics hit the hardest, proportionally, based on exit rates.

Research Question:

Is healthcare still recession proof? Were age and occupation within healthcare important factors in worker’s ability to re-enter the workforce?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Stay rates between the May and October healthcare cohorts followed similar patterns suggesting that healthcare is not acting recession proof. There were some differences across the two cohorts as younger age groups exited faster in the second cohort.

Research Question:

Why did post-peak weekly healthcare claims level off at higher rates than before the pandemic?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

The demand for healthcare jobs is lower than before the pandemic. This level of demand may not increase until the pandemic has passed if it does at all.

Workforce boards may want to consider short term solutions to assist these occupations in the healthcare industry. The solution could be additional training for occupations with higher demand within or outside healthcare.

Research Question:

What was the effect on the UI benefit wage replacement rate?

Was an additional $600 a disincentive for returning to work?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Stay rates of workers in the Accommodation and Food Services industry did

not vary by replacement rates. The story is not as clear for Retail Trade, Health Care and Social Assistance, and Manufacturing: Claimants with high replacement rates tended to stay longer in UI during FPUC. This could be because weaker demand of low-wage (high replacement rate) workers or because FPUC lowered workers willingness to return to work.

Research Question:

Across three representative industries:  Food and Accommodation, Retail Trade, and Manufacturing:

Did COVID result in a disproportionately high number of certified claimants relative to the pre-pandemic averages for young and wise workers?

Were young and wise workers more likely to exhaust their UI benefits?

Data Used:

Illinois Department of Employment Security, UI and QCEW

Preliminary Findings:

Young workers were most likely to lose their employment after the COVID shutdown. With respect to age cohorts, the middle-aged grouping (aged 25-54) accounted for most unemployed workers, reflecting their share of the workforce. The younger and older cohorts were roughly the same size. Each cohort had similar exit rates through the first 12 weeks following program entry, but they diverged after that. Younger workers were most likely to exit the program beginning about week 13 and were the least likely to exhaust benefits. Older workers were the least likely to exit; by week 26, the middle and older cohorts accounted for similar shares of workers exhausting benefits and the younger worker share was lower.

Research Question:

What features of an individual’s employment and earnings history predict the likelihood of being unemployed for two or more quarters in the next two years?

Data Used:

Illinois Department of Employment Services

Preliminary Findings:

Previous quarterly earnings were the strongest predictors of future unemployment.

WORKFORCE TRANSITIONS FOR FORMERLY INCARCERATED INDIVIDUALS

Research Question:

How do treatment and employment affect recidivism rates for offenders with substance abuse history?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security, Chicago Police Department

Preliminary Findings:

Prior substance abuse is predictive of future recidivism. This relationship is attenuated by participation in a rehab program within IDOC.

Research Question:

What are the relevant factors for technical violation that lead to recidivism? Seeks to develop a model that can identify those most likely to recidivate for a technical violation and what those violations entail.

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

Technical violations associated with parole or mandatory supervised release account for as much as 32% of yearly admissions into IDOC. Demographics were important predictors of recidivism, but policy related factors such as prerelease employment plans, substance abuse were also highly predictive.

Research Question:

Can targeted early release programs improve the employment outcomes and reduce the recidivism of formerly incarcerated people.

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

 

 

Demographics play a key part in employment outcomes, but offenders that pursue education and/or training have much higher rates on employment post incarceration. The findings are robust to different labor markets.

Research Question:

Which prisoners should receive priority for interventions that can increase or enhance employment opportunities after release?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

Ex-offenders who were never employed upon release had served longer sentences, were older, were an active gang member, and reported using alcohol and some drugs. Having employment prior to prison and receiving education in prison were related to being employed upon release. 

Research Question:

How can we use administrative data to demonstrate the impacts of geographic access to employment opportunities and social services on the recently incarcerated?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

Most large employers who are willing to hire people with criminal records are not located in the neighborhoods where people live. Non-recidivists were more likely to find employment closer to home in the first four quarters of release. Jobs that had the highest retention rates were further away but were closer for non-recidivists. Non-recidivists also experienced higher wage growth.

Research Question:

How can we provide better information to decision makers? Can we develop a tool that looks across four text heavy information domains: research, policies, programs, and media?

Data Used:

Webscraping, API

Preliminary Findings:

Developed an initial machine-generated feature set that classifies population and non-population words from research articles.

Research Question:

Can the NYPD predict future shooting victims using previous victim arrest and demographic data

Data Used:

NYC shooting incidents, NYPD arrest incidents

Preliminary Findings:

Used a machine learning model to predict shooting victims. Achieved modest precision but will look to connect with other data to improve the features included. Can be key to planning interventions with possible victims.

Research Question:

What are the predictors of women with children re-entering the criminal justice system after exit?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security, aggregated neighborhood characteristics

Preliminary Findings:

Neighborhood factors such as health services and community support systems such as job training and consultation are significantly related to reduced recidivism rates among women.

Research Question:

What is the effect of employment among at-risk youth on their future employment outcomes? What can we learn that will help improve job programs for at-risk youth?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

Early work habits seem to significantly influence adult work outcomes, though the summer job turns out to be almost entirely insignificant.

Research Question:

How well can we currently measure post-SNAP self-sufficiency? How else could we measure, and predict self-sufficiency?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

The current measures being used are insufficient at predicting self-sufficiency after participating in SNAP. Began to develop a better model that included additional features that performed better.

Research Question:

What is the impact of increasing minimum wage on total earnings for all individuals and for those in the fast-food industry?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

Found that the short-term, immediate impact of the increase in minimum wage on earnings was positive and significant. And, while the total dollar value increase was less for those working in the fast-food industry and those receiving public assistance, the magnitude of impact on those subgroups was most likely bigger given the fact that median earnings for those groups are significantly lower than median earnings for the total population. These results

Research Question:

What is impact on wages and Supplemental Nutrition Assistance Program (SNAP) usage of layoff event for manufacturing workers in rural economy?

What is the spillover impact of a mass layoff on those working w/in 5 miles of mass layoff event?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

Found significant earnings losses and no recovery years later for those experiencing a layoff event. There was little response in SNAP participation.

Research Question:

Does Occupational Industry Impact Subsequent Earnings for Cash Assistance Recipients?

Data Used:

New York City Human Resources Administration, NYC Department of Corrections, and the New York State Department of Labor

Preliminary Findings:

Entering the job market in an industry with higher earnings, higher job creation rates, and lower turnover tended to bode well for earnings several years later, though mean earnings even in the most beneficial cluster were below estimated earnings for full-time minimum wage employment in 2016. Conversely, entering the job market in an industry with low earnings, low job creation rates, and low turnover rates did not bode well for future earnings.

Research Question:

Among a cohort of young adults who have received cash assistance, can we predict the likelihood of returning to cash assistance after ending a spell?

Data Used:

New York City Human Resources Administration, NYC Department of Corrections, and the New York State Department of Labor

Preliminary Findings:

Uses ML to build a classification model and identifies important features of welfare dependency in young adults

Research Question:

What are the factors that impact economic stability among young New Yorkers receiving cash assistance (CA)?

Data Used:

New York City Human Resources Administration, NYC Department of Corrections, and the New York State Department of Labor

Preliminary Findings:

Participation in WEP leading up to 2012 positively associated with 3+ quarters employment. Participation in WeCARE during those years negatively associated with employment. These results likely reflect fact that WEP serves clients assessed as ‘able to work w/o limitations’, while WeCARE designed to help CA clients overcome health & mental health barriers to employment.

Research Question:

Do specific characteristics predict an increased likelihood that inmates in the Illinois corrections system will gain successful employment post-release?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security, Missouri Division of Employment Security

Preliminary Findings:

Through clustering analysis, best chances for success are amongst individuals who are: White, Asian, Hispanic; males; with no kids; in their 30s; who have an HS diploma or better

Research Question:

What proportion of ex-offenders obtain steady wages/stable employment post-release? What factors influence whether Illinois ex-offenders find work after release in either Illinois or Missouri?

Data Used:

Illinois Department of Corrections, prison exit table 2010 – 2013; Illinois Department of Employment Security (IDES), wages earned 2005-2015; Missouri Division of Employment Security (MoDOR), wages earned 2005-2015

Preliminary Findings:

Intervention with all individuals who were unemployed before imprisonment increases their chances of having steady wages post-prison. However, this is at a very large scale so it is advised to start off by focusing on 5% of the incarcerated population (near a border county who are expected to find jobs in Missouri).

Research Question:

How does an offender’s wages following release from prison influence their risk of recidivism?

Data Used:

Missouri Department of Corrections, Missouri Department of Employment Security.

Preliminary Findings:

Employment following release was strongly negatively related to recidivism. In addition, release type is predictive of recidivism; with those discharged significantly less likely to return to prison compared to those released on probation, parole, or through conditional release.

Research Question:

Does the Missouri Department of Corrections’ vocational training program have an effect on ex-offenders finding stable employment one year after release?

Data Used:

Missouri Department of Correction, Missouri Department of Employment Security, Missouri Department of Higher Education

Preliminary Findings:

Vocational training in prison positively associated with employment stability after release.

Research Question:

Is there a relationship between high school dropout rate and age at the first jail entry? Does this relationship persist even after controlling for school and neighborhood characteristics?  Can one predict if an individual will be admitted to jail before (s)he reaches 21 years old?

Data Used:

Illinois Department of Correction, Chicago Public Schools, United States Census Bureau

Preliminary Findings:

Individual demographics predict jail entry, even when controlling for neighborhood and school characteristics.

Research Question:

How do state-level criminal and economic factors influence the likelihood of recidivism?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

Individual demographics, residency, and education, as well as state-level employment changes, predict recidivism,

WORKFORCE TRANSITIONS FOR SOCIAL BENEFIT RECIPIENTS

Research Question:

Determine whether it is possible to use these data to identify ABAWDs

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Services

 

Preliminary Findings:

Predictive model is effective in identifying ABAWDs. In order to better track ABAWD-related outcomes, it is recommended to include a direct, time-varying measure of their status.

Research Question:

Exploring features that determine if a TANF recipient returns to TANF within two years – essentially looking at predictors that allow us to determine TANF recipients who have difficulty in transitioning from welfare to work/self-sufficiency.

Data Used:

Illinois Department of Human Services (IDHS), Illinois Department of Employment Security Unemployment Insurance

Preliminary Findings:

Through utilizing multiple Machine Learning models, approximately 60% of Earning-Age adults who return to TANF within two years are correctly identified. Features important in predicting this include gender, recipient age, education level and average earnings before spell.

Research Question:

Can we predict which TANF recipients in Illinois are likely to return to TANF within 2 quarters after exiting the program?

Data Used:

Illinois Department of Human Services (IDHS), Illinois Department of Employment Services (IDES)

Preliminary Findings:

Multiple Machine Learning models are utilized to predict return to TANF.

Research Question:

What are the earnings and employment outcomes of TANF recipients in the year after they exit the program?

Data Used:

Indiana Department of Workforce Development, Family and Social Services Administration, TANF program data (2009-2019)

Preliminary Findings:

Analysis suggests that conditions for Indiana TANF exiter cohorts do not change as they continue to be employed in low-wage jobs in some aspect of the service industry. From a policy standpoint, the study urges self-sufficient wages for TANF recipients who leave the program.

Research Question:

For one parent cases that exit in 2016 Q1-Q4, is the head of household

ever employed in the first four quarters after exit? Stable employment, re-entry rates (after 1 year) are also looked into. Secondly, do these outcomes vary by cohort or employer characteristics?

 

Data Used:

Indiana TANF data, Indiana UI wage data

Preliminary Findings:

Higher earnings before entering TANF indicate stable employment post-TANF. Pre-TANF employment in Healthcare and Social Assistance is linked to stable employment (post-TANF) as well. Developing services is recommended for groups needing additional support such as individuals with lower/zero earnings for a certain time before entering TANF or no high school diploma.

Research Question:

Compare recipients who left TANF and either did or did not return within a specific time period to motivate the idea of developing post-TANF self-sufficiency measures

Data Used:

Indiana TANF data, Indiana UI wage data

Preliminary Findings:

Wages, stable employment, education and employment with a smaller company were all variables that differed when looking at people who left TANF and returned versus those who left TANF and did not return in the 6-month window. Food services industry is dominant in terms of cohort earnings, but it is not suggested to place individuals here primarily because of the on-going pandemic.

Research Question:

Looking at the interaction between TANF participation and Labor Market outcomes while focusing on demographic and employment trends.

Data Used:

Indiana TANF data, Indiana UI wage data

Preliminary Findings:

TANF participants’ use of assistance and employment is complex. Most participants exist but often to sporadic or no employment. In this case however, TANF exit is unrelated to employment as most participants exit but many not to wages.

Research Question:

Looking at the impact of the Great Recession (during and post) and the COVID-19 pandemic on TANF program recipients’ length of stay, churn and recidivism.

Data Used:

TANF program data, Indiana Department of Workforce Development

Preliminary Findings:

TANF length of stay, churn, and recidivism were each much higher in Indiana during the Great Recession than four-years afterward, while short- and long-term post-TANF employment rates were much lower. To the extent that the COVID-19 recession has similar effects on the New Jersey caseload as the Great Recession caseload did on the Indiana caseload, the families we serve will evidence greater needs.

Research Question:

Study characteristics and outcomes of the TANF cohort in Marion County, Indiana, to be able to predict successful path to self-sufficiency

Data Used:

Indiana TANF data, Indiana UI wage data, LEHD Data for Marion County, Indiana

Preliminary Findings:

Average earnings of TANF returners are much lower than that of non-returners in every follow-up quarter. Industries more likely to hire TANF leavers are: administrative and support services, food services and drinking places, merchandize stores, nursing and residential care facilities.

Research Question:

What are the Characteristics Associated with Churn in TANF Recipients?

Data Used:

Indiana TANF data, Indiana UI wage data

Preliminary Findings:

Administrative and Support services have a 50% rate of return for those employed in that industry. Partnering with employers to improve pathways to industries and occupations is suggested in addition to looking at barriers that factor into people returning.

Research Question:

What are the earnings and employment outcomes of TANF recipients? What are the characteristics of employers likely to hire TANF recipients? Comparing Recessions and Expansions

Data Used:

Indiana Department of Workforce Development (Wage Records, 2005-2018), Indiana Family and Social Services Administration (TANF), Indiana Department of Workforce Development (QCEW)

Preliminary Findings:

Findings suggest that a lower proportion of the 2009 (Q1, Recession) cohort was employed for at least one quarter in the year after exiting compared to 2016 (Q4, Expansion). Employment rates are lower in all 4 quarters in 2009 than in 2016. Employers associated with hiring TANF recipients are mostly characterized as small businesses in both years.

Research Question:

What factors predict full-quarter, minimum wage employment for TANF leavers?

Data Used:

Department of Workforce Development, Family and Social Services Administration

Preliminary Findings:

Previous full-time employment positively predicted employment after TANF. Results varied by individual education level and type of employment.

Research Question:

What factors create successful outcomes for TANF leavers in Illinois?

Data Used:

Illinois Department of Employment Services, Illinois Department of Human Services

Preliminary Findings:

The majority of former TANF recipients remained unemployed in the quarter after TANF exit. Results suggested a “benefits cliff”, whereby individuals initially earning higher wages upon TANF exit were less likely to experience wage gains in later years.

Research Question:

What factors predict whether an adult will achieve stable employment in the year after exiting TANF?

Data Used:

Department of Workforce Development, Family and Social Services Administration in Indiana

Preliminary Findings:

The most indicative predictor of stable employment was that the individual was employed during their quarter of TANF entry.

Research Question:

Which demographic or policy factors increase the risk of returning to TANF?

Data Used:

Department of Workforce Development, Family and Social Services Administration in Indiana

Preliminary Findings:

Participation in child support programs and employment in food service, clothing, museum, construction, crops, education service, electronics, or laundry industries are associated with a return to TANF.

Research Question:

What characteristics increase an individual’s risk of returning to TANF in Illinois within one year?

Data Used:

Illinois Department of Employment Services, Illinois Department of Human Services

Preliminary Findings:

Unemployment, single adulthood, and having a young child all increase the risk of returning to TANF.

Research Question:

What are the characteristics of recipients leaving TANF who are at-risk of not finding stable employment?

Data Used:

Illinois Department of Employment Services, Illinois Department of Human Services

Preliminary Findings:

Education level and work in the healthcare/social assistance industry were positively correlated with finding stable employment after leaving TANF.

Research Question:

What is the change in the relationship between CCDF program participation and gainful employment as a result of extending the eligibility and redetermination periods in the CCDF program from 6 months to 12 months?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

Extending the CCDF’s recertification period from 6 to 12 months increased the likelihood of a family receiving the program’s subsidy while also slightly increased the probability of employment and earnings.

Research Question:

Are WIC participants more likely to purchase whole grain products?

Data Used:

IRI InfoScan and IRI Consumer Network data

Preliminary Findings:

WIC participating households are more likely to have purchased a whole grain product in a year’s time. Results vary by region and household structure.

Research Question:

What is the total amount spent and total ounces purchased on WIC-eligible juice in 2017 by method of payment? Do results vary between WIC households and WIC-eligible, but not enrolled (EBNE) households? Do results vary over time? What household factors predict purchase of non-WIC juice items?

Data Used:

IRI InfoScan and IRI Consumer Network data

Preliminary Findings:

WIC households generally purchased WIC-eligible juices with their own money as opposed to through the WIC program. Although WIC households spent more money on WIC-eligible juices, they purchased less by volume than EBNE households. WIC households generally spend more money on all types of juices compared to EBNE household, and primarily purchased sugar-sweetened juice items.

Research Question:

How does the annual proportion of 100% whole wheat bread out of total bread purchases differ between WIC and WIC-eligible households by household size and income? What household characteristics are the best predictors of the differences between WIC and WIC-eligible households?

Data Used:

IRI InfoScan and IRI Consumer Network data

Preliminary Findings:

WIC households were more likely to purchase 100% whole wheat bread, independent of household size. Demographics, including gender, race, and family income, were predictive of not participating in WIC despite eligibility.

Research Question:

What are the prices of nutritionally-eligible breakfast cereals that are and are not WIC approved in three states? Can we create a tool that will help states identify stores that overcharge the State for breakfast cereals?

Data Used:

IRI InfoScan data,

Preliminary Findings:

WIC-approved cereals had a higher average price per ounce than nutritionally-eligible cereals. Prices of breakfast cereals also seem to vary by store type.

Research Question:

Do stores across Indiana offer equal access to 100% whole-wheat bread?

Data Used:

IRI InfoScan data

Preliminary Findings:

Convenience stores charged significantly more for WIC-approved bread. That being said, convenience stores were more likely to carry WIC-approved bread.

Research Question:

What individual characteristics describe Illinois single mother TANF recipients? What characteristics predict the earnings of Illinois single mother TANF recipients? Which Illinois single mother TANF recipients are most likely to garner higher later earnings?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Services

Preliminary Findings:

Findings suggest that efforts designed to promote individuals’ capacity to work and credential attainment may be most productive for reducing dependency on government assistance.

Research Question:

How does the rate of recidivism change for people with drug felonies after the TANF ban? Is there any evidence of administrative errors, fraud, or some other anomaly in the TANF administrative records

Data Used:

Illinois Department of Corrections, Illinois Department of Human Services

Preliminary Findings:

Recidivism rates significantly increased following a ban of TANF benefits for individuals with drug felonies.

Research Question:

What factors predict the continued need of TANF and SNAP benefits?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Services

Preliminary Findings:

Individual demographics, number of spells in benefit programs, and time since last participation in a benefit program significantly predict future participation.

Research Question:

What factors predict interaction with the criminal justice system among participants in King County’s Mental Illness and Drug Dependency (MIDD) program?

Data Used:

King County (Washington) Department of Adult and Juvenile Detention, King County’s Mental Illness and Drug Dependency program

Preliminary Findings:

Individual demographics, age, and prior interaction with the criminal justice system are predictive of future interactions.

Research Question:

Among household assistance benefit recipients residing in a specific metropolitan area, are there demographic, economic, and/or location characteristics that are predictive of transitions from household assistance lasting a minimum of 2 years? Additionally, are there distinguishing characteristics that could be identified which would allow us to differentiate between successful and unsuccessful transitions from assistance?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security

Preliminary Findings:

An association between higher wage and transitioning off of TANF is established using Machine Learning tools. Another important feature is days since last spell in TANF – more days would imply that the recipient had not received benefits in a longer time. This points towards the possibility that they are employed.

Research Question:

How can we better predict which TANF recipients will be successful so that service providers can better understand the factors associated with success?

Data Used:

Illinois Department of Human Services, Illinois Department of Employment Security, Illinois Department of Corrections

Preliminary Findings:

The study finds that the most important features for predicting success in the labor market were previous wages and work history. Education, age and health serve as important predictors too. Residents with health limitations were less likely to find success – there is an emphasis on providing preventative healthcare for welfare recipients.

Research Question:

Studying the access to medical assistance for vulnerable populations in New York City

Data Used:

NYC DOHMH Office of Emergency Preparedness and Response (OEPR), 2014 American Community Survey (ACS)

Preliminary Findings:

Approximately 75% of all NYC residents travel less than one mile to their closest point of dispensing (POD) and 96% are within 2 miles. Although a linear model did support the presence of disparities, there is potential for investigation using a Geographically Weighted Regression.

Research Question:

Do formerly incarcerated individuals release from Illinois Department of Corrections (DOC) take-up SNAP within two years of release?

Data Used:

Illinois Department of Corrections, Illinois Department of Human Services

Preliminary Findings:

SNAP receipt appears more likely for people with repeat Illinois incarcerations and longer lengths of stay. Large differences in SNAP receipt are observed for people who were discharged and recommitted.

Research Question:

Predict successful two-year prison exits and what successful exits look like when facilities are included.

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security

Preliminary Findings:

Important features in predicting successful exits were days in prison, age at first admit, number of kids, prior exits. Facility district or location did not add much to the predictive models. Inclusion of additional wage and employment data is recommended to understand how economic conditions might impact exit success.

Research Question:

What factors might influence released prisoners to enroll in SNAP upon re-entry into community?

Data Used:

Illinois Department of Corrections, Illinois Department of Employment Security, Illinois Department of Human Services

Preliminary Findings:

Prior SNAP enrolment plays an important role in determining future enrolment. People on SNAP before prison are thus more likely to return to it post-prison. However, a small percentage of people exiting prison get connected to food assistance – presence of an information gap about enrolling into public assistance benefits is noted.

MEASUREMENT

Research Question:

What wage-related factors affect firm survivability during an economic crisis?

Data Used:

Missouri State LEHD Wage and Employer Data 

Years of study: 2007, 2010, 2013

Preliminary Findings:

Average number of employees has no impact on short-term survivability but has a negative impact on long-term survivability of firms. Postal services seem to do well in the short-term.

Research Question:

What impact did the Great Recession have on the wages of Missouri’s computer and electronic product manufacturing industry?

Data Used:

Missouri LEHD wage and employer dataset, public NAICS_2007 and NAICS_2012 lookup tables, Bureau of Labor Statistics

Preliminary Findings:

Divided employees into 3 cohorts:

1.     Stayed in industry (exhibited the lowest median wage out of all the cohorts – greatest percent increase in total median wage from 2006 to 2014)

2.     Left industry and returned (Highest median wage in addition to having the highest percentage of middle wage employees)

3.     Left industry and never returned (Largest difference between median wage for all sectors and median wage for NAICS 334 in addition to having the smallest percentage of middle wage employees)

Research Question:

How does the churn rate for Kansas City, Missouri (KCMO) firms differ across time, across industries, and compared to national trends? Can we differentiate between firms with different kinds of churn?

Data Used:

Missouri State LEHD quarterly records of employer and wage data

Preliminary Findings:

The number of jobs churned per quarter is substantial and has grown as the economy has improved. Industries such as Social Assistance, Technical services exhibit lower rates of churn but more churned jobs.

Research Question:

What characteristics of firms are indicators of firm survivability in low-wage industries in Kansas City, MO?

Data Used:

Missouri Department of Labor Wage Records, Missouri Department of Labor Employer Data

Preliminary Findings:

Median wage of firms in low-wage industries that survive is higher than similar firms that closed – wage serves as an important feature for firm survivability.

Research Question:

Relationship between access to public transit and jobs in the Kansas City metro area

Data Used:

General Transit Feed System (GTFS), LODES Files (ADRF)

Preliminary Findings:

Approximately 4000 unique Census block pairs with available transit duration under 30 minutes. Mean trip duration is 22 minutes and mean trip distance is 3 miles.

Research Question:

How did the underemployed in the state of Missouri fare from 2006 through 2015 in terms of wages?

Data Used:

Missouri Labor Force  

Preliminary Findings:

A majority of the low-earners stayed in the low-earner group until 2015. A small percentage of 2006 low-earners progressed to the high earner group.

Research Question:

What is the trajectory of Missouri’s middle class and might utilization of administrative data comprehensively assess its health, or, lack thereof?

Data Used:

Missouri Labor Force (Wage data)

Preliminary Findings:

Extra small firms (less than 10 middle class workers) represent the largest portion of middle-class employers whereas large firms represent the smallest portion of middle-class employers.

Research Question:

What are the geographic distribution, employment outcomes, and characteristics associated with employment success of WIA Youth customers?

Data Used:

 

Chicago Cook Workforce partnership, Illinois Department of Employment Security, American Community Survey (2011-2016), NAICS

Preliminary Findings:

The most common post-exit outcome is continuous employment. Youth saw an 18% increase in average wage over 2 years post-exit. Lastly, most WIA clients work in the service industry.

Research Question:

How do policy makers identify Illinois firms that are likely not to survive the next two years in order to make an effective intervention?

Data Used:

Illinois State LEHD Employers

Preliminary Findings:

13% of firms did not survive from 2011 to 2013. Firms created less than 5 years ago were more likely to fail.

Research Question:

Can we predict which businesses will experience high turnover of low-wage employees in the next year?

Data Used:

Missouri Longitudinal Employer-Household Dynamics (LEHD) Wage and Employer Data 2006-2016

Preliminary Findings:

Number of employees is highly important in predicting employee turnover. Total wage is the least important. In terms of industries, Educational Services, Construction and Public Administration are the most important when looking at turnover.

Research Question:

How well do self-reported wages in the ACS overlap with the firm-reported wages of the EHF? Can we pair EHF/ACS analysis with demographic/geographic information to establish statewide trends?

Data Used:

United States Census Bureau, Illinois Department of Employment Services

Preliminary Findings:

Combining administrative records with Census data has potential to improve Census data products while providing states data to improve their evaluation of programs.