Administra*ve data to examine undergraduate research experiences Jason Owen- Smith University of Michigan Ins*tute for Research on Innova*on and Science (IRIS) jdos@umich.edu
Roadmap Background Learning Analy*cs/Research Analy*cs (LARA) Ques*ons and Data Linkage Preliminary/Sugges*ve Findings Further Ques*ons IRIS: A plaporm for data sharing
Background Two parallel CIC ini*a*ves Learning Analy*cs (Tim McKay) Umetrics (Lane, Owen- Smith, Weinberg) Discussions in Ann Arbor about big data Sloan Founda*on funding for a pair of LARA workshops. Parallel Play Early Results from Michigan Data
Probing Impact of Research Experiences on Reten:on and Persistence at Student Record Data Transcript Research for credit Major Admission porpolio Demographic informa*on Federally Funded Research Records Payroll Grant account Job *tle Propor*on of earnings FY 2002-2014: 93,350 students FY 2002-2014: 49,487 personnel Data assembled: 12,006 students enrolled in research course Data consolidated: 18,720 students had research experience 2,416 students had both research experiences Data assembled: 9,130 students paid on grant 4
Who are the students that are paid on research grants? Category Grant Employed Not Grant Employed % Male 4,211 42,349 9.0 Female 4,919 41,871 10.5 White 5,618 54,038 9.4 Minority 3,162 22,654 12.2 Interna*onal Student 472 6,638 6.6 Pell grant 2,465 15,521 13.7 Total 9,130 84,220 9.8
Who Funds Them? CFDA # Jobs % Na*onal Ins*tutes of Health 3,940 38.2 Na:onal Science Founda:on 2,033 19.7 Other 1,642 15.9 Department of Educa*on 1,471 14.2 Department of Defense 456 4.4 Department of Energy 358 3.5 HHS 329 3.2 Department of Agriculture 91 0.88 The problem of work study.
What jobs do they do? Assistant in Research (Temp) (31.5%) Laboratory Assistant (Temp) (29.0%) Research Assistant I Social Sciences (Temp) (11.1%) Tutor I (Temp) (10.3%) Top four classifica*ons account for ~82% of all undergrad employees
Degree Program (graduates only) Graduates Grant Not Grant % Employed Employed Total 7,054 48,927 12.6 L S & A 4,447 29,512 13.1 Engineering 1,925 8,739 18.1 Nursing 191 1,465 11.5 Kinesiology 141 1,703 7.6 Business Admin 119 3,182 3.6 Education 86 784 9.9 Art and Design 40 842 4.5 Music, Thtre & Dance 34 1,220 2.7 Public Policy 26 259 9.1 Architecture 21 869 2.4 Dental Hygiene 15 310 4.6
PuYng classes and employment together 93,350 enrolled students 12,006 (12.9%) took a class 9,130 (9.8%) paid on a grant 2416 (2.6%) did both 87.2% (2105) of students who did both took the class FIRST
Educa:onal Outcomes Avg. GPA 3.26 Avg. GPA 3.36 Avg. GPA 3.34 Avg. GPA 3.35 N=14,225
Data Sharing Infrastructure Linkages Research Community Next Steps/Needs Stakeholder Input (Universi*es, Students, Employers, Funders) Use Inspired Fundamental Research about (Engineering) Educa*on & Workforce
Ins*tute for Research on Innova*on and Science (IRIS)
An Opportunity... STAR METRICS represents a valuable step toward developing detailed, broadly accessible and na*onally representa*ve data that would allow systema*c and scien*fic analysis of the organiza*on, produc*vity, and at least some of the effects of federally funded research [but]... P. 4-10 1.... STAR METRICS data are largely inaccessible... 2.... data collec*on could usefully be expanded to include more universi*es and other performers... 3.... STAR METRICS data would be more useful if steps were taken to ensure the data can be flexibly linked to other data sources [such as] those maintained by the federal sta*s*cal and science agencies... as well as proprietary data sources... Crea*ng a robust and linkable dataset may require the addi*on of individual and organiza*onal iden*fiers.
Universi*es Contribute Data and Receive Reports/Dashboard Researchers Access De- iden*fied, aggregate product OSU U Chicago Individual Campus Data Boxes Combined Iden:fied Data Combined De- Iden:fied Data VDE CENSUS RDC NPR membership Model 63 researchers have accessed data
Already sharing (at least some) research data Ø Cal Tech Ø Indiana Ø Iowa Ø Michigan Ø Minnesota Ø Northwestern Ø OSU Ø Penn State Ø Purdue Ø Wisconsin Parallel play with learning data Ø Indiana Ø Iowa Ø Maryland Ø Michigan State Ø Michigan Ø Wisconsin Census links for research data Ø Cal Tech Ø Iowa Ø Michigan Ø Minnesota Ø Purdue Ø Penn State Ø Wisconsin
Ques:ons we can soon answer about the engineering workforce? Rich experiences in and out of the classroom Research, co- curriculars, internships Matricula:on and Persistence Gender, URM students Further Educa:on and Careers Clearinghouse data, Census links Comparisons Engineers to non- engineers Interven:ons?
Ques:ons?