a regression analysis of student motivation and the effect of si on student success kathryn beck...

20
A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Upload: damon-king

Post on 03-Jan-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

A Regression Analysis of Student Motivation and the Effect of SI on

Student Success

Kathryn BeckGraduate Student, Applied

Economics

Page 2: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Supplemental Instruction, SI

• Academic Support Program– Historically difficult courses– High DFW rates

• SI Sessions for review and study– Example: Business Statistics 1

Page 3: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Main Questions

• Does attending SI affect student achievement in a course?– How does SI affect the DFW rate?– Should the program be eliminated or extended?

Page 4: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Theories

• Attending SI will improve a student’s final grade and increase understanding in the given course– Aid in future courses and increase graduation

rates• Attending SI may be worse for students who

are better off studying differently

Page 5: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Empirical Issues; Endogeneity

• Motivation– Correlation with attending SI (upward bias)

• At-risk students– Correlation with attending SI (downward bias)

• Unclear as to which direction the bias is causing

Page 6: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Literature Review

• Correlations Only• The International Center for Supplemental

Instruction• Empirical Models

• Blanc, Debuhr, and Martin (Journal of Higher Education, 1983)• Bowles & Jones (Digital Commons @USU, 2003)

Page 7: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

University of Wisconsin, Rock County

• Located in Janesville, WI (pop: 60,000)• One of 13 campuses of the UW Colleges• Enrollment (Fall 2013): 1,120• Average class size: 24• Student profile:

52% part-time33% non-traditional

Page 8: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

SI at UW Rock County

• Since 2011• 18 SI sections offered in conjunction with ten

different courses• 10 SI leaders• 30% participation rate in twice-weekly

sessionsMean Course GPA

SI participants: 2.46Non-SI participants: 1.93

Page 9: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Data

• University of Wisconsin Rock County– Student data from Spring 2011 until Fall 2013– 824 total observations

Page 10: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Model 1: Baseline Model

• Value-Added Education Production Function

• Xijt denotes a vector of student level characteristics for student i in class j in

t semester

• Zjt denotes a vector of course level characteristics for j class in t semester• (βα…) are estimated coefficients

𝐺𝑃𝐴𝑖𝑡 = 𝛽0 +𝛽1𝐴𝑡𝑡𝑒𝑛𝑑+𝛽2𝐺𝑃𝐴𝑖𝑡𝑗−1 +𝛼𝑿𝑖𝑗𝑡 +𝛿𝒁𝑗𝑡 +𝜀𝑖𝑗𝑡

Page 11: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Model 2: IV Estimation

• 1st Stage:

– Where Witj includes the instrumental variables

• 2nd Stage:

𝐿𝑜𝑔𝑆𝐼𝐴𝑡𝑡𝑒𝑛𝑑𝑒𝑑𝑖𝑡𝑗ෳ� = 𝛽0 +𝛽1𝑾𝑖𝑡𝑗 +𝛽2𝐺𝑃𝐴𝑖𝑡𝑗−1 +𝛼𝑿𝑖𝑡𝑗 +𝛿𝒁𝑗𝑡 +𝑣𝑖𝑡𝑗

𝐺𝑃𝐴𝑖𝑡 = 𝛽0 + 𝐿𝑜𝑔𝑆𝐼ෳ� 𝑖𝑡𝑗 +𝛽2𝐺𝑃𝐴𝑖𝑡𝑗−1 +𝛼𝑋𝑖𝑡𝑗 +𝛿𝑍𝑗𝑡 + 𝜇𝑖𝑗𝑡

Page 12: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Results

• Baseline Models:– With all imputations, attendance is significant– Without imputations, attendance not significant

• IV Estimations:– With all imputations, attendance significant– Without imputations, not significant– Without variables that have missing, attendance is

significant

Page 13: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Table 1: Baseline Models  Imputations No Imputations 

SI Attendance 0.0638***(0.010)

0.024(0.015)

ACT 0.0527***(0.014)

0.015(0.02)

HS GPA 0.272***(0.082)

0.212*(0.12)

Class Size 0.002(0.006)

-0.004(0.009)

Female -0.164*(0.087)

-0.134(0.119)

Minority -0.476***(0.138)

-0.356*(0.192)

Credits Enrolled 0.038***(0.015)

0.038*(0.021)

N 706 309

Unit of Observation is the numberof SI attended.The number in parenthesis is the standard error.*,**,***: Significant at the 10,5, and 1% level, respectively

Page 14: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Table 2: IV, First Stage  Imputations No Imputations 

Age 0.100***(0.0289)

0.141(0.158)

Miles (in %) -0.057(0.117)

0.179(0.15)

Likeliness 0.476***(0.121)

0.564***(0.162)

ACT -0.033(0.049)

0.05(0.059)

HS GPA 0.071(0.303)

-0.05(0.388)

Class Size -0.033(0.023)

-0.078**(0.038)

Female 0.973***(0.323)

0.87*(0.449)

Minority 0.228(0.574)

0.758(0.686)

Credits Enrolled 0.036(0.054)

0.071(0.076)

N 705 309F-Test 10.42 4.95

Unit of Observation is the numberof SI attended.The number in parenthesis is the standard error.*,**,***: Significant at the 10,5, and 1% level, respectively

Page 15: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Table 3: IV, Second Stage

  Imputations No Imputations SI Attendance 0.198***

(0.057)0.107(0.089)

ACT 0.057***(0.015)

0.012(0.02)

HS GPA 0.297***(0.083)

0.233(0.121)

Class Size 0.008(0.007)

0.002(0.011)

Female -0.339***(0.121)

-0.235(0.161)

Minority -0.488***(0.147)

-0.414**(0.202)

Credits Enrolled 0.041***(0.016)

0.035(0.021)

N 705 309Over-ID Test 0.0865 0.9145

Unit of Observation is the number of SI attended.The number in parenthesis is the standard error.*,**,***: Significant at the 10,5, and 1% level, respectively

Page 16: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Table 4: IV, First Stage

  ImputationsAge 0.116***

(0.028)Average Size of SI 0.463***

(0.170)

Class Size -0.024(0.023)

Female 1.21***(0.302)

Minority 0.356(0.56)

Credits Enrolled 0.052(0.054)

N 704F-Test 17.07

Unit of Observation is the number of SI attended.The number in parenthesis is the standard error.*,**,***: Significant at the 10,5, and 1% level, respectively

Page 17: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Table 5: IV, Second Stage

  ImputationsSI Attendance 0.232***

(0.072)Average Size of SI -0.086

(0.071)Class Size 0.012

(0.008)Female -0.33**

(0.132)Minority -0.66***

(0.162)Credits Enrolled 0.07***

(0.017)

N 704Over-ID Test 0.0865

Unit of Observation is the number of SI attended.The number in parenthesis is the standard error.*,**,***: Significant at the 10,5, and 1% level, respectively

Page 18: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Summary

• Running Baseline Models:– Attending SI is significant for larger sample with

imputations included– Without imputations and reduced sample, attendance is

no longer significant

• IV Models:– Age, Logmiles, and Likeliness as instruments– Correlated with Attending SI– Uncorrelated with final grade– Smaller Samples, nothing significant– Without missing variables, attendance is significant

Page 19: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

  All Observations 5 or More Sessions 0 Sessions Final Grade 2.07

(1.22)2.79

(0.98)1.89

(1.25)

Female 0.45(0.498)

0.64(0.48)

0.429(0.495)

Minority 0.087(0.281)

0.117(0.323)

0.083(0.276)

Sophomore Status> 0.423(0.494)

0.58(0.496)

0.381(0.486)

Credits Enrolled 12.61(3.2)

12.42(3.33)

12.53(3.26)

ACT 21.04(3.24)

21.08(2.86)

21.07(3.31)

HS GPA 2.93(0.58)

2.95(0.665)

2.93(0.573)

Previous College GPA 2.71(0.47)

2.79(0.433)

2.667(0.48)

Already SI Participant 0.081(0.273)

0.208(0.408)

0.046(0.209)

Required Course 0.406(0.491)

0.506(0.503)

0.379(0.486)

Expected Grade (4pt ) 3.36(0.56)

3.45(0.527)

3.34(0.565)

Female Professor 0.448(0.498)

0.169(0.377)

0.547(0.498)

Class Size 22.89(8.68)

19.62(8.51)

23.853(8.59)

Average Size of SI 2.59(1.22)

3.348(0.897)

2.32(1.22)

Same Day as Class 0.54(0.498)

0.636(0.484)

0.499(0.5)

Female SI Leader 0.577(0.494)

0.753(0.434)

0.513(0.50)

Section Average GPA 2.09(0.338)

2.15(0.359)

2.05(0.33)

N 705 77 483

Appendix

Page 20: A Regression Analysis of Student Motivation and the Effect of SI on Student Success Kathryn Beck Graduate Student, Applied Economics

Variables