Presentation is loading. Please wait.

Presentation is loading. Please wait.

Presented by: Office of Institutional Research (UNCG-IR) November 2017

Similar presentations


Presentation on theme: "Presented by: Office of Institutional Research (UNCG-IR) November 2017"— Presentation transcript:

1 Presented by: Office of Institutional Research (UNCG-IR) November 2017
Factors Impacting 1-Year Retention of Fall 2016 First-time Undergraduate Students at UNCG Presented by: Office of Institutional Research (UNCG-IR) November 2017

2 Executive Summary (Page 1 of 5):
The purpose of this study was to explore the known independent variables (IV) to determine their impact on the dependent variable (DV) of 1-year retention. The study described, in cross tabs, the sample of the Fall 2016 entering first-time full-time undergraduate students at UNCG, and tested their correlations with 1-year retention using Logistic Regression. The binary IV’s explored included: Demographics: female vs. male, White vs. everyone else, Black vs. everyone else, Geography: Guilford County vs. everyone else, Wake County vs. everyone else, Mecklenburg County vs. everyone else, in-state vs. out-state, Pre-college/high school qualifications: SAT total 1,030 or higher vs. below 1,030, high school GPA 3.5 vs. below 3.5, high school (percent) rank 70 vs. below 70, Socio-Economic Status: Pell recipients vs. no Pell, low income vs. not low income, family income $60k or more vs. below, gift aid 10k vs. below, College factors (factors associated with UNCG): on-campus vs. off-campus, honors class, entering fall term gpa 2.5 or better vs. below 2.5, dropped class vs. not in the entering fall term, College of Arts & Sci vs. others, College of Health & Human Services vs. others, School of Nursing vs. others, School of Nursing vs. others.

3 Executive Summary (Page 2 of 5):
Via cross tables comparing the row percent differences of the groups, noticeable differences in 1-year retention were evidenced by: Black students were retained 5-6% more than non-Black students, female students were retained 8% more than male students, White students were retained 3% less than non-White students, North Carolina students were retained 5% more than out of state students, but after equalizing the in- and out-state groups on high school gpa and SAT total, there was no difference between the two groups. We therefore conclude that state residence does not make a significant difference in 1-year retention. Those who brought a high school gpa 3.5 or better were retained 17% more than those who had a high school gpa below 3.5, High school class rank 70 or better were retained 12% more than below 70. Those who had $10k or more in gift aid were retained 8-9% more than those who had below $10k, College of Arts and Sci students were retained 4% less than other students, Those who lived on campus were retained about 4-5% more than those off campus, Honors students were retained at least 10% more than those who were not in Honors, Those who dropped no class in the fall term were retained at least 25% more than those who dropped 1 or more classes, Those who had a fall term gpa 2.5 or better were retained 36% more than those who had below 2.5.

4 Executive Summary (Page 3 of 5):
Having learned about the descriptive statistics, we may follow up to ask: “How do these independent variables interact with one another to impact 1-year retention?” To answer such a question, we ran Logistic Regressions to find that, in an order of large impact to small impact, the factors below significantly correlated with 1-year retention: Fall Term GPA 2.5+ (4.487) Honors Class (4.304) School of Edu (3.249) GiftAid$10k (0.912) School of Nursing (0.648) From North Carolina (-0.536) Dropped Class in fall term (-0.519) High School GPA (0.43) Mecklenburg County (-0.301) While the list represents separate runs of the Logistic Regression, the results are comparable odds ratio estimates. The School of Education was high on the list because it was based on the PSM matched data which considered high school gpa and SAT. Most factors on the list were positive impacts on 1-year retention. However, North Carolina was a negative impact, as after the students were equalized on high school gpa and SAT, in-state students proved to have been retained less than their out-state counterparts. Dropping one or more classes in the entering fall term, compared to dropping no class, was clearly a negative impact on 1-year retention. So was Mecklenburg County. If we want to scrutinize the size of the impact, for example, we see that Fall Term GPA 2.5+ is about four times as large an impact as GiftAid$10k (( )/0.912=3.92).

5 Executive Summary (Page 4 of 5):
As fall term gpa showed itself as the No. 1 impact on 1-year retention, a further question arises: “What courses may have contributed to the term gpa that made the differences in retention? To answer such a question, we pulled the “popular” (enrolling 80 or more students) classes that the Fall 2016 entering first-time full-time students took in Fall 2016 to find that: ENG105, REL101, CHE111, BUS105, and PSY121 all had a DFW rate as high as 33% or more. GEO103, CHE103, BIO111, HHS125 and ENG101 had a 1-year retention rate as high as over 85% compared to REL101, ENG104 and FMS120 that had a 1-year retention rate as low as below 70%. Some classes did an excellent job in retaining both ABC students and DFW students: ENG101, BIO101, BUS105 CHE112, GEO103 set good examples. Some other classes had a relative low retention of both ABC students and DFW students: RCO215 ENG104, FMS120 and REL101 may need an improvement. In yet other classes, the gap between ABC retention and DFW retention was so shockingly large that failing the class meant almost departure from the University: in BIO100 only 5% of the DFWs were retained to the next fall compared to 87% of the ABC students; in BIO110 only 18% of the DFW students were retained compared to 85% of the ABC students; in PSC100 only 19% of the DFWs were retained compared to 83% of the ABS students. Apparently while the high retention rate of the ABC students needs to continue, the low retention of the DFW students needs to be improved.

6 Executive Summary (Page 5 of 5):
Concluding remarks: Retention of the students is the core of the University business, as student success is measured by graduation and a student has to be retained to graduate. There may have been a myth that pre-college factors, particularly standardized test scores defined retention and student success. This study proved that SAT does not matter as much as the academic performance, measured by term gpa that the student earns after arriving on campus. As the entering fall term gpa was the largest impact on 1-year retention, those courses that had a high DFW rate and a low retention rate need to be identified and worked on in order to enhance retention at the University.

7 Demographic variables were formulated and tested.
Descriptives: The significant level of the difference between Black students and non- Black students was tested to be 0.00, which confirms the statistical significance. There is a hypothesis that the seeming difference in 1-year retention between Black and non-Black may be due to other factors than ethnicity, for example high school gpa and SAT score. In response to that, “propensity score matching” (PSM) was applied to find a group in the population with similar high school gpa and SAT total. A comparison of the PSMed groups proves that the difference is significant at the 95% confidence level. Black students were retained 82% compared to 76% of those who were not Black. Demographic variables were formulated and tested.

8 After applying “PSM” on high school gpa and SAT total, there was still no difference between the two groups in each comparison. At the 95% confidence level, as tested by X2, there is no significant difference between those from Guilford Cnty vs. Others, no sig dif between those from Mecklenburg Cnty vs. Others, no sig dif between those from Wake Cnty vs. Others, and no sig dif between those from North Carolina vs. Others, although the 1-year retention of N Carolinians was 5% higher. Descriptives: As home proximity to campus makes a dif in retention. Some geography variables were formulated and tested.

9 At the 95% confidence level, as tested by X2, there is sig difference between those whose high school gpa is 3.5 or greater vs. below, there is sig dif between those whose high school rank is 70 vs. below, but there is no sig dif between those whose SAT total is 1030 vs. below. Descriptives: High school preparations such as gpa, rank and test scores have been proved to have an impact on 1-year retention.

10 Descriptives: At the 95% confidence level, as tested by X2, there is sig difference between those who had $10,000 in gift compared to those who received a gift amount below $10,000. If you further break down the groups, you see that those who received 10-15k were most likely to be retained (84%), compared to 5-10k 80% and 5k or less 72%. After we apply “PSM” to match a similar group on high school gpa and SAT total, the X2 shows there is a significant difference between the Low Income and the Not Low Income, with a numeric difference of 6%. Socio-Economic factors were hypothesized to have an impact on 1-year retention.

11 Descriptives: Students in the College of Arts and Sci were retained at a significantly lower percentage than those who were not in Arts and Sci: 76% vs. 80%. Nursing students were retained more than those who were not in Nursing: 83% vs. 77%. However, there was no sig dif between BE vs. Others, no sig dif between Edu vs. Others, no sig dif between HHuman vs. Others. On campus students are retained more than off campus students: 78.5% vs. 74%, significant difference at the 95% confidence level. However some would argue that the difference may not be due to residence but because on-campus students are of better quality as measured by high school gpa and SAT. To prove that, PSM was applied on high school gpa and SAT. The matched groups were compared to find that there is a 4% difference between the groups, instead of 4.5%, and it is not significant at 95% confidence level. After new students arrive on campus, what happens to them at the University definitely impacts their decision about whether to continue to next year. These variables were tested:

12 Logistic Regression Findings:
A dumb bell away from the Bar of 1 indicates a statistically significant impact, while a dumb bell across the Bar of 1 indicates no significant impact. The left region of the Bar of 1 indicates a negative impact while the right region indicates a positive impact. The length of the dumb bell indicates the power/strength of correlation. Gift2Grp ($10k in gift aid) , DrpCrsYN (drop class in fall term), Mecklbg (coming from Mecklenburg Cnty), HSgpa (high school gpa 3.5), termGPA (entering fall term gpa 2.5) have a significant impact on 1-year retention. DrpCrsYN and Mecklbg are to the left of the Bar of 1, thus a negative impact, while termGPA, Hsgpa, Gift2Grp have a positive impact. As the dumb bell of termGPA is the longest., it is the strongest impact on 1-year retention.

13 Logistic Regression Findings:
The PSM procedure matched the groups on high school gpa and SAT total. Logistic Regression was run on the matched data to find: BecklbgB (Mecklenburg Cnty), StateB (from N. Carolina), EduB (Sch of Education), NursingB (Sch of Nursing), HonorsB (Honors Class) , DrpCrsYN (dropped class in the entering fall term) Gift2Grp ($10k in gift aid) had a significant impact on 1-year retention. While Mecklenburg Cnty, North Carolina and Dropping Class have a negative impact, Sch of Education, Sch of Nursing, Honors Class, and Gift Aid $10K have a positive impact on 1-year retention. Logistic Regression Findings: If you look for the size of the impact, as indicated by the length of the dumb bell and the numeric value, Honors Class (4.304), of all the PSMed variables, is the largest impact on 1-year retention.

14 Summary of Findings:

15 Fall Term GPA to Course Grades:
As it was found that of all the factors that we explored, the fall term GPA exerted the largest impact on 1-year retention of the Fall 2016 entering first-time undergraduate students, natural questions to follow are: 1) What courses awarded more failing grades (DFW) than others? 2) What types of correlation do the grades in these courses have with 1-year retention at the University? To respond to the questions, we pulled “popular” courses, enrolling at least 80 students, that the fall 2016 entering first-time undergraduate students took to test our hypotheses. The following slides show the findings:

16 Some courses really failed many students, as many as 1/3 !!!
Failing Rate: Some courses really failed many students, as many as 1/3 !!!

17 Failing Rate: Retention (to the next fall at the U) varies from course to course: some courses retained just about 2/3 of all the students, others retained over 9/10!!!

18 4 zones: Obs 2: Some courses did a remarkable job of retention in that they retained HIGH not only the ABC students but also the DFW students Obs 3: Most courses just did the “right” thing: they retained HIGH ABC students, over 80% in all the cases. Obs 4: Several courses really need to work on their retention as they not only retained LOW DFW students, but also retained LOW of ABC students. Obs 1: No courses retained DFW students high and ABC low, which is another proof that better grades transfer into higher retention.

19 Dif ABC and DFW: Some courses had a large gap between retention of
ABC students and that of the DFW students: BIO100, BIO110, PSC100, to name just a few.


Download ppt "Presented by: Office of Institutional Research (UNCG-IR) November 2017"

Similar presentations


Ads by Google