FYE PMP Faculty + Peer Mentor Class
Office of Institutional Research
April 16, 2024
Overview
Each semester has different study samples depending on the PMP-participating courses. If any section of the course involves PMP support, the entire section is included in the sample. For example, the ENG 0114 course has some sections with PM assigned. Then, student records in all ENG 0114 sections are included in the study sample for the semester.
For each study sample, there are three types of course sections. The codes of the course section types are A.PMP(sections with PMP support), B. PMPFaculty (sections without PMP support but taught by faculty who were teaching PMP sections concurrently), and C. NotPMP (sections in which no PMP elements are involved).
The descriptive analysis focused on comparing the proportions of the total sample and specific subsets (such as URM students within the sample) represented by each group. This comparison helps determine whether Group A.PMP proportionally reflects particular student populations.
The outcome variable is course-level grade points (the number of units * gpa point, the maximum 12) each student received in the class. And for the statistical test, a non-parametric method, Kruskall-Wallis was used due to the skewness of the grade point data. If the omnibus test returns a significant result, the post-hoc tests and the interaction effect test (differentiated impacts for the URM students) was conducted.
The second outcome variable is retention based on students’ enrollment status in the immediately following semester. The retained case is if the student enrolled in the semester right after taking the course under study or if the student graduated at the end of the semester when taking the course. For the spring semester data, the retained case has one more scenario where the student graduated during the summer after taking the course. Bar charts present retention rates for each of the three groups. The chi-squared test was used to test statistical group differences. If the chi-squared test is significant, standardized residuals were examined to determine if the PMP sections contributed to that group difference. This allows us to conclude that students in the PMP section are more likely to be retained after taking the target course than those in the other section types.
Key Findings
The spring 2020 sample comprised 1,160 students from 60 course sections. Among them, 36% of the total falls into the the PMP group. A notable demographic composition includes 39% of students identified as Not_URM and 38% of students whose parents have college experience. This demographic distribution suggests that the PMP group has a slightly higher representation of students who are not underrepresented minorities (Not_URM) and those whose parents have college experience. The study could not find evidence of differences in the section types from statistical tests of students’ grade points in the course.When it comes to retention in the immediately following semester, there’s no evidence to suggest a difference in retention rates by the course section type(A.PMP: 89.16%, B.PMPFaculty: 90.08, C.Not_PMP: 89.9).
The fall 2020 sample consists of 2,288 students from 125 course sections. The PMP group comprises 11.54% of the total in the study sample. Similar distributions were found across demographic subsets (e.g., 11.29% of the URM subgroup and 11.58% of the first-generation subset). From statistical tests of students’ grade points in the course, the study could not find evidence of a difference by the section type.When it comes to retention in the immediately following semester, there’s no evidence to suggest a difference in retention rates by the course section type(A.PMP: 89.39%, B.PMPFaculty: 92.31, C.Not_PMP: 90.87).
The spring 2021 sample comprised 1,104 students from 48 sections. Among them, 45% of the total enrolled in PMP sections. A notable demographic composition was found among the Pell-eligible subgroup, 42%, implying that the PMP group slightly under-represents students who are eligible for Pell. The study found a statistical group difference favoring the PMP sections; namely, students in the PMP sections achieved higher GPA points compared with the other two types of sections. See results in detail in the Spring 2021 section. The chi-squared test indicates a significant group difference in the spring 2021 data. The standardized residual results point to the A.PMP group contributed most to this group difference, suggesting that students in the PMP sections are more likely to enroll in the immediately following semester (A.PMP: 93.74%, B.PMPFaculty: 87.61%, C.Not_PMP: 89.31%).See results in detail in the spring 2021 section.
The fall 2021 sample comprised 1,373 students from 79 course sections. In the overall sample, students in PMP sections constitute 19% of the total population. However, when the sample is divided into URM and NURM categories, the proportion of PMP section students increases to 20% within the URM category. This indicates a slight improvement in the representation of URM students within group A, especially considering that in previous years, the percentage of URM students was higher than group A’s overall proportion in the sample.The study found a statistical group difference favoring both the PMP sections and not-PMP sections taught by PMP participant faculty. See results in detail in the fall 2021 section. When it comes to retention outcome, there’s no evidence to suggest a difference in retention rates by the course section type (A.PMP: 92.19%, B.PMPFaculty: 96.08%, C.Not_PMP: 91.43%)
The spring 2022 sample comprised 999 students from 44 course sections. Among them, 27.8% of the total falls into the the PMP group. A notable demographic composition includes 22% of first-generation students and 24% of students who are eligible for the Pell grant. This demographic distribution suggests that the PMP group has a slightly lower representation of those student sub populations. The study could not find evidence of differences in the section types from statistical tests of students’ grade points in the course. When it comes to retention outcome, there’s no evidence to suggest a difference in retention rates by the course section type (A.PMP: 88.85%, B.PMPFaculty: 80.43%, C.Not_PMP: 89.33%)
The fall 2022 sample comprised 2,990 students from 160 sections. Among them, 9.83% of the total enrolled in PMP sections. A notable demographic composition was found among the first generation and pell eligible subgroups, 7.88% and 8.46 %, implying that the PMP group slightly under-represents students whose parents have no college experience and who are eligible for Pell. The study found a statistical group difference favoring the PMP sections; namely, students in the PMP sections achieved higher GPA points compared with the other two types of sections. See results in detail in the fall 2022 section. When it comes to retention outcome, there’s no evidence to suggest a difference in retention rates by the course section type (A.PMP: 92.52%, B.PMPFaculty: 92.31%, C.Not_PMP: 92.55%)
The spring 2023 sample comprised 2,876 students from 150 sections. Among them, 15.82% of the total enrolled in PMP sections. A notable demographic composition was found among the first generation subgroup, 12.92 %, implying that the PMP group slightly under-represents students whose parents have no college experience. The study found a statistical group difference favoring the PMP sections; namely, students in the PMP sections achieved higher GPA points compared with the other two types of sections. See results in detail in the spring 2023 section.When it comes to retention outcome, there’s no evidence to suggest a difference in retention rates by the course section type (A.PMP: 90.33%, B.PMPFaculty: 86.36%, C.Not_PMP: 88.44%)
The fall 2023 sample consists of 2,792 students from 141 course sections. The PMP group comprises 21.53% of the total in the study sample. Similar distributions were found across demographic subsets (e.g., 22.51% of the URM subgroup and 21.51% of the first-generation subset).The study found a statistical group difference favoring the not PMP sections taught by faculty who participated in the PMP. See results in detail in the fall 2023 section.When it comes to retention outcome, there’s no evidence to suggest a difference in retention rates by the course section type (A.PMP: 89.02%, B.PMPFaculty: 92.29%, C.Not_PMP: 89.15%)
Spring 2020
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 35.78% | 11.29% | 52.93% |
| ETHNIC_URM1 - 1.URM | 31.48% | 12.14% | 56.38% |
| ETHNIC_URM1 - 2.NURM | 38.87% | 10.68% | 50.45% |
| PARENT_EDUC_STAT - CollegeExperience | 37.72% | 10.25% | 52.03% |
| PARENT_EDUC_STAT - First Generation | 30.4% | 13.07% | 56.53% |
| PARENT_EDUC_STAT - Unknown | 46.15% | 19.23% | 34.62% |
| PARENT_EDUC_STAT - Unreported | 33.33% | 13.33% | 53.33% |
| PELL_ELIGIBLE_DESC - Not Eligible | 36.06% | 11.21% | 52.73% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 35.4% | 11.4% | 53.2% |
| SEX - F | 35.44% | 11.25% | 53.31% |
| SEX - M | 36.32% | 11.21% | 52.47% |
| SEX - N | 33.33% | 33.33% | 33.33% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 2.1254, df = 2, p-value = 0.3455Summary of Results from Statistical Tests
Considering the distribution of grade points, the non-parametric test (Kruskal-Wallis) was selected to test group differences. The test found that there is no statistical difference in median of grade points across course type–PMP sections, Not-PMP sections taught by faculty who is concurrently teaching PMP sections, and NOT-PMP sections.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 0.17669, df = 2, p-value = 0.9154Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Fall 2020
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 11.54% | 2.27% | 86.19% |
| ETHNIC_URM1 - 1.URM | 11.29% | 2.66% | 86.06% |
| ETHNIC_URM1 - 2.NURM | 11.82% | 1.85% | 86.33% |
| PARENT_EDUC_STAT - CollegeExperience | 11.86% | 1.85% | 86.29% |
| PARENT_EDUC_STAT - First Generation | 11.55% | 2.77% | 85.68% |
| PARENT_EDUC_STAT - Unknown | 11.36% | 2.27% | 86.36% |
| PARENT_EDUC_STAT - Unreported | 6.25% | 3.75% | 90% |
| PELL_ELIGIBLE_DESC - Not Eligible | 11.85% | 1.91% | 86.24% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 11.15% | 2.72% | 86.13% |
| SEX - F | 12.14% | 2.66% | 85.2% |
| SEX - M | 10.61% | 1.71% | 87.67% |
| SEX - N | 25% | 0% | 75% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 1.5083, df = 2, p-value = 0.4704Summary of Results from Statistical Tests
Considering the distribution of grade points, the non-parametric test (Kruskal-Wallis) was selected to test group differences. The test found that there is no statistical difference in median of grade points across course type–PMP sections, Not-PMP sections taught by faculty who is concurrently teaching PMP sections, and NOT-PMP sections.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 0.76188, df = 2, p-value = 0.6832Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Spring 2021
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 44.84% | 10.24% | 44.93% |
| ETHNIC_URM1 - 1.URM | 43.93% | 9.72% | 46.36% |
| ETHNIC_URM1 - 2.NURM | 45.69% | 10.72% | 43.59% |
| PARENT_EDUC_STAT - CollegeExperience | 45.87% | 11.06% | 43.07% |
| PARENT_EDUC_STAT - First Generation | 42.36% | 9.12% | 48.53% |
| PARENT_EDUC_STAT - Unknown | 54.84% | 12.9% | 32.26% |
| PARENT_EDUC_STAT - Unreported | 40.91% | 0% | 59.09% |
| PELL_ELIGIBLE_DESC - Not Eligible | 46.56% | 11.09% | 42.34% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 42.46% | 9.05% | 48.49% |
| SEX - F | 44.34% | 10.54% | 45.12% |
| SEX - M | 45.63% | 9.83% | 44.54% |
| SEX - N | 0% | 0% | 100% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 42.689, df = 2, p-value = 5.374e-10## Kruskal-Wallis rank sum test
##
## data: x and group
## Kruskal-Wallis chi-squared = 42.6886, df = 2, p-value = 0
##
##
## Comparison of x by group
## (Benjamini-Hochberg)
## Col Mean-|
## Row Mean | A.PMP B.PMPFac
## ---------+----------------------
## B.PMPFac | 4.929923
## | 0.0000*
## |
## C.NotPMP | 5.587013 -1.525632
## | 0.0000* 0.0636
##
## alpha = 0.05
## Reject Ho if p <= alpha/2## $chi2
## [1] 42.68858
##
## $Z
## [1] 4.929923 5.587014 -1.525632
##
## $P
## [1] 4.113097e-07 1.155038e-08 6.355075e-02
##
## $P.adjusted
## [1] 6.169646e-07 3.465115e-08 6.355075e-02
##
## $comparisons
## [1] "A.PMP - B.PMPFaculty" "A.PMP - C.NotPMP"
## [3] "B.PMPFaculty - C.NotPMP"## $`1.URM `
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 26.857, df = 2, p-value = 1.473e-06
##
##
## $`2.NURM`
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 17.62, df = 2, p-value = 0.0001492Summary of Results from Statistical Tests
The omnibus group difference test found a significant difference in median GPA points among course section types.
Then, the post-hoc test was conducted, and it found that the GPA points in group A.PMP are significantly higher than those in B.PMPFaculty and C.NotPMP.
With those significant group differences between PMP and other sections, the study examined whether this outcome favoring PMP sections is differential according to the student’s URM status. The interaction effect test found a statistical difference in both sub-groups, suggesting no differential impact of the PMP support on a particular learner population, for example, URM students.
Despite no statistically significant interaction, the partitioned box plot indicates a descriptive interaction effect by the PMP program. For example, the difference between A.PMP versus. B.PMPFaculty(2.7) and C.NotPMP(1.8) for the URM subgroup are greater than those for the NURM subgroup (B.PMPFaculty, 2.5 and C.NotPMP, 0.9).
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 7.9116, df = 2, p-value = 0.01914##
## Retained Not_Retained
## A.PMP 2.753416 -2.753416
## B.PMPFaculty -1.385699 1.385699
## C.NotPMP -1.908493 1.908493Summary of Results from Statistical Tests
The chi-squared test indicates a significant group difference in the spring 2021 retention data. The standardized residual results point to the A.PMP group contributed most to this group difference. Taken all together, we can conclude that students in the PMP sections are more likely to enroll in the immediately following semester.
Fall 2021
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 18.65% | 7.43% | 73.93% |
| ETHNIC_URM1 - 1.URM | 19.61% | 6.14% | 74.25% |
| ETHNIC_URM1 - 2.NURM | 17.73% | 8.65% | 73.62% |
| PARENT_EDUC_STAT - CollegeExperience | 19.31% | 9.1% | 71.59% |
| PARENT_EDUC_STAT - First Generation | 19.21% | 4.42% | 76.38% |
| PARENT_EDUC_STAT - Unknown | 7.84% | 9.8% | 82.35% |
| PARENT_EDUC_STAT - Unreported | 14.29% | 5.36% | 80.36% |
| PELL_ELIGIBLE_DESC - Not Eligible | 18.83% | 8.52% | 72.65% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 18.4% | 5.96% | 75.64% |
| SEX - F | 18.83% | 7.11% | 74.06% |
| SEX - M | 18.31% | 7.92% | 73.77% |
| SEX - N | 33.33% | 0% | 66.67% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 10.424, df = 2, p-value = 0.00545## Kruskal-Wallis rank sum test
##
## data: x and group
## Kruskal-Wallis chi-squared = 10.4241, df = 2, p-value = 0.01
##
##
## Comparison of x by group
## (Benjamini-Hochberg)
## Col Mean-|
## Row Mean | A.PMP B.PMPFac
## ---------+----------------------
## B.PMPFac | -0.421654
## | 0.3366
## |
## C.NotPMP | 2.612226 2.234202
## | 0.0135* 0.0191*
##
## alpha = 0.05
## Reject Ho if p <= alpha/2## $chi2
## [1] 10.4241
##
## $Z
## [1] -0.4216547 2.6122267 2.2342025
##
## $P
## [1] 0.336638550 0.004497729 0.012734876
##
## $P.adjusted
## [1] 0.33663855 0.01349319 0.01910231
##
## $comparisons
## [1] "A.PMP - B.PMPFaculty" "A.PMP - C.NotPMP"
## [3] "B.PMPFaculty - C.NotPMP"## $`1.URM `
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 5.5125, df = 2, p-value = 0.06353
##
##
## $`2.NURM`
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 4.3777, df = 2, p-value = 0.112Summary of Results from Statistical Tests
The omnibus group difference test found a significant difference in median GPA points among course section types.
Then, the post-hoc test was conducted, and it found that the GPA points in group A.PMP and P.PMPFaculty are significantly higher than those in C.NotPMP. It is not statistically significant that the B.PMPFaculty group is higer than the A.PMP group.
With those significant group differences among section types, the study examined whether this outcome favoring PMP or PMPFaculty sections is differential according to the student’s URM status. In both subgroups by the URM status, there is no group difference , suggesting no differential impact of the PMP support on a particular learner population, for example, URM students.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 2.7281, df = 2, p-value = 0.2556Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Spring 2022
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 27.83% | 4.6% | 67.57% |
| ETHNIC_URM1 - 1.URM | 26.82% | 5.79% | 67.38% |
| ETHNIC_URM1 - 2.NURM | 28.71% | 3.56% | 67.73% |
| PARENT_EDUC_STAT - CollegeExperience | 31.47% | 5.11% | 63.43% |
| PARENT_EDUC_STAT - First Generation | 21.91% | 4.32% | 73.77% |
| PARENT_EDUC_STAT - Unknown | 28% | 0% | 72% |
| PARENT_EDUC_STAT - Unreported | 11.11% | 5.56% | 83.33% |
| PELL_ELIGIBLE_DESC - Not Eligible | 30.6% | 4.44% | 64.96% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 23.91% | 4.83% | 71.26% |
| SEX - F | 26.69% | 3.46% | 69.85% |
| SEX - M | 29.64% | 6.44% | 63.92% |
| SEX - N | 25% | 0% | 75% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 1.7368, df = 2, p-value = 0.4196Summary of Results from Statistical Tests
Considering the distribution of grade points, the non-parametric test (Kruskal-Wallis) was selected to test group differences. The test found that there is no statistical difference in median of grade points across course type–PMP sections, Not-PMP sections taught by faculty who is concurrently teaching PMP sections, and NOT-PMP sections.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 3.4271, df = 2, p-value = 0.1802Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Fall 2022
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 9.83% | 1.3% | 88.86% |
| ETHNIC_URM1 - 1.URM | 9% | 0.87% | 90.13% |
| ETHNIC_URM1 - 2.NURM | 10.67% | 1.74% | 87.58% |
| PARENT_EDUC_STAT - CollegeExperience | 10.86% | 1.72% | 87.42% |
| PARENT_EDUC_STAT - First Generation | 7.88% | 0.7% | 91.42% |
| PARENT_EDUC_STAT - Unknown | 11.34% | 0% | 88.66% |
| PARENT_EDUC_STAT - Unreported | 9.3% | 1.16% | 89.53% |
| PELL_ELIGIBLE_DESC - Not Eligible | 11.11% | 1.55% | 87.34% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 8.46% | 1.04% | 90.5% |
| SEX - F | 10% | 1.37% | 88.63% |
| SEX - M | 9.66% | 1.25% | 89.1% |
| SEX - N | 7.69% | 0% | 92.31% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 12.272, df = 2, p-value = 0.002163## Kruskal-Wallis rank sum test
##
## data: x and group
## Kruskal-Wallis chi-squared = 12.2722, df = 2, p-value = 0
##
##
## Comparison of x by group
## (Benjamini-Hochberg)
## Col Mean-|
## Row Mean | A.PMP B.PMPFac
## ---------+----------------------
## B.PMPFac | 2.754992
## | 0.0044*
## |
## C.NotPMP | 2.941203 -1.789999
## | 0.0049* 0.0367
##
## alpha = 0.05
## Reject Ho if p <= alpha/2## $chi2
## [1] 12.2722
##
## $Z
## [1] 2.754993 2.941204 -1.790000
##
## $P
## [1] 0.002934671 0.001634696 0.036726975
##
## $P.adjusted
## [1] 0.004402007 0.004904088 0.036726975
##
## $comparisons
## [1] "A.PMP - B.PMPFaculty" "A.PMP - C.NotPMP"
## [3] "B.PMPFaculty - C.NotPMP"## $`1.URM `
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 5.7233, df = 2, p-value = 0.05718
##
##
## $`2.NURM`
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 10.148, df = 2, p-value = 0.006258Summary of Results from Statistical Tests
The omnibus group difference test found a significant difference in median GPA points among course section types.
Then, the post-hoc test was conducted, and it found that the GPA points in group A.PMP are significantly higher than the other two comparison groups.
With those significant group differences among section types, the study examined whether this outcome favoring PMP sections is differential according to the student’s URM status. The interaction effect test found a statistical difference in both sub-groups, suggesting no differential impact of the PMP support on a particular learner population, for example, URM students.
Despite no statistically significant interaction, the partitioned box plot indicates a descriptive interaction effect by the PMP program. For example, the difference between A.PMP versus. B.PMPFaculty(3.24) for the URM subgroup is greater than that for the NURM subgroup (1.65), which suggests that when faculty teach the same course, having peer mentor support is more beneficial to the URM students.However, this descriptive results should be cautiously used because of the small sample size of the B. PMPFaculty (only 39 students from 2 sections).
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 0.0035061, df = 2, p-value = 0.9982Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Spring 2023
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 15.82% | 8.41% | 75.76% |
| ETHNIC_URM1 - 1.URM | 14.6% | 8.77% | 76.63% |
| ETHNIC_URM1 - 2.NURM | 17.02% | 8.06% | 74.91% |
| PARENT_EDUC_STAT - CollegeExperience | 18.15% | 9.07% | 72.78% |
| PARENT_EDUC_STAT - First Generation | 12.92% | 7.28% | 79.79% |
| PARENT_EDUC_STAT - Unknown | 11.19% | 7.46% | 81.34% |
| PARENT_EDUC_STAT - Unreported | 9.88% | 9.88% | 80.25% |
| PELL_ELIGIBLE_DESC - Not Eligible | 16.4% | 8.63% | 74.97% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 15.2% | 8.18% | 76.63% |
| SEX - F | 15.71% | 7.79% | 76.5% |
| SEX - M | 16.16% | 9.24% | 74.6% |
| SEX - N | 0% | 6.67% | 93.33% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 23.417, df = 2, p-value = 8.225e-06## Kruskal-Wallis rank sum test
##
## data: x and group
## Kruskal-Wallis chi-squared = 23.4166, df = 2, p-value = 0
##
##
## Comparison of x by group
## (Benjamini-Hochberg)
## Col Mean-|
## Row Mean | A.PMP B.PMPFac
## ---------+----------------------
## B.PMPFac | 3.555199
## | 0.0003*
## |
## C.NotPMP | 4.676933 -0.616777
## | 0.0000* 0.2687
##
## alpha = 0.05
## Reject Ho if p <= alpha/2## $chi2
## [1] 23.41656
##
## $Z
## [1] 3.5551991 4.6769337 -0.6167775
##
## $P
## [1] 1.888465e-04 1.455982e-06 2.686907e-01
##
## $P.adjusted
## [1] 2.832697e-04 4.367945e-06 2.686907e-01
##
## $comparisons
## [1] "A.PMP - B.PMPFaculty" "A.PMP - C.NotPMP"
## [3] "B.PMPFaculty - C.NotPMP"## $`1.URM `
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 2.3666, df = 2, p-value = 0.3063
##
##
## $`2.NURM`
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 27.674, df = 2, p-value = 9.787e-07Summary of Results from Statistical Tests
The omnibus group difference test found a significant difference in median GPA points among course section types.
Then, the post-hoc test was conducted, and it found that the GPA points in group A.PMP are significantly higher than the other two comparison groups.
With those significant group differences among section types, the study examined whether this outcome favoring PMP sections is differential according to the student’s URM status. The interaction effect test found a statistical difference in only the NURM sub-group, suggesting a differential impact of the PMP support on grade points for students who are not URM background. a particular learner population, for example, URM students. The partitioned box plot echoes this result that there is group difference in student course outcomes favoring the A.PMP group for the group of not URM students.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 2.5924, df = 2, p-value = 0.2736Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
Fall 2023
| Demographic_Level | A.PMP | B.PMPFaculty | C.NotPMP |
|---|---|---|---|
| Total | 21.53% | 19.05% | 59.42% |
| ETHNIC_URM1 - 1.URM | 22.51% | 17.89% | 59.6% |
| ETHNIC_URM1 - 2.NURM | 20.32% | 20.48% | 59.2% |
| PARENT_EDUC_STAT - CollegeExperience | 21.84% | 18.75% | 59.41% |
| PARENT_EDUC_STAT - First Generation | 21.51% | 18.6% | 59.89% |
| PARENT_EDUC_STAT - Unknown | 20% | 21.74% | 58.26% |
| PARENT_EDUC_STAT - Unreported | 18.39% | 26.44% | 55.17% |
| PELL_ELIGIBLE_DESC - Not Eligible | 20.36% | 20.36% | 59.28% |
| PELL_ELIGIBLE_DESC - Pell Eligible | 22.6% | 17.86% | 59.55% |
| SEX - F | 20.66% | 20.21% | 59.13% |
| SEX - M | 22.41% | 17.69% | 59.9% |
| SEX - N | 30.77% | 15.38% | 53.85% |
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 21.478, df = 2, p-value = 2.169e-05## Kruskal-Wallis rank sum test
##
## data: x and group
## Kruskal-Wallis chi-squared = 21.4775, df = 2, p-value = 0
##
##
## Comparison of x by group
## (Benjamini-Hochberg)
## Col Mean-|
## Row Mean | A.PMP B.PMPFac
## ---------+----------------------
## B.PMPFac | -3.545984
## | 0.0003*
## |
## C.NotPMP | 0.333663 4.555421
## | 0.3693 0.0000*
##
## alpha = 0.05
## Reject Ho if p <= alpha/2## $chi2
## [1] 21.47754
##
## $Z
## [1] -3.5459843 0.3336636 4.5554219
##
## $P
## [1] 1.955747e-04 3.693167e-01 2.614027e-06
##
## $P.adjusted
## [1] 2.933621e-04 3.693167e-01 7.842080e-06
##
## $comparisons
## [1] "A.PMP - B.PMPFaculty" "A.PMP - C.NotPMP"
## [3] "B.PMPFaculty - C.NotPMP"## $`1.URM `
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 11.697, df = 2, p-value = 0.002884
##
##
## $`2.NURM`
##
## Kruskal-Wallis rank sum test
##
## data: GRADE_POINTS by group
## Kruskal-Wallis chi-squared = 6.5643, df = 2, p-value = 0.03755Summary of Results from Statistical Tests
The omnibus group difference test found a significant difference in median GPA points among course section types.
Then, the post-hoc test was conducted, and it found that the GPA points in group B.PMPFaculty are significantly higher than the other two comparison groups.Namely, sections that were taught by faculty who participated in the PMP without PM assigned generated the best course outcomes.
With those significant group differences among section types, the study examined whether this outcome favoring Non-PMP sections taught by PMP-opt-in faculty is differential according to the student’s URM status. The interaction effect test found a statistical difference in both sub-groups, suggesting no differential impact on a particular learner population, for example, URM students.
Retention
##
## Pearson's Chi-squared test
##
## data: tab
## X-squared = 4.7246, df = 2, p-value = 0.0942Summary of Results from Statistical Tests
There is no significant difference between A.PMP, B.PMPFaculty, and C.NotPMP groups in the retention rate.
