Donate
Independent, objective, nonpartisan research

Improving College Pathways in California, Technical Appendix

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3711 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(20) "1117ngr_appendix.pdf" ["wpmf_size"]=> string(6) "292336" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(28477) "Improving College Pathways in California Technical Appendices CONTENTS Appendix A: Pipeline Model Appendix B: Descriptive Summary of Our Sample Appendix C: Additional Figures for Our Pathway Analysis Niu Gao, Hans Johnson 2 5 8 Supported with funding from the Dirk and Charlene Kabcenell Foundation, the Evelyn and Walter Haas, Jr. Fund, the James Irvine Foundation, the Leona M. and Harry B. Helmsley Charitable Trust, and the Sutton Family Fund Appendix A: Pipeline Model Developing a Pipeline Model To estimate where students fall off the pathway to college readiness and completion, we develop a pipeline model that identifies key transitions as students move through their educational trajectories. The pipeline model shown in Figure 1 represents a synthetic cohort of 9th graders in public schools in California as they move through high school, complete the a–g courses, enter college (public or private, in California or another state), transfer from community college, and complete a bachelor’s degree. The primary assumption of the model is that current rates of high school completion, a–g completion, college enrollment, transfer from community college, and college completion prevail throughout the cohort’s high school and college years. Pipelines are developed separately for four large ethnic groups (Latinos, whites, Asian Americans, and African Americans) and two genders (female and male). Four key transitions are identified and estimated in the model: (1) 9th grade to high school graduation (including a–g completion), (2) high school graduation to college enrollment (including community colleges and four-year colleges), (3) enrollment in four-year colleges (either as freshmen or transfer students), and (4) college enrollment to college completion (of a bachelor’s degree). • The first transition is based on cohort graduation rates provided by the California Department of Education for the 2015–16 graduating class. Completion of the a–g courses is based on rates among public high school graduates in 2015–16, also provided by the California Department of Education. • The second transition is based on enrollment rates of recent high school graduates in community colleges and four-year colleges. College enrollment rates are calculated separately by type of college. California State University, the University of California, and the California Community Colleges provided 2015 data on public high school of origin for incoming freshmen. Data on private and out-of-state college enrollment rates was calculated from 2014 IPEDS data, adjusted to reflect public high school graduates. IPEDS data do not include information on ethnicity and gender of students based on their state of residence. Therefore, estimates of enrollment rates to private and out-of-state colleges by ethnicity and gender were derived by applying ethnic and gender distributions of young college-enrolled migrants (students who had left California) based on American Community Survey data. • The third transition is derived from data on 2014–15 transfers from Community Colleges to UC, CSU, private colleges, and out-of-state colleges as provided by institutional research offices of UC, CSU, and the community colleges. • The final transition to college completion is based on six-year graduation rates of incoming freshmen and four-year graduation rates of transfer students. Rates are calculated separately by gender and ethnicity for UC, CSU, private colleges in California, and out-of-state colleges. FIGURE 1 Among those not obtaining a bachelor’s degree, most fall off the pathway before making it to a four-year college Attended a four-year college, but did not earn a degree Attended college, but did not attend a four-year college Finished high school, but did not attend college Did not finish high school 0% 18% 31% 26% 25% 5% 10% 15% 20% 25% 30% 35% Student Pathways to College FIGURE 2 Latino and African American students are underrepresented in the pathways to college 100% 80% 3% 26% 4% 26% 4% 30% 9% 22% 11% 21% 16% 21% % of students 60% 40% 52% 52% 42% 49% 47% 26% 4% 20% 0% 7% 12% Grade 9 enrollment 7% 11% 5% 20% 6% 13% 4% 17% HS graduates a-g completion CCC first-time freshmen CSU first-time freshmen Asian African American Latino White Other 33% UC first-time freshmen SOURCE: Grade 9 enrollment: California Department of Education, 2011–12; a–g completion: California Department of Education, 2015–16, CCC first-time freshmen enrollment: California Community College, 2016; CSU first time freshmen from California high schools: California State University, 2015, UC first time freshmen: University of California, 2016. The need to focus on a–g courses Statewide, at current rates of high school completion and preparation for college, only 30% of 9th graders will earn a high school diploma and complete the a–g college preparatory courses required for entry to UC and CSU. These rates of completion and readiness vary substantially by gender and ethnicity. For example, only 25 percent of African American 9th graders will complete both high school and the a–g courses, compared to 65 percent of Asian Americans. Male 9th graders are less likely to finish high school prepared for college than females (32% and 44% respectively). While high school dropout is an important problem and an obvious reason that some students are not prepared for college, completing the a–g college preparatory courses is an even larger challenge. As shown in Table 1, for every ethnic group and for females and males, the share of high school students who are not ready for college because they did not complete the a–g courses is substantially higher than the share not ready because they did not finish high school. TABLE 1 Most 9th graders will not complete high school ready for college, with large differences by gender and ethnicity Female African American Among the cohort of 9th graders in the class of 2016: Did not finish high school Did not complete the a–g courses Total share not ready for college 21% 47% 68% Latino 16% 48% 63% White 9% 39% 48% Asian American 5% 23% 28% All females 13% 43% 56% Male African American Latino White Asian American All males 33% 24% 14% 8% 20% 48% 52% 46% 33% 48% 81% 77% 60% 41% 68% Appendix B: Descriptive Summary of Our Sample K–12 Sample Our K–12 sample includes 472,324 high school students in 24 diverse districts from 2007 to 2014 school years. Geographically, these districts are well spread across the state (inland and coastal, northern and southern). As seen in Table 1, African American and low-income students are overrepresented in our sample when compared to statewide average. Students in our sample also tend to be low-performing. Note that we also have more other/missing (including inconsistent) values for ethnicity in our dataset. Due to the differences in student characteristics between our sample and statewide averages, our findings can only be applied to similar settings, e.g., schools with a large concentration of low-income and low performing students. This is not necessarily a drawback of our study, as recent reforms (e.g., Every Student Succeeds Act and its predecessor No Child Left Behind) have focused on turning around low-performing schools. TABLE 1 Descriptive summary of key student characterIstics in our k–12 sample Sample State Female 49% 49% Asian American 9% 8% African American 13% 8% Latino 34% 45% White 29% 34% Other/missing 15% 5% Eligible for free/reduced price lunch 64% 50% Parent education: college degree or above 31% - CST: math 278 312 CST: English 315 346 % graduates completing a–g 20% 38% N Graduates 141,307 2,832,012 SOURCE: Cal-PASS Plus, 2007 – 2014; California Department of Education, 2007–2014. Community College Sample We matched the high school graduates in our sample to 18 nearby community colleges and followed them over time. Our analytical sample includes 16,792 first-time college students from 2011–2015. Since we restricted our sample to high school graduates (1) who have complete four-year transcripts (so we can identify their a–g completion status), and (2) who enroll in a nearby community college immediately following high school graduation (due to data limitations), our sample is a very selective one; students in our sample have better outcomes than state averages. For instance, 45 percent of students in our sample take at least one transfer-level course in Math or English, and only a third of students are ever enrolled in a developmental education course during their two years in college (compared to 80 percent statewide within six years). Latino and African American students are underrepresented in our sample, but more students in our sample are first generation college students. A descriptive summary of our community college sample is in Table 2 below. TABLE 2 Descriptive summary of key student characterIstics in our Community college sample Variable Asian American CC statewide 7% 15% African American White Latino Female Pell grant recipients BOGG recipients free/reduced lunch eligible First generation CCC outcomes % taking transfer-level math/English Average passing rate of transfer math/English High school preparation 4% 51% 31% 52% 24% 42% 34% 56% 45% 77% 6% 26% 44% 53% 22% 49% 40% - # a–g math taken # a–g science taken # a–g English taken # a–g social science taken # a–g foreign language taken # a–g art taken 323222- Average GPA in a–g math Average GPA in a–g science Average GPA in a–g English Average GPA in a–g social science Average GPA in a–g foreign language Average GPA in a–g art Highest math taken: higher math Highest math taken: algebra 2 or equivalent Highest math taken: geometry or equivalent 2.3 2.5 2.6 2.7 2.7 3.3 42% 30% 14% - N 16,792 2,355,774 SOURCE: Cal-PASS Plus, 2007–2014; California Community College Data Mart, various years. CSU Sample A descriptive summary of our CSU sample has been summarized in Table 3. Our analytical sample includes 3,004 first-time freshmen students in three CSU campuses (with complete high school transcripts) from 2011 to 2015 school years. Latino students are underrepresented in our sample, and more students in our sample are lowincome and first generation college students. First-year and second-year persistence is lower among our sample, which is not surprising given that most students in our sample are low-income or first generation college students. TABLE 3 Descriptive summary of key student characteristics in our CSU sample Asian American Our Sample 30% CSU system wide (2014) 17% African American White Latino 13% 32% 22% 4% 23% 43% Female Low-income students First generation CSU outcomes First-year persistence Second-year persistence Average year end units earned Average year end GPA High school preparation Average GPA in a–g math Average GPA in a–g science Average GPA in a–g English 60% 64% 68% 85% 57% 36% 77% 61% 21 2.39 84% 76% - 2.71 2.93 3.18 - Average GPA in a–g social studies Average GPA in a–g foreign languages Average GPA in a–g art Average GPA in a–g elective Took a–g math in 12th grade Took a–g science in 12th grade 3.19 3.15 3.64 3.19 67% 54% - Took a–g social science in 12th grade Highest level by 11th grade: higher math 82% 40% - N 3,004 62,523 Source: Cal-PASS Plus, 2011–2015; California State University, 2014. Note: We use students’ eligibility for free/reduced price lunch in high school to determine their low-income status; and CSU uses a different metric—Pell grant recipients—which contributes to the difference in low-income students counts. About 12 percent of students in our CSU sample (not the K–12 sample) miss their free/reduced price lunch information; and about 20 percent of students miss their parental education, which may skew our share of low-income, first generation students. Appendix C: Additional Figures for Our Pathway Analysis TABLE 1 % high school graduates completing each subject area requirements with a C or better, by student characteristics, 2007–2014 Social Science (a) English (b) Math (c ) Science (d) Foreign Language (e) Female 60% 40% 45% 56% 56% Male 54% 32% 39% 51% 46% Asian American 71% 47% 49% 75% 70% African American 44% 23% 31% 41% 36% White 65% 45% 46% 59% 58% Latino 49% 27% 39% 46% 42% Free/reduced price lunch eligible First generation college students 47% 51% 25% 29% 38% 40% 45% 47% 41% 43% Overall 57% 36% 42% 54% 51% SOURCES: Authors’ calculations using Cal-PASS Plus data. TABLE 2 Average passing rate in a–g courses, by subject area and by student characteristics, 2007 –2014 Art (f) Electives (g) 77% 71% 79% 64% 80% 69% 68% 70% 74% 84% 81% 80% 79% 84% 83% 82% 82% 82% First Asian African Low- generation All Female American American Latino White income college Social Science (a) 88% 91% 95% 81% 85% 91% 83% 86% English (b) 87% 90% 95% 80% 83% 91% 81% 84% Math (c) 82% 85% 92% 73% 78% 87% 77% 78% Science (d) 88% 90% 95% 77% 84% 92% 84% 85% Foreign Language (e) 91% 93% 96% 83% 89% 94% 88% 89% Art (f) 93% 95% 97% 87% 91% 95% 89% 91% Electives (g) 89% 92% 95% 82% 86% 93% 85% 87% SOURCE: Authors’ calculations using Cal-PASS Plus data. NOTE: (1). Sample includes 472,324 high school students from 2007 to 2014 school years. (2). Overall passing rate is higher in later grades (e.g., Grades 11 and 12). (3). The same conclusion holds for all subgroups in all subject areas. PPIC.ORG Technical Appendices Improving College Pathways in California 8 TABLE 3 % schools offering a–g courses, by subject areas and by grade, 2007–2014 Math Science English Social Science Grade 9 96% 96% 96% 95% Grade 10 97% 97% 97% 97% Grade 11 97% 96% 96% 96% Grade 12 97% 97% 97% 96% NOTE: (1). Sample includes 105 regular high schools from 2007 to 2014 school years. (2). Social science sequences includes one year of world history, cultures, and historical geography and one year of US history, both of which are typically offered in grades 10 or above. TABLE 4 Math progression, by student characteristics, 2007–2014 Overall Female Male Asian African American American Earned A or B in algebra 1, took geometry Earned A or B in geometry , took algebra 2 Earned A or B in algebra 2, took higher math 69% 68% 59% 70% 68% 58% 69% 68% 59% 72% 73% 66% 66% 54% 48% SOURCE: Authors’ calculations using CalPASS Plus data. NOTE: Sample restricted to schools offering the full a–g math sequence. TABLE 5 Math progression, by pathway and student characteristics, 2007 – 2014 White 64% 67% 57% Latino 68% Lowincome First generation 74% 75% 63% 68% 69% 53% 57% 58% Female Asian African American American Pathway I (start algebra 1 before 9th grade) % algebra 1 completers moving to geometry 82% 91% % geometry completers moving to algebra 2 66% 70% % algebra 2 completers moving to higher level math 52% 58% Pathway II (start algebra 1 in 9th grade) % algebra 1 completers moving to geometry 66% 72% % geometry completers moving to algebra 2 56% 63% % algebra 2 completers moving to higher level math 24% 31% Pathway III (start algebra 1 in 10th grade) % algebra 1 completers moving to geometry 56% 62% % geometry completers moving to higher level math 26% 32% Pathway IV (start algebra 1 in 11/12th grade) % algebra 1 completers moving to geometry 32% 37% NOTE: Sample includes 472,324 high school students from 2007 to 2014 school years. TABLE 6 Science and English pathways, by student characteristics, 2007–2014 74% 59% 40% 64% 49% 21% 57% 26% 34% White 77% 60% 44% 62% 53% 19% 49% 20% 26% Latino Low- First income generation 77% 63% 49% 79% 65% 51% 67% 54% 25% 70% 55% 26% 56% 23% 59% 25% 34% 35% 79% 66% 52% 70% 55% 25% 59% 25% 35% PPIC.ORG Technical Appendices Improving College Pathways in California 9 Female Asian American Black White Hispanic Lowincome First generation Science Passed 9th grade, moved to 10th 77% 70% 67% 68% 65% Passed 10th grade, moved to 11th 68% 75% 65% 68% 61% Passed 11th grade, moved to 12th 43% 55% 37% 42% 37% English Passed 9th grade, moved to 10th 90% 73% 75% 75% 77% Passed 10th grade, moved to 11th 82% 86% 79% 85% 75% Passed 11th grade, moved to 12th 82% 87% 78% 83% 75% TABLE 7 % schools offering each a–g course/sequence, by subject area and school characteristics, 2007–2014 72% 62% 37% 88% 76% 76% 73% 62% 38% 88% 77% 81% Overall High poverty High minority High first generation Math Algebra I (or equivalent) 96% 93% 93% 94% Geometry (or equivalent) 96% 93% 93% 94% Algebra II (or equivalent) 96% 93% 93% 94% Entire math sequence 96% 93% 93% 94% Science Biology (regular or advanced) 96% 93% 93% 94% Chemistry (regular or advanced) 95% 92% 91% 93% Physics (regular or advanced) 90% 86% 86% 86% Entire science sequence 90% 85% 84% 84% English 9th grade 96% 93% 93% 94% 10th grade 96% 93% 93% 94% 11th grade 95% 92% 91% 93% 12th grade 94% 92% 89% 91% Entire English sequence 93% 90% 86% 90% Social Science World history 96% 93% 93% 94% US history 95% 93% 91% 93% Entire social science sequence 95% 93% 91% 93% N high schools 105 59 44 70 SOURCE: Cal-PASS Plus, 2007 – 2014. NOTE: (1). Sample includes all regular high schools. (2).We define high poverty and high African American/Latino minority schools according to NCES definitions. Specifically, high-poverty schools are schools where more than 75%of students are eligible for free/reduced price lunch; and high-minority schools are those where more than 75% of students are African American or Latino. High first generation schools refer to those where more than 75 percent of students are first generation college students, i.e., their parents do not have a bachelor’s degree or higher. PPIC.ORG Technical Appendices Improving College Pathways in California 10 TABLE 8 Probit estimates of the partial effects of student characteristics on course-taking decisions in community college DV: Taking development education course (math/ English) DV: Taking transfer level course (math / English) VARIABLES Overall Female Latino Low income First generation Overall Female Latino Low income First generation Female 0.0166*** 0.0163*** (0.0032) (0.0050) Asian American 0.0393*** 0.0442*** -0.0477 (0.0095) (0.0095) (0.0626) African American 0.0335*** 0.0364*** 0.0213 (0.0082) (0.0072) (0.0409) Latino 0.0243*** 0.0236*** (0.0045) (0.0030) American Indian -0.0194* -0.0430** 0.1037*** (0.0116) (0.0169) (0.0326) Eligible for free/reduced lunch price 0.0204*** 0.0258*** 0.0265*** (0.0022) (0.0032) (0.0041) Parented education: BA or above -0.0029 -0.0028 -0.0125 (0.0029) (0.0039) (0.0097) Pell grant recipients 0.0079*** 0.0052 0.0131* (0.0028) (0.0050) (0.0070) # a–g English courses taken -0.0018* -0.0020** -0.0023** (0.0010) (0.0010) (0.0010) Average grade in a–g English -0.0071*** -0.0120*** -0.0119** (0.0018) (0.0019) (0.0057) Highest math: higher level -0.0463*** -0.0428*** -0.0731*** (0.0116) (0.0143) (0.0183) Highest math: algebra 2 or equivalent -0.0402*** -0.0411*** -0.0496*** 0.0218*** (0.0042) 0.0368*** (0.0066) 0.0345*** (0.0124) 0.0319*** (0.0049) -0.0291 (0.0211) -0.0073 (0.0100) 0.0088 (0.0075) -0.0031** (0.0014) -0.0083*** (0.0028) -0.0573*** (0.0166) -0.0425*** PPIC.ORG 0.0179*** (0.0024) 0.0364*** (0.0052) 0.0306*** (0.0086) 0.0314*** (0.0045) -0.0230*** (0.0085) 0.0230*** (0.0037) 0.0071** (0.0027) -0.0022* (0.0013) -0.0082*** (0.0026) -0.0548*** (0.0143) -0.0451*** -0.0079*** -0.0069* (0.0023) (0.0038) 0.0087** 0.0049 0.0205 (0.0042) (0.0052) (0.0467) -0.0217** -0.0080 0.0025 (0.0103) (0.0178) (0.0231) 0.0019 0.0016 (0.0015) (0.0027) -0.0062 -0.0117 0.0043 (0.0082) (0.0168) (0.0461) -0.0140*** -0.0144*** -0.0144*** (0.0028) (0.0032) (0.0049) 0.0085*** 0.0056** 0.0133*** (0.0018) (0.0022) (0.0030) 0.0089* 0.0071 0.0135** (0.0052) (0.0064) (0.0054) 0.0010 0.0006 -0.0002 (0.0009) (0.0010) (0.0008) 0.0162*** 0.0184*** 0.0147*** (0.0041) (0.0033) (0.0046) 0.0659*** 0.0613*** 0.0692*** (0.0097) (0.0125) (0.0145) 0.0360*** 0.0344*** 0.0386*** -0.0044 (0.0029) 0.0099 (0.0071) -0.0214* (0.0114) 0.0020 (0.0020) -0.0064 (0.0078) 0.0061*** (0.0019) 0.0121** (0.0051) 0.0014 (0.0010) 0.0117*** (0.0033) 0.0649*** (0.0081) 0.0344*** -0.0042** (0.0021) 0.0130 (0.0091) -0.0294* (0.0166) 0.0006 (0.0032) -0.0059 (0.0051) -0.0123*** (0.0037) 0.0101** (0.0047) 0.0010 (0.0012) 0.0163*** (0.0036) 0.0658*** (0.0125) 0.0358*** Technical Appendices Improving College Pathways in California 11 DV: Taking development education course (math/ English) Highest math: geometry or equivalent Grade in highest math course Dual enrollment (ever) (0.0083) -0.0103 (0.0091) -0.0043* (0.0022) 0.0446 (0.0276) (0.0128) -0.0108 (0.0138) -0.0052** (0.0023) 0.0418 (0.0333) (0.0098) -0.0095 (0.0111) -0.0096*** (0.0029) 0.1241 (0.0814) (0.0115) -0.0068 (0.0104) -0.0074** (0.0029) 0.0679 (0.0535) (0.0106) -0.0078 (0.0119) -0.0050** (0.0021) 0.0681 (0.0434) DV: Taking transfer level course (math / English) (0.0056) -0.0064 (0.0067) 0.0094*** (0.0007) -0.0030 (0.0396) (0.0081) -0.0045 (0.0111) 0.0052*** (0.0010) -0.0043 (0.0430) (0.0111) 0.0031 (0.0112) 0.0121*** (0.0020) -0.0221 (0.0356) (0.0035) -0.0040 (0.0053) 0.0097*** (0.0007) -0.0091 (0.0311) (0.0083) 0.0002 (0.0090) 0.0096*** (0.0006) -0.0079 (0.0316) Community college fixed effects X XX X X XXX X X School year fixed effects X XX X X XXX X X School term fixed effects X XX X X XXX X X N students 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 SOURCE: Cal-PASS Plus. NOTE: (1). Standard errors (in parentheses) adjusted for clustering at the community college campus level. (2). *** p<0.01, ** p<0.05, * p<0.1 (3).We did not report the results for African Americans because the sample is too small. PPIC.ORG Technical Appendices Improving College Pathways in California 12 TABLE 9 Probit estimates of the partial effects of student characteristics on passing transfer math/English in community college VARIABLES Overall Female Latino Low income First generation Female Asian American African American Latino American Indian Eligible for free/reduced lunch price Parented education: BA or above Pell grant recipients # a–g English courses taken Average grade in a–g English Highest math: higher level Highest math: algebra 2 or equivalent Highest math: geometry or equivalent Grade in highest math course Dual enrollment (ever) -0.0042*** (0.0016) 0.0125*** (0.0038) -0.0178* (0.0101) 0.0010 (0.0015) -0.0057 (0.0067) -0.0147*** (0.0027) 0.0085*** (0.0010) 0.0084* (0.0044) 0.0011* (0.0006) 0.0288*** (0.0026) 0.0476*** (0.0059) 0.0248*** (0.0064) -0.0051 (0.0072) 0.0081*** (0.0006) 0.0061 (0.0280) 0.0079** (0.0031) -0.0104 (0.0153) 0.0002 (0.0019) -0.0147* (0.0085) -0.0144*** (0.0027) 0.0068*** (0.0015) 0.0062 (0.0053) 0.0005 (0.0004) 0.0294*** (0.0020) 0.0557*** (0.0092) 0.0351*** (0.0082) 0.0030 (0.0114) 0.0058*** (0.0012) 0.0042 (0.0325) -0.0020 (0.0026) 0.0179 (0.0281) 0.0081 (0.0242) -0.0564 (0.0375) -0.0128*** (0.0037) 0.0135*** (0.0033) 0.0153*** (0.0048) 0.0002 (0.0008) 0.0243*** (0.0033) 0.0498*** (0.0075) 0.0286*** (0.0095) 0.0089 (0.0098) 0.0085*** (0.0021) -0.0134 (0.0228) -0.0021 (0.0025) 0.0116* (0.0067) -0.0149 (0.0096) 0.0016 (0.0015) -0.0078 (0.0112) 0.0059*** (0.0014) 0.0114*** (0.0043) 0.0014*** (0.0005) 0.0222*** (0.0015) 0.0478*** (0.0026) 0.0310*** (0.0043) 0.0047 (0.0032) 0.0075*** (0.0006) 0.0046 (0.0233) -0.0018 (0.0017) 0.0162* (0.0087) -0.0165 (0.0161) -0.0004 (0.0028) -0.0073 (0.0093) -0.0130*** (0.0028) 0.0101*** (0.0038) 0.0014* (0.0008) 0.0263*** (0.0023) 0.0481*** (0.0066) 0.0272*** (0.0060) 0.0026 (0.0065) 0.0082*** (0.0007) 0.0039 (0.0207) PPIC.ORG Technical Appendices Improving College Pathways in California 13 VARIABLES Overall Female Latino Low income First generation Community college fixed effects X X X X X School year fixed effects X X X X X School term fixed effects X X X X X N Students 16,792 16,792 16,792 16,792 16,792 SOURCE: Cal-PASS Plus. NOTE: (1). Standard errors (in parentheses) adjusted for clustering at the community college campus level. (2). *** p<0.01, ** p<0.05, * p<0.1 (3). Sample includes 16,792 first-time college freshmen from 2011–2015 school years (4).We did not report the results for African Americans because the sample is too small. FIGURE 1 Likelihood of taking transfer courses based on high school record D e n s ity 0 2 46 8 0 .1 .2 .3 P re d ic te d P ro b a b ility T a k e rs N on-Tak ers Cond SOURCE: Cal-PASS Plus. NOTE: (1). Sample includes 16,792 first-time college freshmen from 2011–2015 school years. (2). Predicted probability is based on a probit model where the dependent variable is whether a student takes transferable course (in a school year/term); and the independent variables include student demographics (gender and race/ethnicity), low-income status (as proxied by free/reduced lunch price eligibility), parental education, disability status, financial aid status, community college fixed effects, term, and school year fixed effects. PPIC.ORG Technical Appendices Improving College Pathways in California 14 TABLE 10 The effects of student characteristics on CSU persistence, units earned, and average GPA Persistence Units Female Asian American African American Latino Eligible for free/reduced price lunch (in high school) Parent education: BA or above GPA in a–g math GPA in a–g science GPA in a–g English GPA in a–g social science GPA in a–g foreign language GPA in a–g art GPA in a–g elective Took a–g math in 12th grade Took a–g science in 12th grade Highest math by 11th grade: higher-level math Model CSU fixed effects -0.0211*** (0.0019) -0.0024 (0.0258) -0.0001 (0.0082) 0.0469*** (0.0174) 0.0142*** (0.0044) 0.0285 (0.0189) 0.0313*** (0.0006) -0.0137 (0.0650) 0.0347 (0.0419) 0.0149* (0.0080) 0.0160 (0.0128) 0.0473** (0.0207) 0.0040 (0.0292) 0.0312*** (0.0037) 0.0315*** (0.0033) 0.0029 (0.0118) Probit X 1.0171 (0.6951) 0.7950 (0.5556) -0.4250 (0.7982) 0.7207 (0.4249) -0.7879 (0.6739) 0.2024 (0.0995) 0.3450 (0.5162) -0.4535 (0.6018) 0.2891 (0.5596) -0.0545 (0.0234) 0.0900 (0.1391) 0.1349 (0.2860) 0.8101 (0.5039) 0.5916* (0.1760) 0.0288 (0.0736) 1.4819 (1.1000) OLS X GPA -0.0051 (0.0571) -0.1651 (0.0838) -0.2644* (0.0833) -0.1231 (0.0557) -0.1936* (0.0568) 0.1178*** (0.0102) 0.1382*** (0.0094) 0.1282 (0.0896) 0.1591** (0.0297) 0.0559 (0.0337) 0.0730*** (0.0043) 0.1480*** (0.0120) 0.0567** (0.0097) 0.1334** (0.0297) 0.0745*** (0.0022) 0.1547** (0.0300) OLS X PPIC.ORG Technical Appendices Improving College Pathways in California 15 Persistence Units GPA School year effects X XX N Students 3,004 3,004 3,004 R-squared 0.169 0.230 NOTE: (1). We did not include a–g course-taking in English or social science because more than 90%of students in our sample took at least one course in each subject in grade 12. (2). The raw gender gap favors females, but reversed after we factor in high school preparation. This is perhaps due to males slightly outperformed females in our sample. For instance, while most females and males completed a higher math course, the share is slightly higher among males. TABLE 11 % schools offering a–g courses, by subject area, 2016–17 Math Science English Social Science All 88% 86% 86% 90% High poverty 86% 86% 86% 89% High minority 87% 87% 86% 89% Small 65% 61% 64% 75% Rural 77% 73% 78% 85% N High Schools 1607 540 577 401 244 SOURCE: California Department of Education, 2016–17. National Center for Education Statistics, 2013–14. NOTE: (1). Sample includes all regular high schools in California. (2).We define high poverty and high African American/Latino minority schools according to NCES definitions. Specifically, high-poverty schools are schools where more than 75 percent of students are eligible for free/reduced price lunch; and high-minority schools are those where more than 75 percent of students are African American or Latino. (3). Small schools are those in the bottom 25th percentile of the enrollment distribution. (4). Rural districts are based on NCES locality measures. PPIC.ORG Technical Appendices Improving College Pathways in California 16 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research. Public Policy Institute of California 500 Washington Street, Suite 600 San Francisco, CA 94111 T: 415.291.4400 F: 415.291.4401 PPIC.ORG PPIC Sacramento Center Senator Office Building 1121 L Street, Suite 801 Sacramento, CA 95814 T: 916.440.1120 F: 916.440.1121" } ["___content":protected]=> string(162) "

Improving College Pathways in California, Technical Appendix

" ["_permalink":protected]=> string(91) "https://www.ppic.org/publication/improving-college-pathways-in-california/1117ngr_appendix/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(12733) ["ID"]=> int(12733) ["post_author"]=> string(1) "4" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-11-27 16:48:52" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(12616) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(60) "Improving College Pathways in California, Technical Appendix" ["post_type"]=> string(10) "attachment" ["slug"]=> string(16) "1117ngr_appendix" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(20) "1117ngr_appendix.pdf" ["wpmf_size"]=> string(6) "292336" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(28477) "Improving College Pathways in California Technical Appendices CONTENTS Appendix A: Pipeline Model Appendix B: Descriptive Summary of Our Sample Appendix C: Additional Figures for Our Pathway Analysis Niu Gao, Hans Johnson 2 5 8 Supported with funding from the Dirk and Charlene Kabcenell Foundation, the Evelyn and Walter Haas, Jr. Fund, the James Irvine Foundation, the Leona M. and Harry B. Helmsley Charitable Trust, and the Sutton Family Fund Appendix A: Pipeline Model Developing a Pipeline Model To estimate where students fall off the pathway to college readiness and completion, we develop a pipeline model that identifies key transitions as students move through their educational trajectories. The pipeline model shown in Figure 1 represents a synthetic cohort of 9th graders in public schools in California as they move through high school, complete the a–g courses, enter college (public or private, in California or another state), transfer from community college, and complete a bachelor’s degree. The primary assumption of the model is that current rates of high school completion, a–g completion, college enrollment, transfer from community college, and college completion prevail throughout the cohort’s high school and college years. Pipelines are developed separately for four large ethnic groups (Latinos, whites, Asian Americans, and African Americans) and two genders (female and male). Four key transitions are identified and estimated in the model: (1) 9th grade to high school graduation (including a–g completion), (2) high school graduation to college enrollment (including community colleges and four-year colleges), (3) enrollment in four-year colleges (either as freshmen or transfer students), and (4) college enrollment to college completion (of a bachelor’s degree). • The first transition is based on cohort graduation rates provided by the California Department of Education for the 2015–16 graduating class. Completion of the a–g courses is based on rates among public high school graduates in 2015–16, also provided by the California Department of Education. • The second transition is based on enrollment rates of recent high school graduates in community colleges and four-year colleges. College enrollment rates are calculated separately by type of college. California State University, the University of California, and the California Community Colleges provided 2015 data on public high school of origin for incoming freshmen. Data on private and out-of-state college enrollment rates was calculated from 2014 IPEDS data, adjusted to reflect public high school graduates. IPEDS data do not include information on ethnicity and gender of students based on their state of residence. Therefore, estimates of enrollment rates to private and out-of-state colleges by ethnicity and gender were derived by applying ethnic and gender distributions of young college-enrolled migrants (students who had left California) based on American Community Survey data. • The third transition is derived from data on 2014–15 transfers from Community Colleges to UC, CSU, private colleges, and out-of-state colleges as provided by institutional research offices of UC, CSU, and the community colleges. • The final transition to college completion is based on six-year graduation rates of incoming freshmen and four-year graduation rates of transfer students. Rates are calculated separately by gender and ethnicity for UC, CSU, private colleges in California, and out-of-state colleges. FIGURE 1 Among those not obtaining a bachelor’s degree, most fall off the pathway before making it to a four-year college Attended a four-year college, but did not earn a degree Attended college, but did not attend a four-year college Finished high school, but did not attend college Did not finish high school 0% 18% 31% 26% 25% 5% 10% 15% 20% 25% 30% 35% Student Pathways to College FIGURE 2 Latino and African American students are underrepresented in the pathways to college 100% 80% 3% 26% 4% 26% 4% 30% 9% 22% 11% 21% 16% 21% % of students 60% 40% 52% 52% 42% 49% 47% 26% 4% 20% 0% 7% 12% Grade 9 enrollment 7% 11% 5% 20% 6% 13% 4% 17% HS graduates a-g completion CCC first-time freshmen CSU first-time freshmen Asian African American Latino White Other 33% UC first-time freshmen SOURCE: Grade 9 enrollment: California Department of Education, 2011–12; a–g completion: California Department of Education, 2015–16, CCC first-time freshmen enrollment: California Community College, 2016; CSU first time freshmen from California high schools: California State University, 2015, UC first time freshmen: University of California, 2016. The need to focus on a–g courses Statewide, at current rates of high school completion and preparation for college, only 30% of 9th graders will earn a high school diploma and complete the a–g college preparatory courses required for entry to UC and CSU. These rates of completion and readiness vary substantially by gender and ethnicity. For example, only 25 percent of African American 9th graders will complete both high school and the a–g courses, compared to 65 percent of Asian Americans. Male 9th graders are less likely to finish high school prepared for college than females (32% and 44% respectively). While high school dropout is an important problem and an obvious reason that some students are not prepared for college, completing the a–g college preparatory courses is an even larger challenge. As shown in Table 1, for every ethnic group and for females and males, the share of high school students who are not ready for college because they did not complete the a–g courses is substantially higher than the share not ready because they did not finish high school. TABLE 1 Most 9th graders will not complete high school ready for college, with large differences by gender and ethnicity Female African American Among the cohort of 9th graders in the class of 2016: Did not finish high school Did not complete the a–g courses Total share not ready for college 21% 47% 68% Latino 16% 48% 63% White 9% 39% 48% Asian American 5% 23% 28% All females 13% 43% 56% Male African American Latino White Asian American All males 33% 24% 14% 8% 20% 48% 52% 46% 33% 48% 81% 77% 60% 41% 68% Appendix B: Descriptive Summary of Our Sample K–12 Sample Our K–12 sample includes 472,324 high school students in 24 diverse districts from 2007 to 2014 school years. Geographically, these districts are well spread across the state (inland and coastal, northern and southern). As seen in Table 1, African American and low-income students are overrepresented in our sample when compared to statewide average. Students in our sample also tend to be low-performing. Note that we also have more other/missing (including inconsistent) values for ethnicity in our dataset. Due to the differences in student characteristics between our sample and statewide averages, our findings can only be applied to similar settings, e.g., schools with a large concentration of low-income and low performing students. This is not necessarily a drawback of our study, as recent reforms (e.g., Every Student Succeeds Act and its predecessor No Child Left Behind) have focused on turning around low-performing schools. TABLE 1 Descriptive summary of key student characterIstics in our k–12 sample Sample State Female 49% 49% Asian American 9% 8% African American 13% 8% Latino 34% 45% White 29% 34% Other/missing 15% 5% Eligible for free/reduced price lunch 64% 50% Parent education: college degree or above 31% - CST: math 278 312 CST: English 315 346 % graduates completing a–g 20% 38% N Graduates 141,307 2,832,012 SOURCE: Cal-PASS Plus, 2007 – 2014; California Department of Education, 2007–2014. Community College Sample We matched the high school graduates in our sample to 18 nearby community colleges and followed them over time. Our analytical sample includes 16,792 first-time college students from 2011–2015. Since we restricted our sample to high school graduates (1) who have complete four-year transcripts (so we can identify their a–g completion status), and (2) who enroll in a nearby community college immediately following high school graduation (due to data limitations), our sample is a very selective one; students in our sample have better outcomes than state averages. For instance, 45 percent of students in our sample take at least one transfer-level course in Math or English, and only a third of students are ever enrolled in a developmental education course during their two years in college (compared to 80 percent statewide within six years). Latino and African American students are underrepresented in our sample, but more students in our sample are first generation college students. A descriptive summary of our community college sample is in Table 2 below. TABLE 2 Descriptive summary of key student characterIstics in our Community college sample Variable Asian American CC statewide 7% 15% African American White Latino Female Pell grant recipients BOGG recipients free/reduced lunch eligible First generation CCC outcomes % taking transfer-level math/English Average passing rate of transfer math/English High school preparation 4% 51% 31% 52% 24% 42% 34% 56% 45% 77% 6% 26% 44% 53% 22% 49% 40% - # a–g math taken # a–g science taken # a–g English taken # a–g social science taken # a–g foreign language taken # a–g art taken 323222- Average GPA in a–g math Average GPA in a–g science Average GPA in a–g English Average GPA in a–g social science Average GPA in a–g foreign language Average GPA in a–g art Highest math taken: higher math Highest math taken: algebra 2 or equivalent Highest math taken: geometry or equivalent 2.3 2.5 2.6 2.7 2.7 3.3 42% 30% 14% - N 16,792 2,355,774 SOURCE: Cal-PASS Plus, 2007–2014; California Community College Data Mart, various years. CSU Sample A descriptive summary of our CSU sample has been summarized in Table 3. Our analytical sample includes 3,004 first-time freshmen students in three CSU campuses (with complete high school transcripts) from 2011 to 2015 school years. Latino students are underrepresented in our sample, and more students in our sample are lowincome and first generation college students. First-year and second-year persistence is lower among our sample, which is not surprising given that most students in our sample are low-income or first generation college students. TABLE 3 Descriptive summary of key student characteristics in our CSU sample Asian American Our Sample 30% CSU system wide (2014) 17% African American White Latino 13% 32% 22% 4% 23% 43% Female Low-income students First generation CSU outcomes First-year persistence Second-year persistence Average year end units earned Average year end GPA High school preparation Average GPA in a–g math Average GPA in a–g science Average GPA in a–g English 60% 64% 68% 85% 57% 36% 77% 61% 21 2.39 84% 76% - 2.71 2.93 3.18 - Average GPA in a–g social studies Average GPA in a–g foreign languages Average GPA in a–g art Average GPA in a–g elective Took a–g math in 12th grade Took a–g science in 12th grade 3.19 3.15 3.64 3.19 67% 54% - Took a–g social science in 12th grade Highest level by 11th grade: higher math 82% 40% - N 3,004 62,523 Source: Cal-PASS Plus, 2011–2015; California State University, 2014. Note: We use students’ eligibility for free/reduced price lunch in high school to determine their low-income status; and CSU uses a different metric—Pell grant recipients—which contributes to the difference in low-income students counts. About 12 percent of students in our CSU sample (not the K–12 sample) miss their free/reduced price lunch information; and about 20 percent of students miss their parental education, which may skew our share of low-income, first generation students. Appendix C: Additional Figures for Our Pathway Analysis TABLE 1 % high school graduates completing each subject area requirements with a C or better, by student characteristics, 2007–2014 Social Science (a) English (b) Math (c ) Science (d) Foreign Language (e) Female 60% 40% 45% 56% 56% Male 54% 32% 39% 51% 46% Asian American 71% 47% 49% 75% 70% African American 44% 23% 31% 41% 36% White 65% 45% 46% 59% 58% Latino 49% 27% 39% 46% 42% Free/reduced price lunch eligible First generation college students 47% 51% 25% 29% 38% 40% 45% 47% 41% 43% Overall 57% 36% 42% 54% 51% SOURCES: Authors’ calculations using Cal-PASS Plus data. TABLE 2 Average passing rate in a–g courses, by subject area and by student characteristics, 2007 –2014 Art (f) Electives (g) 77% 71% 79% 64% 80% 69% 68% 70% 74% 84% 81% 80% 79% 84% 83% 82% 82% 82% First Asian African Low- generation All Female American American Latino White income college Social Science (a) 88% 91% 95% 81% 85% 91% 83% 86% English (b) 87% 90% 95% 80% 83% 91% 81% 84% Math (c) 82% 85% 92% 73% 78% 87% 77% 78% Science (d) 88% 90% 95% 77% 84% 92% 84% 85% Foreign Language (e) 91% 93% 96% 83% 89% 94% 88% 89% Art (f) 93% 95% 97% 87% 91% 95% 89% 91% Electives (g) 89% 92% 95% 82% 86% 93% 85% 87% SOURCE: Authors’ calculations using Cal-PASS Plus data. NOTE: (1). Sample includes 472,324 high school students from 2007 to 2014 school years. (2). Overall passing rate is higher in later grades (e.g., Grades 11 and 12). (3). The same conclusion holds for all subgroups in all subject areas. PPIC.ORG Technical Appendices Improving College Pathways in California 8 TABLE 3 % schools offering a–g courses, by subject areas and by grade, 2007–2014 Math Science English Social Science Grade 9 96% 96% 96% 95% Grade 10 97% 97% 97% 97% Grade 11 97% 96% 96% 96% Grade 12 97% 97% 97% 96% NOTE: (1). Sample includes 105 regular high schools from 2007 to 2014 school years. (2). Social science sequences includes one year of world history, cultures, and historical geography and one year of US history, both of which are typically offered in grades 10 or above. TABLE 4 Math progression, by student characteristics, 2007–2014 Overall Female Male Asian African American American Earned A or B in algebra 1, took geometry Earned A or B in geometry , took algebra 2 Earned A or B in algebra 2, took higher math 69% 68% 59% 70% 68% 58% 69% 68% 59% 72% 73% 66% 66% 54% 48% SOURCE: Authors’ calculations using CalPASS Plus data. NOTE: Sample restricted to schools offering the full a–g math sequence. TABLE 5 Math progression, by pathway and student characteristics, 2007 – 2014 White 64% 67% 57% Latino 68% Lowincome First generation 74% 75% 63% 68% 69% 53% 57% 58% Female Asian African American American Pathway I (start algebra 1 before 9th grade) % algebra 1 completers moving to geometry 82% 91% % geometry completers moving to algebra 2 66% 70% % algebra 2 completers moving to higher level math 52% 58% Pathway II (start algebra 1 in 9th grade) % algebra 1 completers moving to geometry 66% 72% % geometry completers moving to algebra 2 56% 63% % algebra 2 completers moving to higher level math 24% 31% Pathway III (start algebra 1 in 10th grade) % algebra 1 completers moving to geometry 56% 62% % geometry completers moving to higher level math 26% 32% Pathway IV (start algebra 1 in 11/12th grade) % algebra 1 completers moving to geometry 32% 37% NOTE: Sample includes 472,324 high school students from 2007 to 2014 school years. TABLE 6 Science and English pathways, by student characteristics, 2007–2014 74% 59% 40% 64% 49% 21% 57% 26% 34% White 77% 60% 44% 62% 53% 19% 49% 20% 26% Latino Low- First income generation 77% 63% 49% 79% 65% 51% 67% 54% 25% 70% 55% 26% 56% 23% 59% 25% 34% 35% 79% 66% 52% 70% 55% 25% 59% 25% 35% PPIC.ORG Technical Appendices Improving College Pathways in California 9 Female Asian American Black White Hispanic Lowincome First generation Science Passed 9th grade, moved to 10th 77% 70% 67% 68% 65% Passed 10th grade, moved to 11th 68% 75% 65% 68% 61% Passed 11th grade, moved to 12th 43% 55% 37% 42% 37% English Passed 9th grade, moved to 10th 90% 73% 75% 75% 77% Passed 10th grade, moved to 11th 82% 86% 79% 85% 75% Passed 11th grade, moved to 12th 82% 87% 78% 83% 75% TABLE 7 % schools offering each a–g course/sequence, by subject area and school characteristics, 2007–2014 72% 62% 37% 88% 76% 76% 73% 62% 38% 88% 77% 81% Overall High poverty High minority High first generation Math Algebra I (or equivalent) 96% 93% 93% 94% Geometry (or equivalent) 96% 93% 93% 94% Algebra II (or equivalent) 96% 93% 93% 94% Entire math sequence 96% 93% 93% 94% Science Biology (regular or advanced) 96% 93% 93% 94% Chemistry (regular or advanced) 95% 92% 91% 93% Physics (regular or advanced) 90% 86% 86% 86% Entire science sequence 90% 85% 84% 84% English 9th grade 96% 93% 93% 94% 10th grade 96% 93% 93% 94% 11th grade 95% 92% 91% 93% 12th grade 94% 92% 89% 91% Entire English sequence 93% 90% 86% 90% Social Science World history 96% 93% 93% 94% US history 95% 93% 91% 93% Entire social science sequence 95% 93% 91% 93% N high schools 105 59 44 70 SOURCE: Cal-PASS Plus, 2007 – 2014. NOTE: (1). Sample includes all regular high schools. (2).We define high poverty and high African American/Latino minority schools according to NCES definitions. Specifically, high-poverty schools are schools where more than 75%of students are eligible for free/reduced price lunch; and high-minority schools are those where more than 75% of students are African American or Latino. High first generation schools refer to those where more than 75 percent of students are first generation college students, i.e., their parents do not have a bachelor’s degree or higher. PPIC.ORG Technical Appendices Improving College Pathways in California 10 TABLE 8 Probit estimates of the partial effects of student characteristics on course-taking decisions in community college DV: Taking development education course (math/ English) DV: Taking transfer level course (math / English) VARIABLES Overall Female Latino Low income First generation Overall Female Latino Low income First generation Female 0.0166*** 0.0163*** (0.0032) (0.0050) Asian American 0.0393*** 0.0442*** -0.0477 (0.0095) (0.0095) (0.0626) African American 0.0335*** 0.0364*** 0.0213 (0.0082) (0.0072) (0.0409) Latino 0.0243*** 0.0236*** (0.0045) (0.0030) American Indian -0.0194* -0.0430** 0.1037*** (0.0116) (0.0169) (0.0326) Eligible for free/reduced lunch price 0.0204*** 0.0258*** 0.0265*** (0.0022) (0.0032) (0.0041) Parented education: BA or above -0.0029 -0.0028 -0.0125 (0.0029) (0.0039) (0.0097) Pell grant recipients 0.0079*** 0.0052 0.0131* (0.0028) (0.0050) (0.0070) # a–g English courses taken -0.0018* -0.0020** -0.0023** (0.0010) (0.0010) (0.0010) Average grade in a–g English -0.0071*** -0.0120*** -0.0119** (0.0018) (0.0019) (0.0057) Highest math: higher level -0.0463*** -0.0428*** -0.0731*** (0.0116) (0.0143) (0.0183) Highest math: algebra 2 or equivalent -0.0402*** -0.0411*** -0.0496*** 0.0218*** (0.0042) 0.0368*** (0.0066) 0.0345*** (0.0124) 0.0319*** (0.0049) -0.0291 (0.0211) -0.0073 (0.0100) 0.0088 (0.0075) -0.0031** (0.0014) -0.0083*** (0.0028) -0.0573*** (0.0166) -0.0425*** PPIC.ORG 0.0179*** (0.0024) 0.0364*** (0.0052) 0.0306*** (0.0086) 0.0314*** (0.0045) -0.0230*** (0.0085) 0.0230*** (0.0037) 0.0071** (0.0027) -0.0022* (0.0013) -0.0082*** (0.0026) -0.0548*** (0.0143) -0.0451*** -0.0079*** -0.0069* (0.0023) (0.0038) 0.0087** 0.0049 0.0205 (0.0042) (0.0052) (0.0467) -0.0217** -0.0080 0.0025 (0.0103) (0.0178) (0.0231) 0.0019 0.0016 (0.0015) (0.0027) -0.0062 -0.0117 0.0043 (0.0082) (0.0168) (0.0461) -0.0140*** -0.0144*** -0.0144*** (0.0028) (0.0032) (0.0049) 0.0085*** 0.0056** 0.0133*** (0.0018) (0.0022) (0.0030) 0.0089* 0.0071 0.0135** (0.0052) (0.0064) (0.0054) 0.0010 0.0006 -0.0002 (0.0009) (0.0010) (0.0008) 0.0162*** 0.0184*** 0.0147*** (0.0041) (0.0033) (0.0046) 0.0659*** 0.0613*** 0.0692*** (0.0097) (0.0125) (0.0145) 0.0360*** 0.0344*** 0.0386*** -0.0044 (0.0029) 0.0099 (0.0071) -0.0214* (0.0114) 0.0020 (0.0020) -0.0064 (0.0078) 0.0061*** (0.0019) 0.0121** (0.0051) 0.0014 (0.0010) 0.0117*** (0.0033) 0.0649*** (0.0081) 0.0344*** -0.0042** (0.0021) 0.0130 (0.0091) -0.0294* (0.0166) 0.0006 (0.0032) -0.0059 (0.0051) -0.0123*** (0.0037) 0.0101** (0.0047) 0.0010 (0.0012) 0.0163*** (0.0036) 0.0658*** (0.0125) 0.0358*** Technical Appendices Improving College Pathways in California 11 DV: Taking development education course (math/ English) Highest math: geometry or equivalent Grade in highest math course Dual enrollment (ever) (0.0083) -0.0103 (0.0091) -0.0043* (0.0022) 0.0446 (0.0276) (0.0128) -0.0108 (0.0138) -0.0052** (0.0023) 0.0418 (0.0333) (0.0098) -0.0095 (0.0111) -0.0096*** (0.0029) 0.1241 (0.0814) (0.0115) -0.0068 (0.0104) -0.0074** (0.0029) 0.0679 (0.0535) (0.0106) -0.0078 (0.0119) -0.0050** (0.0021) 0.0681 (0.0434) DV: Taking transfer level course (math / English) (0.0056) -0.0064 (0.0067) 0.0094*** (0.0007) -0.0030 (0.0396) (0.0081) -0.0045 (0.0111) 0.0052*** (0.0010) -0.0043 (0.0430) (0.0111) 0.0031 (0.0112) 0.0121*** (0.0020) -0.0221 (0.0356) (0.0035) -0.0040 (0.0053) 0.0097*** (0.0007) -0.0091 (0.0311) (0.0083) 0.0002 (0.0090) 0.0096*** (0.0006) -0.0079 (0.0316) Community college fixed effects X XX X X XXX X X School year fixed effects X XX X X XXX X X School term fixed effects X XX X X XXX X X N students 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 16,792 SOURCE: Cal-PASS Plus. NOTE: (1). Standard errors (in parentheses) adjusted for clustering at the community college campus level. (2). *** p<0.01, ** p<0.05, * p<0.1 (3).We did not report the results for African Americans because the sample is too small. PPIC.ORG Technical Appendices Improving College Pathways in California 12 TABLE 9 Probit estimates of the partial effects of student characteristics on passing transfer math/English in community college VARIABLES Overall Female Latino Low income First generation Female Asian American African American Latino American Indian Eligible for free/reduced lunch price Parented education: BA or above Pell grant recipients # a–g English courses taken Average grade in a–g English Highest math: higher level Highest math: algebra 2 or equivalent Highest math: geometry or equivalent Grade in highest math course Dual enrollment (ever) -0.0042*** (0.0016) 0.0125*** (0.0038) -0.0178* (0.0101) 0.0010 (0.0015) -0.0057 (0.0067) -0.0147*** (0.0027) 0.0085*** (0.0010) 0.0084* (0.0044) 0.0011* (0.0006) 0.0288*** (0.0026) 0.0476*** (0.0059) 0.0248*** (0.0064) -0.0051 (0.0072) 0.0081*** (0.0006) 0.0061 (0.0280) 0.0079** (0.0031) -0.0104 (0.0153) 0.0002 (0.0019) -0.0147* (0.0085) -0.0144*** (0.0027) 0.0068*** (0.0015) 0.0062 (0.0053) 0.0005 (0.0004) 0.0294*** (0.0020) 0.0557*** (0.0092) 0.0351*** (0.0082) 0.0030 (0.0114) 0.0058*** (0.0012) 0.0042 (0.0325) -0.0020 (0.0026) 0.0179 (0.0281) 0.0081 (0.0242) -0.0564 (0.0375) -0.0128*** (0.0037) 0.0135*** (0.0033) 0.0153*** (0.0048) 0.0002 (0.0008) 0.0243*** (0.0033) 0.0498*** (0.0075) 0.0286*** (0.0095) 0.0089 (0.0098) 0.0085*** (0.0021) -0.0134 (0.0228) -0.0021 (0.0025) 0.0116* (0.0067) -0.0149 (0.0096) 0.0016 (0.0015) -0.0078 (0.0112) 0.0059*** (0.0014) 0.0114*** (0.0043) 0.0014*** (0.0005) 0.0222*** (0.0015) 0.0478*** (0.0026) 0.0310*** (0.0043) 0.0047 (0.0032) 0.0075*** (0.0006) 0.0046 (0.0233) -0.0018 (0.0017) 0.0162* (0.0087) -0.0165 (0.0161) -0.0004 (0.0028) -0.0073 (0.0093) -0.0130*** (0.0028) 0.0101*** (0.0038) 0.0014* (0.0008) 0.0263*** (0.0023) 0.0481*** (0.0066) 0.0272*** (0.0060) 0.0026 (0.0065) 0.0082*** (0.0007) 0.0039 (0.0207) PPIC.ORG Technical Appendices Improving College Pathways in California 13 VARIABLES Overall Female Latino Low income First generation Community college fixed effects X X X X X School year fixed effects X X X X X School term fixed effects X X X X X N Students 16,792 16,792 16,792 16,792 16,792 SOURCE: Cal-PASS Plus. NOTE: (1). Standard errors (in parentheses) adjusted for clustering at the community college campus level. (2). *** p<0.01, ** p<0.05, * p<0.1 (3). Sample includes 16,792 first-time college freshmen from 2011–2015 school years (4).We did not report the results for African Americans because the sample is too small. FIGURE 1 Likelihood of taking transfer courses based on high school record D e n s ity 0 2 46 8 0 .1 .2 .3 P re d ic te d P ro b a b ility T a k e rs N on-Tak ers Cond SOURCE: Cal-PASS Plus. NOTE: (1). Sample includes 16,792 first-time college freshmen from 2011–2015 school years. (2). Predicted probability is based on a probit model where the dependent variable is whether a student takes transferable course (in a school year/term); and the independent variables include student demographics (gender and race/ethnicity), low-income status (as proxied by free/reduced lunch price eligibility), parental education, disability status, financial aid status, community college fixed effects, term, and school year fixed effects. PPIC.ORG Technical Appendices Improving College Pathways in California 14 TABLE 10 The effects of student characteristics on CSU persistence, units earned, and average GPA Persistence Units Female Asian American African American Latino Eligible for free/reduced price lunch (in high school) Parent education: BA or above GPA in a–g math GPA in a–g science GPA in a–g English GPA in a–g social science GPA in a–g foreign language GPA in a–g art GPA in a–g elective Took a–g math in 12th grade Took a–g science in 12th grade Highest math by 11th grade: higher-level math Model CSU fixed effects -0.0211*** (0.0019) -0.0024 (0.0258) -0.0001 (0.0082) 0.0469*** (0.0174) 0.0142*** (0.0044) 0.0285 (0.0189) 0.0313*** (0.0006) -0.0137 (0.0650) 0.0347 (0.0419) 0.0149* (0.0080) 0.0160 (0.0128) 0.0473** (0.0207) 0.0040 (0.0292) 0.0312*** (0.0037) 0.0315*** (0.0033) 0.0029 (0.0118) Probit X 1.0171 (0.6951) 0.7950 (0.5556) -0.4250 (0.7982) 0.7207 (0.4249) -0.7879 (0.6739) 0.2024 (0.0995) 0.3450 (0.5162) -0.4535 (0.6018) 0.2891 (0.5596) -0.0545 (0.0234) 0.0900 (0.1391) 0.1349 (0.2860) 0.8101 (0.5039) 0.5916* (0.1760) 0.0288 (0.0736) 1.4819 (1.1000) OLS X GPA -0.0051 (0.0571) -0.1651 (0.0838) -0.2644* (0.0833) -0.1231 (0.0557) -0.1936* (0.0568) 0.1178*** (0.0102) 0.1382*** (0.0094) 0.1282 (0.0896) 0.1591** (0.0297) 0.0559 (0.0337) 0.0730*** (0.0043) 0.1480*** (0.0120) 0.0567** (0.0097) 0.1334** (0.0297) 0.0745*** (0.0022) 0.1547** (0.0300) OLS X PPIC.ORG Technical Appendices Improving College Pathways in California 15 Persistence Units GPA School year effects X XX N Students 3,004 3,004 3,004 R-squared 0.169 0.230 NOTE: (1). We did not include a–g course-taking in English or social science because more than 90%of students in our sample took at least one course in each subject in grade 12. (2). The raw gender gap favors females, but reversed after we factor in high school preparation. This is perhaps due to males slightly outperformed females in our sample. For instance, while most females and males completed a higher math course, the share is slightly higher among males. TABLE 11 % schools offering a–g courses, by subject area, 2016–17 Math Science English Social Science All 88% 86% 86% 90% High poverty 86% 86% 86% 89% High minority 87% 87% 86% 89% Small 65% 61% 64% 75% Rural 77% 73% 78% 85% N High Schools 1607 540 577 401 244 SOURCE: California Department of Education, 2016–17. National Center for Education Statistics, 2013–14. NOTE: (1). Sample includes all regular high schools in California. (2).We define high poverty and high African American/Latino minority schools according to NCES definitions. Specifically, high-poverty schools are schools where more than 75 percent of students are eligible for free/reduced price lunch; and high-minority schools are those where more than 75 percent of students are African American or Latino. (3). Small schools are those in the bottom 25th percentile of the enrollment distribution. (4). Rural districts are based on NCES locality measures. PPIC.ORG Technical Appendices Improving College Pathways in California 16 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research. Public Policy Institute of California 500 Washington Street, Suite 600 San Francisco, CA 94111 T: 415.291.4400 F: 415.291.4401 PPIC.ORG PPIC Sacramento Center Senator Office Building 1121 L Street, Suite 801 Sacramento, CA 95814 T: 916.440.1120 F: 916.440.1121" ["post_date_gmt"]=> string(19) "2017-11-28 00:48:52" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(16) "1117ngr_appendix" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-11-27 16:49:20" ["post_modified_gmt"]=> string(19) "2017-11-28 00:49:20" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(59) "http://www.ppic.org/wp-content/uploads/1117ngr_appendix.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }