Professional Documents
Culture Documents
The purpose of this study was to examine how a student’s first-year success, measure
with first-year GPA and retention, is impacted by participation in Advanced Placement
coursework, with considerations of socioeconomic background. The population for this
study included cohorts of first-year, full-time students with freshman class standing
from 2011-2017. Hypotheses presented assess the statistical significance between
Advanced Placement and Zero-Credit Start students for incoming academic profile, first-
term GPA, first-year GPA and first-year retention. Additionally, a multiple regression
model is constructed for both student populations, and controlled for socioeconomic
factors to assess the impact of socioeconomic characteristics on first-year success
variance. Findings show that AP students have consistently higher-achievement across
academic factors, with limited impact on retention. Regression models show that
academic factors account for a majority of first-year success variance, with the Zero-
Credit Start students experiencing an overall lower performance with larger variance
with socioeconomic factors. Limitations and opportunities for future research are
discussed.
Andrew Clum
SOUTHERN OREGON UNIVERSITY
i
Running Head:
Are they worth their weight? The impact of Advanced Placement courses on first-year success
Andrew Clum
1|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
Table of Contents
List of Tables …………………………………………………………………………………..... ii
List of Charts .…………………………………………………………………………………... iii
List of Equations ……………………………………………………………………………….. iv
Chapter 1: Introduction…………………………………………………………………………... 1
Chapter 2: Review of Related Literature ………………………………………………………... 7
Chapter 3: Methods …………………………………………………………………………….. 13
Chapter 4: Results …………………………………………………………………………….... 18
Chapter 5: Discussion of Findings……………………………………………………………… 32
References ……………………………………………………………………………………… 39
i
THE IMPACT OF ADVANCED PLACEMENT COURSES
List of Tables
Table 4.3 Year-One to Year-Two (Fall to Fall) retention rates for AP and ZCS Students…………….. 24
Table 4.4 Year-Two to Year-Three (Fall to Fall) retention rates for AP and ZCS Students ………….. 24
Table 4.5 Hypothesis testing for incoming GPA for AP and ZCS Students …………………. ……… 24
Table 4.6 Hypothesis testing for converted SAT scores for AP and ZCS students …………................ 24
Table 4.7 2-Sample test for Variance for incoming GPA and converted SAT composite score for AP and
Table 4.8: Summary of hypothesis testing for the mean first-term academic performance of AP and ZCS
students ……………………………………………………………………………………………….. 27
Table 4.9 Summary of hypothesis test for variability of first-term performance of AP and ZCS
students………………………………………………………………………………………………….. 27
Table 4.10: Summary of hypothesis testing for the mean first-year academic performance of AP and ZCS
students ……………………………………………………………………………………………..…. 28
Table 4.11 Summary of hypothesis test for variability of first-year performance of AP and ZCS students
…………………………………………………………………………………………………………... 28
Table 4.12 Summary of regression coefficients for first-year GPA of AP students, including
socioeconomic factors………………………………………………………………………..………….. 29
Table 4.13 Summary of regression coefficients for first-year GPA of ZCS students, including
socioeconomic factors……………………………………………………………………………..…….. 31
Table 4.15: Descriptive Summary of AP and ZCS students by controlling for ethnicity and first-
ii
THE IMPACT OF ADVANCED PLACEMENT COURSES
List of Charts
Chart 4.4 Scatterplot for AP students of predicted values of gpa_y1 with regression………. 30
Chart 4.5 Scatterplot for ZCS students of predicted values of gpa_y1 with regression using all
variables, excluding academic index ………………………………………………………… 33
Chart 4.6 Scatterplot for ZCS students of predicted values of gpa_y1 with regression using only
academic index ………………………………………………………………………………. 33
iii
List of Equations
Equation 4.1: Hypothesis for mean values of “incoming_gpa” for AP and ZCS students…...... 24
Equation 4.2: Hypothesis for mean values of maxc for AP and ZCS students ……………....... 24
Equation 4.3: Hypothesis test for variability of “incoming_gpa” for AP and ZCS students…... 25
Equation 4.4: Hypothesis test for variability of maxc for AP and ZCS students………………. 25
Equation 4.5: Hypothesis test for mean values of first-term academic performance in AP and
ZCS students………………………………………………………………………………….... 26
Equation 4.6: Hypothesis test for variability of first-term academic performance in AP and ZCS
students......................................................................................................................................... 26
Equation 4.7: Hypothesis test for mean values of first-year academic performance in AP and
ZCS students................................................................................................................................. 27
Equation 4.8: Hypothesis test for variability of first-year academic performance in AP and ZCS
students......................................................................................................................................... 27
Equation 4.9: Hypothesis testing of proportions for f2_ret for AP and ZCS students................. 28
Equation 4.10: Hypothesis testing of proportions for f3_ret for AP and ZCS students............... 28
iv
THE IMPACT OF ADVANCED PLACEMENT COURSES
Chapter 1: Introduction
persistence and completion despite their growing numbers in higher education. Additionally, even for
the respectively small percentage of low-income families, when compared to more advantaged
populations, the likelihood of degree attainment is low (Engle & Tinto, 2008; Cataldi, Bennet & Chen,
2018; Redford & Hoyer, 2017). However, an interesting shift in the recent years of research are showing
that low-income students are enrolling at a higher rate than middle-income students, as defined by the
National Center for Education Statistics (NCES). While only a short time ago in 2012, low-income
students were enrolling at a rate nearly 15% lower than that of middle-income students, in 2016 low-
income students showed a significant climb, overtaking the middle-income students, who remained
relatively stagnant. (NCES, 2018a). Despite this increase in enrollment, and while trends suggest that
low-income enrollment will increase, they face higher attrition rates than their peers, and commonly
struggle with familial and financial obligations when attempting to finish their coursework, often taking
lighter course loads that their peers (Engle & Tinto, 2008). If these hardships weren’t enough already,
students as a whole, face a question of college preparation overall. National assessments, collected
through the NCES, have shown either stagnation or decline in areas of reading, mathematics and science
for 12th grade students (NCES, 2018b). The point still remains however, if more low-income students
are making the journey into higher education, why are they not persisting or completing degrees?
Colleges and universities have long studied the science of retention through numerous research
articles attempting to analyze trends to predict the likelihood of students returning to the same campus
the following year, with the goal being to confer a degree. Researchers often try to draw links between
specific traits or populations and overall retention. These can range from best curricular programs
2|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
(Adelman, 2006), to the impact of socioeconomic and first-generation status (Engle & Tinto, 2008;
Lohman & Jarvis, 2000; Vuong et al., 2010) and ethnic inequity (Walton & Cohen, 2007). Tinto (2017)
shared the view that many theories revolving around student retention and continuation take a slant from
the university lens, with a large focus being to increase the bottom line; however, he also sheds light on
the importance of student experience, where the term persistence truly takes form. It is a strange idea for
a student to think of themselves as retaining when they are indeed persisting, that is to overcome
obstacles, both personal and academic, to achieve progress to the next mark in their respective
program(s). To understand student success in a challenging academic environment, there must be a clear
Persistence takes on an active form for students, and represents a sense of motivation. This belief
a student may have to overcome academic obstacles, despite fears or shortcomings for some student
populations, is a form of courage on themselves. It would be natural to expect a university to ask how
they can foster this motivation students have. Students have many different forces which assist in
creating who they will become, and to best prepare students to curb the effects of adversity, a focus
deals with an individual’s belief they can succeed in a task at-hand (Bandura & Estes, 1977). It is no
surprise that students with high self-efficacy can persist in challenging tasks, and see them to
completion, whereas those with low self-efficacy can find themselves easily discouraged (Chemers et al.
2001; Vuong et al., 2010). While universities may offer an array of student support services, this does
not mean a struggling student will opt to partake in asking for help.
It is no secret, there exist students who struggle and find intrinsic reasons to not ask for help,
whether it be fear of being the only one, not being able to make it, or something else. This concept of
self-efficacy and confidence can be positively influenced with focusing on student engagement. Tinto
3|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
(2017) shares that sense of belonging can be another key area for assisting students. This non-cognitive
trait can be expressed as a way that students feels bound, in a way, to a group or a community in both
encouraging and adverse times. When students feel a connection to a university, they can persist, and
these connections can help provide motivation to connect with others, and in turn help more students.
When a sense of belonging is weak for a student, they can become withdrawn, both in-and-out of the
classroom, and undermine student performance (Walton & Cohen, 2007). The postsecondary education
landscape has long been riddled with challenges that both students and campus professionals must face,
and ultimately work through together. Students can often face and be forced to overcome risk factors
Students who encounter these factors in higher education are often referred to as at-risk, and can
have to overcome academic barriers and stopping out, or dropping out, of higher education, according to
The Glossary of Higher Education Reform (2018). There are a multitude of risk factors facing collegiate
students, especially during their first year. Some common examples of these risk factors are being
secondary knowledge, and not understanding the responsibilities of financial independence, though this
list is not exhaustive. For a student to be ready to succeed in a post-secondary environment, they must
also have other securities in place: financially, emotionally, and physically though physical well-being
won’t be a focus in this study. One of the well-studied areas of risk factors for students is the
income family is home to a first-generation student. For the National Center for Education Statistics
(NCES), there are several key risk factors which impinge on a student’s likelihood to succeed, especially
those which are more prominent in low-income or first-generation families: attending part-time, being
4|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
financially independent from family, having children or being a single parent (Engle & Tinto, 2008). It
is easy to link the before mentioned risk factors to financial needs and socioeconomic characteristics for
students. The financial needs for students have shown to be a constant concern for students, with the
cost for 4-year public institutions climbing higher every year, while financial aid distributions cover
about one-third to one-half of the cost of tuition for low- and middle-income families (NCES, 2018c).
In addition to these attributes, a student must also have an academic background to be successful in a
collegiate environment, altogether creating a pathway to the baseline goal of higher education,
completing a degree.
Adelman (2006), in his seminal paper on national higher education, examined the key factors
that lead to the attainment of a bachelor’s degree, but also tried to capture an idea of academic rigor an
preparation in its relations to college success, across a wide array of demographics. First-year academic
performance was shown to be a key factor in persistence for students in terms of their likelihood to
complete a bachelor’s degree (Adelman, 2006). Programs that promote advanced standing for high
school students have shown positive impacts on the first-year performance of students in a variety of
college settings (Ewing & Howell, 2015; North & Jacobs, 2010; Adelman, 2006). Because of this,
students are constantly encouraged to participate in these programs, and participation in Advanced
Placement (AP) coursework showed a positive correlation with student pursuit and transition into
colleges. In the years since Adelman’s 2006 report, there have been millions of AP exams given each
year nationwide, with over five-million tests being given about three-million students in the 2018
(CollegeBoard, 2017).
One of the identifiable factors contributing to collegiate success of students is their preparation
which provide education comparable to a collegiate environment, is common among public and private
5|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
secondary schools across the nation. Collegiate preparatory coursework is a venue by which students
can help themselves prepare for the transition to a collegiate environment. Chen (2012), showed that
consistent with earlier works(Tinto 1987, 1992), and Pascarella and Terenzini’s (2005) review, there is a
strong relationship between academic preparation and college experience and predicting collegiate drop-
out. Though this study is limited by the parameters of the data not tracking students who successfully
complete after transferring, this paper seems to be the first of its kind in its modeling. Nevertheless,
research shows a strong connection between secondary education rigor and postsecondary success,
resulting in programmatic partnerships across the country between secondary institutions, colleges, and
universities to foster academic preparation. Often times, these programs offer accelerated curriculums
through courses like AP, or by utilizing these courses to partner with local institutions to receive college
credit.
With the rising cost of attendance in public and private collegiate education in the United States,
students have to struggle to create a sense of financial security, not just for attendance, but living as well
(CollegeBoard, 2017; Silva et al. 2015). Though many college students have resources and support, both
inside, and outside of the classroom, it requires a campus-wide approach to address the issues which
impact students at the foundational level, rather than case-by-case solutions. Institutions of higher
education, as a whole, are often questioned in terms of their ability to prepare students for the
expectations of a new professional world, in a globalized economy. Within higher education, there is an
abundance of resources in the importance of the first year a student attends and ultimately their degree
complete college credit, or college equivalent courses, before attending a postsecondary institution.
Students will often have an added benefit of saving money by earning college credits through
6|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
accelerated coursework, such as Advanced Placement. Though there exist other credit-by-examination
For high school students, accelerated programs which offer an opportunity for college credit
have gained popularity, especially in regards to AP coursework (Suldo et al., 2018). Given its growing
popularity, many states have begun linking policy initiatives to accelerated coursework, such as AP, in
order to address a needed avenue for increased academic rigor (Spalding, Eden, & Heppner, 2012;
McBride Davis, Slate, Moore & Barnes, 2015). Policy development has gained attention in the area of
high school education, with 283 state bills being introduced in the last year relating to high school
education, with over half being directly related to dual or concurrent enrollment, AP, and college
While the AP program is studied often, with its incredible reach in the United States, the
program has not often been analyzed on an institutional level, especially with regard to a regional
university. In regards to dual credit, there is an inherent difficulty in separating out true institutional
courses versus the courses which students are awarded credit for taking an advanced course (such as AP)
in their secondary curriculum, for that reason dual credit will not be studied explicitly. With states
focusing on participation in these advanced programs, there remains a need to clarify where resources
should be placed, and if students in these accelerated courses need additional resources, and if these
Though AP tests remain incredibly popular with high school students in the United States, there
has been little-to-no research involving the progression of students at Southern Oregon University
(SOU). Many students of SOU enter with college-level credit, though there is no unique tracking of
first-time full-time students at SOU in comparison to a group of students who have no college-level
credit, or in the scope of this paper, traditional first-year students. The goal of this study will be to
7|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
assess how the amount of credit earned by incoming students will impact their first-year success, as
Many students and educators know the AP program of today to be one of the prominent “children” of
CollegeBoard, who also administers the SAT exam, but the first iteration of the program was not
actually property of the CollegeBoard. The Advanced Placement program originates from America,
specifically near and following the conflicts of the Cold War and Korean War in the 1950s wherein the
public urged for the upgrading of education in order to combat the spread of communism (Rothschild,
1999). In the early 1950s, the Ford Foundation heard the cry of the public and began creating the Fund
for the Advancement of Education (FAE). Following the formation of the FAE, the history of Advanced
Placement becomes somewhat clouded, as described by Rothschild (1999). It isn’t clear if there were
crucial educational theories that existed at the core of the AP program, or if it simply out of a perceived
public need to accelerate gifted students. It seems clear that in 1951, 11 university presidents worked
with administrators from secondary schools to create desirable accelerated courses in an effort to allow
quick advancement for a high school student, targeting students at the age of 16, roughly sophomores or
juniors. There is a disagreement about when CollegeBoard officially assumed administrative control of
the program, but it is clear that under the management of CollegeBoard, Advanced Placement grew into
As the program grew out of the creative university administrators who developed it, AP spread across
the United States fiercely. Additionally, with little competition, as IB didn’t arrive in the US until the
1970s, the AP program had no market challenge. College professors are the primary source of course
content for the AP curriculum, and throughout much of its lifespan, the AP courses have been criticized
8|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
for having a broad scope, and little depth in areas (Tierney, 2012; Drew, 2011). Through its widespread
growth, there has been a back-and-forth argument regarding the definition of the “quality” of advanced
placement programs, ultimately resulting in researchers arguing that there was blatant manipulation of
data to favor arguments (Camara et al., 2000; Licthen, 2000). Being a widely accessible, the AP
program is often a subject of many researchers’ projects. The abundance of available data brings the
reports filed by CollegeBoard under scrutiny for their claims of financial benefits, accuracy through
omission, such as reports claiming success by multiple-choice scores only and failing to include the
written portion of exams in reported scores (Licthen 2000, 2007, 2010; Klopfenstein, 2010; Lacy, 2010).
Even in the face of such scrutiny, this has not stopped the program from growing from a handful of
As of 2019, Advanced Placement is comprised of 38 unique subject exams offered to high school
students. In its origination, the AP program was meant to merely incentivize college placement, but it
has since turned into a factor in the college admissions process, wherein 60% on colleges and
universities have policies allowing college credit equivalency for strong exam scores (Byrd et al., 2007).
Unlike its main competitor IB, the AP program boasts millions of test takers each year, and has for
several years. Over time, the exam has grown more inclusive, showing a quintupled increase in low-
income participants from 2003-2013, and along with that growth, there remains a consistent, sizeable
pool of students, over 40%, who are not able to achieve a score of 3 (scale of 1-5) which is considered
the minimum score for passing, or the academic equivalent of a C. Unfortunately, the number of
students who are scoring below 3 is increasing each year, growing by 2% overall from 2006-2016
which represent just over 900,000 students in 2006 and nearly 2MM in 2016 (CollegeBoard, 2014,
2017).
9|Page
THE IMPACT OF ADVANCED PLACEMENT COURSES
One may naturally assume that if a student is accelerated academically, they may be more
prepared for challenging courses, but often times, these students need support. Accelerated course
participants are in need of crucial academic and emotional support, similar to the general student body
(Suldo, et al., 2018). Furthermore, knowing that students with advanced curriculum may not be able to
cope with stress as effectively as the general student body, institutions can exercise extra precaution
when considering student services. With Chen’s (2012) model considering both student and
institutional-level concerns, he is also able to show contrast between factors which impact students, and
factors which impact the institution, but all factors relate directly to student persistence. While there are
financial barriers for students, there stand financial hardships for universities and the possible
misallocation of institutional funds. Millea et al. (2018) suggested that institutions should invest in
smaller class sizes, Chen (2012) found that institutional funds towards instruction are less value than
Student services and class size can be considered, but first there must be an understanding of how these
accelerated students are received by their host institutions. While AP courses are accepted by a wide
array of schools, there remains a question of if students are truly ready for a next-level course following
credit received for an AP course (e.g. taking BIO 102 after getting credit for BIO 101). There are
millions of AP exams taken by millions of students each year, and of this population there are many
students who will not pass the exam with scores sufficient to receive credit; however, there is a sense of
consistency to how AP scores are received and credited at different institutions. To understand the
differences in how these scores are interpreted and impact student’s academic efficacy, it is necessary to
10 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
AP examinations and their predictive validity and have been well-studied and have shown some
sort of positive correlation, either weak or strong, when associated with college attendance, completion,
and overall performance (Patterson & Ewing, 2013; Shaw, Mattern, & Marini, 2012; Morgan & Klaric,
2007). However, even with AP testing having a positive impact in the broad population of students,
exam performance often doesn’t highlight some of the disadvantages experienced by certain
demographics, or even geographical location. Rural school districts often struggle to offer AP
examinations, and Gagnon and Mattingly (2016) reveal that only a slight majority of rural districts enroll
AP students, which falls significantly below towns, suburban areas, and urban areas. Additionally, even
when rural districts can offer AP courses, there is still a 30% difference in enrollment between rural
areas and towns than suburban and urban areas. Suburban and urban areas showed significantly higher
enrollment and completion rates than rural areas and towns, with these outcomes amplified by poverty
With more students taking courses in AP curriculum, there has been a growing focus
surrounding the efficacy of students who participate in these courses, and if the outcome is worth the
work. Adelman (2006) suggested that participation in accelerated courses (AP) was a key indicator of
post-secondary success, even though more recent studies have argued that rather than participation, the
end-of-course exam was a stronger indication of success (Ackerman, Kanfer & Calderwood, 2013). A
recent study asserts that a focus on mere academic completion is too narrow a focus when considering
student performance. Research should include factors such as emotional well-being through quality of
life and mental health, especially considering that students who do participate in accelerated courses are
generally more stressed than non-participants (Sulda et al., 2018, Sulda et al., 2009). There are many
students who participate in accelerated coursework who do not fit the “profile” of gifted students.
Students who participate in these courses can have GPAs below 3.0, low emotional well-being, mental
11 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
health concerns, and may not earn a passing grade. This would contradict the assumption that
As the AP program has grown into a nearly standard piece of curriculum in school districts who
can afford it, the program has seldom addressed the consistent decrease in mean score for test-takers
over time, showing the mean national score for tests hasn’t been at a “3” since 2002 (CollegeBoard,
2017). This would seem to give weight to the controversial claim made by Lichten (2000, 2007, 2010)
that as the program has grown, the quality of the program has decreased. As AP program spread to over
21,000 schools in the US, the Senior Vice President for AP Trevor Packer, made a point to become more
inclusive of low-income families, growing the population of low-income test takers from just below
10% in 2003 to nearly 20% in 2018 (Mathews, 2018). Though this growth is admirable, as the
population grows and time moves forward, there becomes a need to revisit the curriculum required for
the courses. In the years leading up to 2013, the AP program has come under fire for the slip in
curriculum, and has watched as colleges and universities begin to value the AP exam differently over
time. Top schools like Massachusetts Institute of Technology, Brown University, and Dartmouth have
begun to question the quality of skills that come out of AP courses, some schools choosing to not allow
One of the causes of this review on AP credits in colleges may be attributed to the amount of
material that was required for AP teachers to cover in a course which is spread over a year. Christopher
Drews, a reporter for The Times (2011) interviewed AP teachers and inquired about the difficulty of
meeting the curricular demands of AP and how it may impact students. Teachers of AP Biology, the
second most popular STEM exam offered, reported that the material for AP Biology had grown over
years from a 36 chapter book to 56 chapter covering 1400 pages of material, from which any of the AP
Biology questions could be drawn (Drew, 2011). However, as more public and private high schools
12 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
began criticizing the exams, along with colleges and universities, CollegeBoard did take the feedback
seriously. Trevor Packer, aimed to revisit and rollout new curriculum for the courses, aimed to span
2014-2018 for all courses, and a renewed focus on concept comprehension, rather than just
memorization (Mathews, 2018). Even with new materials being released within the AP program, there
stands an abundance of research looking into racial inequity in testing, access to rural schools, and the
In recent study targeting Hispanic students in different states over a period of 15 years (1997-
2012), noted that the AP Language and Composition and Literature and Composition tests showed a
dramatic lack of preparation for Hispanic students, with only one-third of test takers actually passing the
exam with a 3 or better (Koch, Slate & Moore, 2016). AP Language (or Literature) and Composition are
two of the most popular exams for students across the nation, and are commonly used to assist students
with college-level writing practices. Additionally, with concern to an equity-gap, Wilson, Slate, Moore
and Barnes (2014) demonstrated that AP scores in several common subjects for Black males, in states
with high numbers passing AP scores, failed to reach passing-level with 65% of all Black males from
the study did not earn a passing grade. Along with these sorts of preparatory courses, it is common to
see AP math courses, such as AP Calculus (AB or BC) be utilized to help students standout on not only
Calculus (AB or BC) to search for measurable changes in performance of the ACT. Findings indicated
that while both exams showed a small-positive correlation with performance, the study has limitations
with multiple-subject learners and trying to discern success from AP courses versus high achieving
students (Warne et al, 2015). One of the areas which creates difficulty in measuring success is the
comparison between what the ACT and AP exams attempt to assess. In theory, the ACT tests for college
13 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
readiness and the AP exam tests for college-level performance, creating a mismatch in desired
outcomes, as a high-ACT score would demonstrate strong college-readiness, and high-AP scores would
Ideally, this study aimed to shed light on areas around the academic and retention benefits of AP credit
for students in college, specifically their first-year of academic performance and first two years of
retention. To do so, the study created central hypotheses to allow for targeted outcomes to be isolated,
such as first-term and first-year GPA, incoming academic profile, and retention of incoming AP and the
H1: AP students will consistently have a stronger incoming academic profile than ZCS students,
measured by incoming high school GPA and converted SAT composite score.
H2: AP students will achieve a better first-term GPA more consistently GPA than ZCS students.
H3: AP students will achieve a better first-year GPA more consistently GPA than ZCS students.
In addition to these hypotheses, the study aimed to investigate the relationship of socioeconomic and
academic preparation with regards to first year academic performance with the following research
question (RQ):
RQ: How does socioeconomic status impact first year academic performance?
Chapter 3: Methods
14 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
This study carried out a statistical analysis of incoming first-time, full-time students at Southern
Oregon University who have a freshman class standing. Two groups are determined by students who
have participated in an AP course, and those who have not. The model of this study was a descriptive
correlational design and the aim of this project is to assess the academic readiness and overall student
success and retention of students entering SOU, based on assigned group. An analysis between students
who are zero-credit starts [ZCS] and those who have advanced standing from coursework such as AP
coursework was performed to assess first-year performance, through retention and first-year GPA.
Additionally, this study aimed to track these groups of students through a subsequent year of enrollment
at SOU, in order to gauge the persistence of these groups through their time at SOU. The potential
benefit of this study is to gain an insight on presumed academic preparedness of students who complete
accelerated coursework prior to starting at SOU. This may better allow the campus as a whole to assess
the quality and transferability of the accelerated coursework. For the students, or community-at-large,
there is a benefit in acknowledging the performance outcomes of this program, and to essentially answer
Data has been provided by the Office of Institutional Research is comprised of first-time students, with
freshman class standing, who begin in the fall term of a given cohort. The collection of data provided
includes both academically descriptive and socioeconomic factors collected through surveys, the
Department of Education via the Free Application for Federal Student Aid (FAFSA) and institutional
tracking. The students will be classified as first-time, full-time freshman students, transferring no more
than 44 credits into the university. Data was provided for fall-term cohorts from 2008 through 2017, and
do not include mid-year starting students (i.e. those who begin any other term than fall).
15 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Due to inaccurate and inconsistent reporting of first-generation status, the cohorts from 2008-2010 were
removed, leaving the 2011-2017 cohorts for this study. For confidentiality all student majors were
removed, and replaced with an indicator to show whether a major was declared at the time of admission,
or not. In order to accurately assess financial consideration, within the parameters of this data, students
who did not submit a FAFSA to SOU were excluded, and factors such as veteran status, honors college,
and remedial math placement were also excluded, as they were not directly a part of the analysis, though
they may be considered a potential third variable. Financial data such as unmet financial need, Pell grant
awards, and student loans were all recoded to also include indicators, where a value of 1 would indicate
a nonzero monetary value, during the analysis, either a continuous value will be displayed, or a binary
introductory first-year courses were also excluded. If a student was not identifiable in either group, such
as a transfer student, or a student with credit from an International Baccalaureate, College Level
Equivalency Program, or other course, they were excluded. The final data set yielded N=3098 students,
ranging from the 2011-2017 cohorts. From this data two groups were formed to represent ZCS and AP
students, respectively.
Many self-reported data and survey data tables are known to be incomplete, and have missing values. In
the case of this data, there were missing values in some variables. First, missing data (less than 5%) was
present for the variables tracking first-year and first-term GPA. Given the small sample of missing data,
these values were replaced with a “0.00”, if both variables were missing. There was also missing data in
the variable representing unmet financial need, which lead to a value being assigned of the mean unmet
financial need for that student population (AP or ZCS). Values for the GPA were calculated using the
quality points and credit hours completed for each student, and then calculating a given GPA for the
designated population. The statistical analysis in this study was conducted with the Stata 14.1 software
16 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
package. Given the small percentage of missing data, the addition of the missing values did not impact
Ethical Considerations
Given the nature of the dataset, and its internal use, all student identifying information was
reacted, including student’s major, math placement level, honors college designation, residency status,
and veteran status. After the reaction of data had occurred this study received approval for institutional
The two variables which served as the basis for this study are first-year grade-point average (gpa_y1)
and year-two fall retention (f2_ret). The study allowed for the inclusion of a third dependent variable,
year-three fall retention (f3_ret) as well. Gpa_y1 is a continuous variable, measured by academic
performance in the entirety of the first-year, even if a student stopped-out after their first fall term.
F2_ret and f3_ret are both binary indicators, given that a student who successfully enrolls in any credits
in a subsequent fall term is considered to have retained and would be represented with a value of 1. For
this study, there were 13 independent variables, which included both continuous and binary responses.
Additionally, for the final research question, a new independent variable was introduced to control for
The design for this study was a series of hypothesis tests conducted using both a one-sided, two
sample t-test assuming unequal variance, and an F-test to investigate variance of a given population. In
17 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
these cases, both the null and alternative hypotheses are presented. For the investigation of retentions
between the two populations, a two-sample proportions test was conducted, and excluded a test to
evaluate variance. For the research question proposed in the study, a multilinear regression model built
for the continuous gpa_y1 response variable, and a binary regression model for the f2_ret response
variable. For all the f3_ret data, it is a given that all student-observations in the f2_ret variable were a
value of 1. In order to accurately assess a complete set of data, a set of 80% of the data was removed to
use in building a regression model, to be tested against the remaining 20% of data. Once a complete data
set was constructed, descriptive statistics for both groups of students (AP and ZCS) were created. Next,
regression models were built for all response variables, as specified above.
Independent Variables:
Dependent Variables:
18 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Chapter 4: Results
The central idea of this study was to evaluate the extent to which accelerated coursework, namely
Advanced Placement credits, influenced a student’s incoming, first-term, and first-year performance, if
at all. Below, descriptive statistics are presented to outline the AP students, and the ZCS students, along
with any assumption of population properties. In both populations, the missing data for the first-term
GPA (gpa_f1) were found to be from students who completely withdrew from coursework in their first
term. Because of this, if both the gpa_f1 value and the first-year GPA (gpa_y1) were missing, the
student was assigned a value of “0.0” in both areas. If only the gpa_f1 variable was missing, the value
It is worth noting that the mean value for the SAT composite score (maxc) is computed using the both
the original SAT I and SAT II scorings, along with converted ACT scores. Any missing values of maxc
were replaced with the mean score from that particular population. Additionally, all binary data were
treated as having a Bernoulli distribution when calculating their mean and variance. After a descriptive
overview, the core hypotheses of the study are presented, and are listed below:
H1: AP students will consistently have a stronger incoming academic profile than ZCS students,
measured by incoming high school GPA and converted SAT composite score.
H2: AP students will achieve a better first-term GPA more consistently GPA than ZCS students.
H3: AP students will achieve a better first-year GPA more consistently GPA than ZCS students.
In addition to these hypotheses, the study aimed to investigate the relationship of socioeconomic and
academic preparation with regards to first year academic performance with the following research
question:
RQ: How does socioeconomic status impact first year academic performance?
19 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Table 4.1 describes the populations of AP Students in this study. With 721 students considered in the AP
student population, it is evident the participants had a high credit-completion rate (hrs_ratio_f1) of
95.82% during their first term, accompanied with a healthy enrollment of about 15 credits, an amount
which is recommended at quarter-based schools for timely graduation. The mean maxc of these students
was 1230.80 out of a possible 1600. Of these students, over a quarter of the population were either
recipients of the Pell grant (Low Income) or of an ethnic background (ethn_nonwht), with one-in-five
students being first-generation (first_gen) students. As students progressed through the first-year, a
marginal decrease in academic performance occurred, as students has a high school GPA
(incoming_gpa) of 3.52, with a narrow standard deviation of only 0.36. The population of AP students
then progressed downward to GPAs of 3.47 and 3.36 for gpa_f1 and gpa_y1, respectively, with a
marked increase in standard deviation. Graphically, the AP students’ GPA profile is represented with
Chart 4.1. The distribution of these GPA’s was non-normal, but treated as normal in the analysis
presented.
0.8
Frequency
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
GPA
20 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Table 4.2 describes the population of the ZCS students in this study. With 2375 students considered in
the ZCS student population, it is evident the participants had a moderate credit-completion rate of
88.90% during their first term, accompanied with a fair enrollment of about 14 credits, an amount which
is below the recommended baseline at quarter-based schools for timely graduation. The mean maxc of
these students was 1076.12 out of a possible 1600. Of these students, nearly half (44.46%) of the
population were recipients of the Pell grant. Additionally, first-generation students and those of an
ethnic background were present in over a third of the population (34.06% and 37.5%, respectively). As
students progressed through the first-year, a non-marginal decrease in academic performance occurred,
as students has a high school GPA of 3.19, with a narrow standard deviation of only 0.38. The
population of ZCS students then progressed downward to GPAs of 2.98 and 2.77 for gpa_f1 and
gpa_y1, respectively, with a marked increase in standard deviation. Graphically, the ZCS students’ GPA
21 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
0.8
Frequency
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
GPA
The research question posed by this study is based on the overall retention and performance of both
groups of students, compared to each other. Below, Table 4.3 and 4.4 illustrate the year-one to year-two,
and year-two to year-three retention (f2_ret and f3_ret, respectively). Overall, students from the AP
group retained at a higher rate (78.09%) versus the ZCS students who achieved a rate of 67.24% from
their first to second year in college. While the difference in f2_ret was noticeable at 10.84%, in favor of
the AP students, this gap narrowed to 3.14% for f3_ret, wherein AP and ZCS students retained from the
previous year at 60.03% and 56.86%, respectively. There were a total of 1834 (59%) students lost over
22 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Table 4.3 Year-One to Year-Two (Fall to Fall) retention rates for AP Table 4.4 Year-Two to Year-Three (Fall to Fall) retention rates for
AP and ZCS Students and ZCS Students
# of
Retention # Students Retention Retention Total Loss
Students
Rate Retained of f2_ret from start
at start
f2_ret (AP) 78.09% 563 721 f3_ret(AP) 60.03% 47.43% -379
f2_ret
67.24% 1597 2375 f3_ret(ZCS) 56.86% 38.74% -1455
(ZCS)
The first hypothesis presented in this study was to determine if the incoming profile of the AP students,
from high school, proved to be stronger than those of the ZCS group, measured with the converted SAT
composite score and high school GPA. The null and alternative hypotheses were:
Equation 4.1: Hypothesis for mean values of “incoming_gpa” for AP and ZCS students.
Equation 4.2: Hypothesis for mean values of maxc for AP and ZCS students
𝐻𝑜 : 𝜇𝑚𝑎𝑥𝑐,𝐴𝑃 − 𝜇𝑚𝑎𝑥𝑐,𝑍𝐶𝑆 = 0
𝑃𝑎𝑟𝑡 2 − {
𝐻𝑎 : 𝜇𝑚𝑎𝑥𝑐,𝐴𝑃 − 𝜇𝑚𝑎𝑥𝑐,𝑍𝐶𝑆 > 0
The first part of the hypothesis was tested using a one-tailed t-test for two samples having unequal
Table 4.5 Hypothesis testing for incoming GPA for AP and ZCS Table 4.6 Hypothesis testing for converted SAT scores for AP and
ZCS Students students
Hypothesis 1, Part 1: Incoming GPA of AP and ZCS Hypothesis 1, Part 2: Converted SAT scores of AP and
Students ZCS Students
t-Test: Two-Sample Assuming Unequal Variances t-Test: Two-Sample Assuming Unequal Variances
incoming_gpa incoming_gpa(AP) maxc(ZCS) maxc(AP)
(ZCS)
Mean 3.185915789 3.521969487 Mean 1076.122947 1230.804438
Variance 0.146814399 0.130062505 Variance 18112.45666 14229.62976
Observations 2375 721 Observations 2375 721
Mean Difference 0.34 Mean Difference 154
Degrees of 1253 Degrees of Freedom 1324
Freedom
t Stat Value -43.43967635 t Stat Value -59.01122056
23 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
It was shown that the null hypothesis was rejected, affirming that incoming students from the AP
population had a stronger academic performance as presented by high school GPA. With this test
confirming the first part of the hypothesis, it was necessary to test the students’ converted SAT
composite scores, as presented in Table 4.6. Again, results affirmed that the null hypothesis is rejected,
and that the students from the AP population have a stronger incoming test scores. In order to have a
stronger understanding of the relationship between the AP and ZCS students upon entering college, the
variance of these two populations were tested through a two variance F-test to check if both categories
of academic performance (incoming GPA and converted test scores) were consistently higher in the AP
population than the ZCS population. The hypotheses for these tests are:
Equation 4.3: Hypothesis test for variability of “incoming_gpa” for AP and ZCS students.
𝐻𝑜 : 𝜎 2 𝑚𝑎𝑥𝑐,𝐴𝑃 = 𝜎 2 𝑚𝑎𝑥𝑐,𝑍𝐶𝑆
𝑃𝑎𝑟𝑡 2 − {
𝐻𝑎 : 𝜎 2 𝑚𝑎𝑥𝑐,𝐴𝑃 < 𝜎 2 𝑚𝑎𝑥𝑐,𝑍𝐶𝑆
Table 4.7 2-Sample test for Variance for incoming GPA and converted SAT composite score for AP and ZCS students
24 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
The p-values for the F-tests in Table 4.7 show that in both measures of academic profile strength
presented in this hypothesis, the AP students were performing at a higher-level of achievement, more
consistently.
In the second hypothesis the study tested the performance of the AP students to the ZCS students in their
first-term performance in terms of their first-term performance, academically. The hypothesis testing for
this comparison wass two-fold. Beginning with a study of overall performance through a one-tailed t-
test, similar to above, to establish a higher mean GPA for AP students, the study then tested the idea of
decreased variability (increased consistency) in the first-year performance for both groups. The
Equation 4.5: Hypothesis test for mean values of first-term academic performance in AP and ZCS students.
Equation 4.6: Hypothesis test for variability of first-term academic performance in AP and ZCS students.
After the hypothesis tests we conducted, results showed that there was a statistically significant
difference in the mean performance of AP students during their first-term (p-value <.00001), allowing
the null hypothesis to be rejected. Similarly, the variability of both samples was tested in a two-sample
variance F-test, which yielded a statistically significant results (p-value < .00001), affirming that a
higher variability exists in the ZCS students than the AP students, during their first-term academic
performance. Tables 4.8 and 4.9 summarize the results of the tests.
25 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Table 4.8: Summary of hypothesis testing for the mean first-term Table 4.9 Summary of hypothesis test for variability of first-term
performance of AP and ZCS students academic performance of AP and ZCS students
Hypothesis 2: First-Term GPA of AP and ZCS Hypothesis 2: First-Term GPA of AP and ZCS Students
Students
F-Test Two-Sample for Variances
t-Test: Two-Sample Assuming Unequal Variances
gpa_f1 gpa_f1 (AP)
gpa_f1 (ZCS)
gpa_f1 (AP)
(ZCS) Mean 2.975251051 3.469307251
Mean 2.975251051 3.469307251
Variance 0.865677774 0.444832677
Variance 0.865677774 0.444832677
Observations 2375 721
Observations 2375 721
Degrees of Freedom 2374 720
Mean Difference 0.49
F 1.946075049
Degrees of Freedom 1648
P(F < f) one-tail <.00001
t Stat Value -13.258
F Critical one-tail 1.105767345
P(T<t) one-tail <.00001
In the third hypothesis the study tested the performance of the AP students to the ZCS students in their
first-year performance in terms of their performance, academically. The hypothesis testing for this
comparison was two-fold. Beginning with a study of overall performance through a one-tailed t-test,
similar to above, to establish a higher mean GPA for AP students, the study then tested the idea of
decreased variability (increased consistency) in the first-year performance for both groups. The
Equation 4.7: Hypothesis test for mean values of first-year academic performance in AP and ZCS students.
Equation 4.8: Hypothesis test for variability of first-year academic performance in AP and ZCS students.
After the hypothesis tests we conducted, results showed that there was a statistically significant
difference in the mean performance of AP students during their first-term (p-value <.00001), allowing
the null hypothesis to be rejected. Similarly, the variability of both samples was tested in a two-sample
26 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
variance F-test, which yielded a statistically significant results (p-value < .00001), affirming that a
higher variability exists in the ZCS students than the AP students, during their first-term academic
performance. Tables 4.10 and 4.11 summarize the results of the tests.
Table 4.10: Summary of hypothesis testing for the mean first-year Table 4.11 Summary of hypothesis test for variability of first-
year performance of AP and ZCS students academic performance of AP and ZCS students
Hypothesis 3: First-Year GPA of AP and ZCS students Hypothesis 3: First-Year GPA of AP and ZCS students
F-Test Two-Sample for Variances One-Sided T-Test: Two sample Assuming Unequal
Variances
gpa_y1 gpa_y1 (AP) gpa_y1 (ZCS) gpa_y1 (AP)
(ZCS)
Mean 2.766943158 3.355520111 Mean 2.766943158 3.355520111
Variance 0.729084081 0.414195597 Variance 0.729084081 0.414195597
Observations 2375 721 Observations 2375.00 721.00
Degrees of Freedom 2374 720 Degrees of 1560.00
Freedom
F 1.76024102 Mean Difference 0.59
P(F < f) one-tail <.00001 t Stat Value 39.70
F critical one-tail 1.105767345 P(T < t) one-tail <.00001
t Critical one-tail 1.65
The last hypothesis of the study aimed to investigate the relationship of retention rates in the second and
third years of study during their collegiate career (f2_ret and f3_ret, respectively). Given that these were
measured as a percentage, or proportion (𝑝̂ ), of retained students, a two-sample proportions z-test was
determined to be the most appropriate test. The values for these proportions we summarized in Table 4.2
and Table 4.3 for f2_ret and f3_ret, respectively. The hypotheses tested were:
Equation 4.9: Hypothesis testing of proportions for f2_ret for AP and ZCS students.
Equation 4.10: Hypothesis testing of proportions for f3_ret for AP and ZCS students.
27 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
The results of these two tests showed to have contradictory results, wherein the hypothesis tested in
Equation 4.9 yield a statistically significant result, (p-value <.00001) in favor of the alternative
hypothesis, suggesting that the f2_ret proportion for the AP students was higher than the ZCS students.
On the contrary, for the f3_ret proportions, tested in Equation 4.10, showed to have no significant
The last area investigated in this study is to determine the impact of financial and demographic
background on the first-year performance for AP and ZCS students. A robust multiple linear regression
model was employed on two separate cases. The first instance incorporated all variables into the
regression of the gpa_y1 variable. The subsequent model included attempts to combine academic
factors, and exclude socioeconomic information. Table 4.12 summarizes the outcome of the regression
model for AP students, including all variables. Difficulty in the modelling surfaced when many of the
hypothetical tests for the coefficients returned their lack of significance; however, the fit of the model
seemed relatively strong. Only the coefficient for gpa_f1 returned a significant value, with ethn_nonwht
Table 4.12 Summary of regression coefficients for first-year GPA of AP students, including socioeconomic factors.
28 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
In the second model, it was shown that the robust linear regression with just Academic Index showed a
strong correlation (.8764) with the gpa_y1 variable. The regression model had a strong fit with an
R2=.7681, and a statistically significant regression coefficient (p-value < .0001). The regressions of both
cases are presented in Charts 4.3 and 4.4, and detail the plot of predicted values of gpa_y1 and actual
gpa_y1_predicted gpa_y1_predicted
Chart 4.3 Scatterplot for AP students of predicted values of Chart 4.4 Scatterplot for AP students of predicted values of
gpa_y1 with regression using all variables, excluding gpa_y1 with regression. R2 = 0.7681 for inclusion of only the
academic index. R2 = 0.8224 for inclusion of all variables. academic index for scholastic factors.
Similarly, for the ZCS students, a robust multiple linear regression model was employed on two separate
cases. The first instance incorporated all variables into the regression of the gpa_y1 variable. The
29 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
subsequent model included attempts to combine academic factors, and exclude socioeconomic
information. Table 4.13 summarizes the outcome of the regression model for ZCS students, including all
variables.
Table 4.13 Summary of regression coefficients for first-year GPA of ZCS students, including socioeconomic factors.
There was less difficulty in the modelling for this population as more of the hypothetical tests for the
coefficients returned their significance. Strong significance was shown with gpa_f1, hrs_ratio_f1,
incoming_gpa, and unmet (p-values < .0001), and moderate significance with enrl_hrs_f1,
ethn_nonwht, and loan_amt_f1 (p-value < .05). The fit of the model seemed relatively strong, having an
R2 value of .8069. In the model for predicting first-year GPA without the socioeconomic factors,
including only the academic index, the overall fit was somewhat weaker than the case for the AP
students, yielding an R2 of .7426, and a correlation of .8614. Based on the output of both models, it was
determined that multicollinearity was present, and that more descriptive investigation was warranted
30 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
into the effects of socioeconomic status. The regressions of both cases are presented in Charts 4.5 and
4.6, and detail the plot of predicted values of gpa_y1 and actual values for the ZCS students.
gpa_y1 gpa_y1
With the sample being comprised of students who had filed a FAFSA, and had it submitted to SOU, the
most convenient metric for income was found to be the estimated family income (efc). In order to
determine the effect of efc on each student population, quartiles of were constructed from the overall
distribution of efc values. Tables 4.13 shares the results of the investigation. The results indicated a
statistically significant difference in performance of the first-year GPA of the populations of ZCS and
AP students (p-value < .05) across all quartiles. First-year retention was found to be in favor of the AP
students across all EFC quartiles, though no statistical test was performed. It is clear that the uppermost
quartile, in each population, experienced higher performance with their first-year GPA and academic
index. For the upper-three quartiles in the AP population, there was not a strongly varied presence of
ethnic minorities; however, there was a stout growth in the presence of first-generation students as the
EFC lowered, ranging from 11% at the 4th quartile and 31.9% in the 1st. Similarly, for ZCS students,
there was a marked growth in the presence of first-generation students ranging from 14% in the 4th-
31 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
f2_ret f3_ret gpa_y1 majr_undl athlete ethn_nonwht first_gen efc unmet Low Income Academic Index
EFC Quartile 4
ZCS (N=472) 75.1% 52.9% 2.94 12.1% 18.8% 26.8% 14.0% 53038 -527 0.0% 6.65
AP (N=144) 85.5% 57.3% 3.43 9.7% 13.1% 29.7% 11.0% 52847 -1302 0.0% 6.93
EFC Quartile 3
ZCS (N=473) 67.0% 59.2% 2.82 15.5% 23.5% 34.7% 25.2% 12930 2538 0.0% 6.48
AP (N=144) 76.6% 61.6% 3.33 12.5% 9.7% 27.8% 16.0% 21182 488 0.0% 6.71
EFC Quartile 2
ZCS (N=473) 61.3% 55.3% 2.72 13.3% 17.2% 44.1% 43.0% 2740 4726 79.9% 6.5
AP (N=144) 70.8% 55.3% 3.43 17.4% 4.2% 27.8% 21.5% 9164 2725 13.8% 6.94
EFC Quartile 1
ZCS (N=472) 63.9% 61.2% 2.59 14.8% 13.5% 43.3% 52.2% 4 4750 98.3% 6.28
AP (N=144) 75.2% 64.5% 3.25 12.5% 8.3% 36.8% 31.9% 887 4767 98.6% 6.57
Table 4.15: Descriptive Summary of AP and ZCS students by controlling for ethnicity and first-generation status.
N f2_ret f3_ret gpa_y1 majr_undl athlete efc unmet Low Income Academic Index
Ethnicity & First Gen = 0
ZCS 84467.9% 56.2% 2.82 15.2% 14.2% 23720 2268 31.6% 6.53
AP 335 79.1% 58.1% 3.44 14.3% 7.2% 23921 1413 23.6% 6.95
Ethnicity = 1, First Gen = 0
ZCS 41170.6% 53.1% 2.8 13.9% 23.4% 17145 2893 39.9% 6.44
AP 125 80.0% 60.4% 3.21 12.4% 10.4% 21039 1276 27.2% 6.51
Ethnicity = 0, First Gen = 1
ZCS 342 62.3% 64.5% 2.71 14.0% 14.6% 10811 3359 58.5% 6.46
AP 65 69.2% 60.9% 3.36 10.8% 7.7% 13298 2419 36.9% 6.87
Ethnicity & First Gen = 1
ZCS 293 64.8% 58.1% 2.62 10.2% 27.0% 5875 4006 72.0% 6.39
AP 51 70.6% 69.4% 3.09 7.8% 17.6% 11763 3436 49.0% 6.37
32 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Chapter 5: Discussion
Discussion of findings
The first hypothesis of this study was to assess the claim that AP participation can positively
influence an incoming student’s academic profile, specifically in terms of their converted SAT
composite score (maxc) and incoming GPA(incoming_gpa) from high school. As presented in Tables
4.5 and 4.6, hypothesis testing confirmed that there is a statistically significant difference in the
performance of incoming AP and ZCS students, skewed in favor of the AP students which suggests that
the average incoming performance of AP students is, on average, higher in both incoming_gpa and
maxc. The findings related to incoming_gpa tend to agree with the literature suggesting that AP students
are naturally high-achieving students (Ewing & Howell, 2015; North & Jacobs, 2010; Adelman, 2006).
It also aligns with the findings that support that AP students tend to graduate at higher rates than non-
participants (O’Keefe, 2009). AP student’s performance with regards to converted SAT score agrees
with the findings on previous studies, suggesting that passing an AP exam leads to a higher performance
on an ACT test (Warne, et al., 2015). While these studies, and many overall, consider the GPA
performance of students in both secondary and postsecondary education, little look into the consistency
of performance.
One area this study addressed which is excluded from much of the literature, is the topic of consistency
relationship, the test of variances showed that the variance on the incoming_gpa and maxc for ZCS
students were higher than those for the AP students, suggesting that not only do AP students have a
stronger admittance profile, but outperform ZCS students more consistently. This would suggest that
students who participate in AP courses will tend to have stronger academic profiles, and additionally, do
so more consistently. However, it is worth noting that while AP student tend to outperform ZCS students
33 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
in these areas, there is not enough evidence to suggest that a student’s participation in AP coursework
alone was the reason they achieved these feats. Next, this study aimed to investigate impact on first-term
The second hypothesis of this study was to determine the impact that AP coursework had on a student’s
first-term performance in college. Equations 4.5 and 4.6 outline both of the tests performed to assess
both mean academic performance, and variance of academic performance, respectively. Both of these
tests demonstrated favor to the alternative hypothesis, confirming that AP students, on average, perform
better in their first-term, and do so more consistently. Given that this study included only students who
received credit for AP courses (i.e. a credit-bearing score), these results agree with Fara’s (2010)
findings linking AP scores, along with incoming high school GPA and college admission test scores, to
college success. With little literature analyzing the impact of AP courses on the first-term (rather than
first-year) performance of college students, this study aims to add to the literature with the findings of
the second hypothesis. Next, this study evaluated the impact of AP coursework on first-year
The third hypothesis aimed to assess the supposed advantage that AP students experience in terms of
their first-year performance. Not surprisingly, first-year GPA, on average, was shown to be stronger for
the AP students when compared to the ZCS students. Tests confirm that AP students achieve a higher
mean first-year GPA, more consistently, which is closely related with first-term performance a priori.
AP examinations and their predictive validity and have been well-studied and have shown some sort of
positive correlation, either weak or strong, when associated with college attendance, completion, and
overall performance (Patterson & Ewing, 2013; Shaw, Mattern, & Marini, 2012; Morgan & Klaric,
2007). The outcome of this hypothesis is in agreement with previous literature, claiming that programs,
34 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
like AP, lead to first-year success (Ewing & Howell, 2015; North & Jacobs, 2010; Adelman, 2006).
Lastly, this study aimed to assess the difference in first- and second-year retention for both populations.
The last hypothesis assessed the difference in the retention rates of AP and ZCS students. At first glance,
it seems that the AP student population of students outperformed the ZCS students in every academic
category in their first-year at SOU. This can lead to the natural assumption that higher performance
begets higher retention rates, which proved to be true in the first-year for both populations. This gives
weight to the findings which show accelerated coursework can have a positive impact on student’s
collegiate success (Ewing & Howell, 2015; North & Jacobs, 2010; Adelman, 2006). Surprisingly, this
effect was stymied by the end of the second year for the AP students, in comparison of the ZCS
populations, where the retention of AP students was not shown to have any statistical difference.
The research question in this study attempts to describe the impact of socioeconomic factors on first-
year success. In this case, the socioeconomic considerations are split into quartiles summarized in Table
4.13 and 4.14. It is clear that first-year retention varied by EFC quartile, and is statistically significant
between AP and ZCS students in each quartile (p-values < .05), and in agreement with the last
hypothesis of the study. Second-year retention is generally not statistically different, except at the 4th-
quartile of both populations, wherein AP students, on average, retain at a higher rate in the second year
also (p-value < .005). While the participation in AP participation of low-income and first-generation
students have increased over time, there seems to be a limited benefit when assessing retention
(Mathews, 2018).
Other factors which presented over the quartiles include the relative population of first-generation and
ethnic minority students, unmet financial need, and academic index. As EFC decreases, there is an
increased presence of first-generation and minority students, and the increase in unmet financial need.
The academic index seemed to have little variability in the upper three quartiles, but showed a relatively
35 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
larger decrease in the lowest quartile. Interestingly, at the lowest quartile for ZCS students, the academic
index was at its lowest, yet second-year retention showed to be strong by relation to first-year retention,
by comparison to ZCS students at other quartiles. This meant that second-year retention was the highest
for the lowest-income population (for both AP and ZCS students), a surprising fact, which could show
the strength of support services for low-income and first-generation students. This ideology aligns with
the literature about how support services impact improvement of low-income and first-generation
students (Cataldi et al., 2018; Chemers et al., 2001; Crystal et al., 1994; Engle & Tinto, 2008; North &
With consideration to ethnicity and first-generation status, this study explores the descriptive
relationship between these factors and first-year performance. Table 4.14 summarizes the categories
analyzed. In the ZCS population, 58.4% of students were either first-generation or minority students (or
both), and these students only comprise 41.8% of the AP population. Interestingly, 72% of ZCS students
and 49% of AP students who are both first-generation and from an underrepresented background are
classified as low income (receiving the Pell Grant). In order to gain more insight into how scholastic and
socioeconomic factors impact first-year performance, multiple regressions were conducted on both AP
and ZCS students for all variables and then for only scholastic variables contained within the academic
Figures 4.1 and 4.2 show that the scholastic variables can account for much of the variability in the first-
year GPA of AP students, but not as much as the model which includes socioeconomic variables.
However, due to multicollinearity of all predictors, it is difficult to discern the impacts of the individual
variables, though the overall fit of the model contained in Figure 4.1 is strong. There is a strong
clustering near the regression line for higher values of first-year GPA suggesting that socioeconomic
factors may cause larger variance as GPA is decreased; however, there is no conclusive evidence of this.
36 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Similar observations can be made for the models contained in Figures 4.3 and 4.4 for the ZCS students.
Overall, both models have a good fit for the first-year GPA. As with the AP populations, there is an
increase in model variance for lower first-year GPA; however, in both populations, it is possible that
Conclusions
It is clear throughout the study that AP students are generally higher-achieving in academics, both on
arrival and in first-year performance. Additionally, their ability to retain is shown to be higher through
the first-year, though this benefit seems to be rendered insignificant for the second-year. Socioeconomic
factors seem to be less of a concern for AP students throughout the first-year, suggesting that AP
coursework may be able to close the equity and opportunity gap for low income and first-generation
students with similar incoming background. While much of the variance in first-year performance can
be accounted for with the variables presented in this study, it is clear that socioeconomic factors impact
AP and ZCS students differently, seemingly more impactful for ZCS students. The most likely source
for the unaccounted variance in the regression models for first-year GPA in AP and ZCS students are
more psychosocial instances which may positively (tutoring, peer mentoring, campus employment, etc)
or negatively (life event, food or housing insecurity, etc) impact student performance.
This study is not without limitations. The data provided is a combination of both reported variables from
the institution, and self-reported data from sources such as informational cards admissions application.
For example, ethnicity may not always be accurately reported, and a student’s academic major will
change over their time in college. There is no data to track student intervention, or any sort of large-scale
37 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
personal matter which may have impacted a student’s ability to learn or participate in courses.
Additionally, while we track some financial data, there is not a variable to track financial literacy of
students, especially in the first-generation population. First-year students often have to address the factor
of living in campus housing, and having a randomly assigned roommate, which is an additional
Another key area where this model could be improved is the AP test subject content, and amount of
credits earned by AP coursework, or failure to include students who took an AP course, but did not
addition to the data, there are concerns of the model used in this study. Many of the hypotheses tested in
this study have an expectation of data normality, or mean normality, which is not necessarily true in
each case. Multicollinearity is known to negatively impact multiple regression models, and while the
academic index was established to help control for scholastic multicollinearity, there was no successful
Future studies may look to incorporate other types of accelerated such as International Baccalaureate or
transfer students as other comparison groups to search for discrepancies in these areas. Additionally,
tracking the amount of credits earn, housing status, and any indicators of large-scale personal events for
students. Tracking additional areas such as campus employment and use of academic resources like
tutoring, TRiO, or other student support services could be added to help describe some of the variance in
student performance and retention. For first-generation students especially, tracking things like financial
literacy, entrance counselling and engagement in peer mentoring programs could help create a more
robust model.
38 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
References
Adelman, C. (2006). The toolbox revisited: paths to degree completion from high school through
college. Washington, D. C.: Office of Vocational and Adult Education, U.S. Dept. of Education.
research/pubs/toolboxrevisit/toolbox.pdf
Bandura, A., & Estes, William K. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Camara, W., Dorans, N., Morgan, R., & Myford, C. (2000). Advanced Placement: Access Not
doi:http://dx.doi.org/10.14507/epaa.v8n40.2000Institute.
Cataldi, E., Bennett, C., & Chen, X. (2018). First-generation students: College access, persistence, and
https://nces.ed.gov/pubs2018/2018421.pdf
Chemers, M. M., Li-tze, H., & Garcia, B. F. (2001). Academic self-efficacy and first-year college
Chen, R. (2012). Institutional characteristics and college student dropout risks: A multilevel event
http://dx.doi.org/10.1007/s11162-011-9241-4
CollegeBoard. (2014). The 10th annual AP report to the nation (Rep.). New York, NY: CollegeBoard.
CollegeBoard. (2017). AP score distributions - All subjects 1996-2006 (Rep.). New York, NY:
CollegeBoard.
39 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Crystal, D., Chen, C., Fuligni, A., Stevenson, H., Hsu, C., Ko, H., . . . Kimura, S. (1994). Psychological
Drew, C. (2011, January 7). Rethinking Advanced Placement. Times: Education Life. Retrieved from
https://www.nytimes.com/2011/01/09/education/edlife/09ap-t.html
Education Commission of the States (2019). 2019 State Education Policy Watch List. Retrieved from:
https://www.ecs.org/state-education-policy-watch-list/
Engle, J., & Tinto, V. (2008). Moving beyond access: College success for low-income, first generation
students. Pell Institue for the Study of Opportunity in Higher Education. Retrieved March 31,
Ewing, M. & Howell, J. (2015) Is the relationship between AP participation and academic performance
really meaningful? (College Board Research Brief No. 2015-01). Retrieved from
research.collegebaord.org
Fara, K. J. (2010). The relationship of college credit earned while in high school to first-semester
college GPA and persistence to the second college year(Unpublished doctoral dissertation). Iowa
https://lib.dr.iastate.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=2829&context=etd
Gagnon, D. J., & Mattingly, M. J. (2016). Advanced Placement and rural schools. Journal of Advanced
Klopfenstein, K. (2010). Does the Advanced Placement Program save taxpayers money? The effect of
Klopfenstein (Eds.), AP: A critical examination of the Advanced Placement Program (pp 189-
40 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Koch, B., Slate, J. R., & Moore, G. W. (2016). Advanced Placement English exam scores. Education
Lichten, W. (2000). Whither Advanced Placement?. education policy analysis archives, 8, 29.
doi:http://dx.doi.org/10.14507/epaa.v8n29.2000
Licthen, W. (2007). Equity and excellence in the CollegeBoard Advanced Placement Program. Teachers
Licthen, W. (2010). Whither Advanced Placement – Now? In P.M. Sadler, G. Sonnert, R.H.Tai & K.
Klopfenstein (Eds.), AP: A critical examination of the Advanced Placement Program (pp.233-
Lohman, B., & Jarvis, J. (2000). Adolescent stressors, coping strategies, and psychological health
studied in the family context. Journal of Youth and Adolescence, 29(1), 15-43.
Mathews, J. (2004). Why Colleges Think They're Better Than AP. Washingtonpost.com. Retrieved from
https://link-galegroup-com.glacier.sou.edu
McBride Davis, C., Slate, J., Moore, G. & Barnes, W. (2015). Advanced Placement exams, incentive
programs, and cost effectiveness: A lack of equity and excellence for black students in Texas,
New York, and Florida. Journal of Begro Education, 84, 139-153. doi:
10.7709/jnegroeducation.84.2.0139
McNamara, S. (2000). Stress in young people : What's new and what can we do? London ; New York:
Continuum.
Millea, M., Wills, R., Elder, A., & Molina, D. (2018). What matters in college student success?
41 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Morgan, R., & Klaric, J. (2007) AP students in college: An analysis of five-year academic careers.
National Center for Educational Statistics. (2018a). The condition of education 2018. (NCES 2018-114).
https://nces.ed.gov/pubs2018/2018144.pdf
National Center for Educational Statistics. (2018b). Digest of education statistics 2016. Washington,
https://nces.ed.gov/programs/digest/d16/index.asp
National Center for Educational Statistics. (2018c). Price of attending an undergraduate institution.
https://nces.ed.gov/programs/coe/indicator_cua.asp
North, T., & Jacobs, J. (2010) Dual credit in Oregon: An analysis of students taking dual credit in high
www.ous.edu/dept/ir/reports
O'Keefe, L. M. (2009). The effect of participation in Advanced Placement and/or dual credit on four-
year graduation rates. (Master's thesis, University of North Texas, 2009). Denton: University of
https://digital.library.unt.edu/ark:/67531/metadc12172/m1/39/.
Patterson, B., & Ewing, M. (2013) Validating the use of AP exam scores for college course
Redford, J., & Hoyer, K.M. (2017). First-generation and continuing-generation college students: A
comparison of high school and postsecondary experiences. (NCES 2018-009). U.S. Department
42 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
of Education. Washington, DC: National Center for Education Statistics. Retrieved from
https://nces.ed.gov/pubs2018/2018009.pdf.
Shaw, E., Mattern, K., & Marini, J. (2013) Considering practical uses of Advanced Placement
http://research.collegeboard.org
Spalding, A., Eden, A., & Heppner, R. (2012). Implementing the AP for All Movement in two Florida
high schools. In B. Smeardon & K. Borman (Eds.) Pressing Forward: Increasing and expanding
rigor and relevance in America’s high schools (pp. 49-82). Charlotte, NC: Information Age
http://www.pellinstitute.org/files/COEMovingBeyondReport Final.pdf
Suldo, S., Shaunessy-Dedrick, E., Ferron, J. & Dedrick, R. (2018). Predictors of success among high
school students in Advanced Placement and International Baccalaureate programs. Gifted Child
Tierney, J. (2012, October 13). AP Classes Are a Scame. The Atlantic. Retrieved from
https://www.theatlantic.com/national/archive/2012/10/ap-classes-are-a-scam/263456/
Doi: 10.5204/ssj.v8i2.376.
Vuong, M., Brown-Welty, S., & Tracz, S. (2010). The effects of self-efficacy on academic success of
first-generation college sophomore students. Journal of College Student Development, 51(1), 50-
64.
Walton, G. M., & Cohen, G. L. (2007). A question of belonging: Race, social fit, and achievement.
3514.92.1.82
43 | P a g e
THE IMPACT OF ADVANCED PLACEMENT COURSES
Warne, R. T., Larsen, R., Anderson, B., & Odasso, A. J. (2015). The impact of participation in the
Advanced Placement program on students college admissions test scores. The Journal of
Wilson, J. L., Slate, J. R., Moore, G. W., & Barnes, W. (2014). Advanced Placement scores for Black
male students from Connecticut, Florida, Maryland, Massachusetts, and Texas. Education
44 | P a g e