BIBLIOGRAPHY CARIAS, JULIET N. and...
BIBLIOGRAPHY
CARIAS, JULIET N. and LUISITO, MELDRID A., April 2008. Logistic
Regression Analysis on the Decision of Freshmen College Students to Pursue Their
Enrolled Course. Benguet State University, La Trinidad, Benguet.
ADVISER: CRISTINA B. OCDEN
ABSTRACT
This study was conducted to determine the odds on the decision of freshmen
college students to pursue their enrolled course and to determine the variables that affect
the decision of freshmen college students to pursue their enrolled course.
The data were obtained from randomly selected freshmen college students
enrolled in degree programs of the Benguet State University with the aid of a
questionnaire as guide. The Logistic Regression Analysis was used to determine the
factors that affect freshmen college students whether to pursue or not to pursue their
enrolled course.
Based on the result of the study, 78 percent of the students are decided to pursue
their enrolled course as opposed to 22 percent students who are decided not to pursue
their enrolled course. The variables that affect significantly in pursuing their course are:
gender, family size, and scholarship.
With all the results obtained, further study on the decision of freshmen college
students is recommended to identify other factors that influence freshmen college
students to pursue or not to pursue their enrolled course.
TABLE OF CONTENTS
Pages
Bibliography……………………………………………………………… i
Abstract……………………………………………………………………. i
Table of Contents…………………………………………………………. ii
INTRODUCTION
Background of the Study …………………………………………. 1
Objectives of the Study …………………………………………... 3
Importance of the Problem ……………………………………… 3
Scope and Delimitation of the Study ……………………………. 4
REVIEW OF RELATED LITERATURE……………………………….. 5
THEORETICAL FRAMEWORK ……………………………………….
12
Definition of Terms ………………………………………………. 16
METHODOLOGY
Respondents of the study………………………………………… 18
Data Collection…………. ………………………………............ 18
Data Gathered………………………………………………........ 18
Data Analysis……………………………………………………. 21
RESULTS AND DISCUSSION……………………………………....... 22
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS............ 33
LITERATURE CITED………………………………………………….. 36
APPENDICES…………………………………………………………... 38
ii
1
INTRODUCTION
Background of the Study
Going to college is one of the most important decisions a high school
student can make. Contrary to popular belief, preparing for college is even more
important than simply going to college. How a student prepares for college has a
great influence whether or not the student will actually go to college. Many
students and parents underestimate the importance of college preparation. Due to
this miscalculation; students go to college without a clear choice of concentration
and the right choice of college.
Choice of concentration or course required inquiry and reasoned
judgement and some creative research on the part of students. Because choice of
concentration is a complex decision, it is not surprising that a number of students
change concentrations after first semester or after the first year of college.
Aside from choosing a course, choosing a college is one of the most
difficult and important decisions/choices a student will ever have to make. It is
important to select a college that first, has the major that is interests you and
second, has a good program. Even if students are not sure what course to take,
they must visit different colleges to have information on the courses offered.
According to a survey conducted at University of California, Los Angeles
(UCLA), the percentage of freshmen choosing their college because of low tuition
fee rose sharply to its second highest point ever, 31.3 percent compared to 27.7
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
2
percent in 1995. And those going elsewhere because they were not offered
financial aid by their first college reached a high 5.7 percent up from 5.2 percent
in 1989, when this question last was asked.
Benguet State University is one of the 81 State Universities and Colleges
(SUC’s) in the Philippines located at La Trinidad, Benguet where students,
Filipinos or foreigners choose to study. It is known for its popularity as the center
of excellence in agricultural education, offering high quality education and
assessing low cost tuition.
The courses offered are Bachelor of Science in Agriculture (BSA),
Bachelor of Science in Agricultural Engineering (BSAE), Bachelor of Science in
Applied Statistics (BSAS), Bachelor of Science in Environmental Science
(BSES), Bachelor of Science in Information Technology (BSIT), Bachelor of
Science in Nursing (BSN), Doctor of Veterinary Medicine (DVM), Bachelor of
Science in Elementary Education (BEE), Bachelor of Science in Home
Economics (BSHE), Bachelor of Science in Nutrition Dietetics (BSND), Bachelor
of Science in Forestry (BSF), Bachelor of Science in Agri-business (BSAB),
Diploma in Agro-Forestry (DAF), and a 6 months Automotive Technology.
Choice of college and choice of concentration or course go together.
Changing concentration and changing of college is a waste of time and money. To
avoid this, it is important to study the factors that affect the decision of freshmen
college students to pursue their enrolled course.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
3
Objectives of the Study
The main objective of this study is to determine the odds on the decision
of freshmen college students to pursue their enrolled course. Specifically, it aims
to determine the variables that affect freshmen college students in their decision to
pursue their enrolled course.
Importance of the Study
The result of this study would serve as a guideline to identify the factors
that affect the decision of freshmen college students to pursue their enrolled
course and provide insight on the importance of self - evaluation in making
decisions in life. For high school students, it would serve as an enlightment for
them to take their studies seriously. It could also give information to parents in
order for them to guide and support their children in their studies.
Furthermore, the result of the study could guide teachers, school
administrators and guidance counselors in structuring programs or activities for
the incoming freshmen college students to help them in choosing their course as
well as choosing the college. It could also serve as basis or reference for future
related studies.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
4
Scope and Delimitation of the Study
The study was conducted at the Benguet State University campus. The
respondents were the freshmen college students and they were selected randomly
from the different courses of Benguet State University.
The number of respondents was determined using stratified random
sampling technique. The study was conducted from November 2007 to March
2008.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
5
REVIEW OF RELATED LITERATURE
Application of Logistic Regression Analysis
Logistic regression analysis is a type of predictive model that can be used
when the target variable is a categorical variable with two categories for example:
live / die, has disease/doesn’t have disease, purchase product / doesn’t purchase,
wins race / doesn’t win, etc. A logistic regression model does not involve decision
trees and is more a kin to nonlinear regression such as fitting a polynomial to a set
of data values. Logistic regression can be used only with two types of target
variables; 1) a categorical target variable that has exactly two categories (i.e. a
binary or dichotomous variable) 2) a continuous target variable that has values in
the range 0.0 to 1.0 representing probability values or proportions
(Http:www.dtreg.com/logistic.htm)
Logistic regression is a part of a categorical model called generalized
linear models. This broad class of models includes ordinary regression and
ANOVA, as well as Multivariate Statistical Technique like MANOVA, ANOVA
and Loglinear regression. An excellent generalized model is presented by Agresti
(1996).
The logistic regression model can be extended to more factorial designs.
Graphical models can be used in checking the logistic regression and the local
mean deviant plots are used for detecting the over all lack of fit mentioned by
Elizon Rogel (1998).
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
6
Comparing logistic regression model and discriminant analysis, it was
found that logistic regression model is more preferred because the simple model
for P (X/Y) is ambiguous when the number of explanatory variables are moderate
or large and that the parameters to be estimated are only in the conditional
contributions of y given x in logistic regression as cited by Rubin (1984).
Logistic regression allows one to predict a discrete outcome such as group
membership, from one set of variable that may be continuous, or a mix of any of
these. Generally, the dependent or response variable is dichotomous, such as
presence/absence or success/ failure. Discriminant Analysis is also used to predict
group membership with only two groups. However, Discriminant Analysis can
only be used with continuous independent variables. Thus, in instances where the
independent variables are categorical or a mix of continuous and categorical,
Logistic Regression is preferred.
Logistic Regression aims to correctly predict the category of outcome for
individual cases using the most parsimonious model.
To accomplish this goal, a model is created that includes all predictor
variables that are useful in predicting the response variable. Several different
options are available during the model creation. Variables can be entered into the
model specified by the researcher. The fitted model can be tested after each
coefficient is added or deleted in the procedure called stepwise regression.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
7
For the uses of logistic regression, this is used for the prediction of group
membership. Since logistic regression calculates the probability or success over
the probability of failure, the results of the analysis are in the form of odds ratio.
The logistic regression also provides knowledge of the relationship and
strengths among the variables.
Studies on Choosing and Enrolling a Course
Many students come to school uncertain about the field of study they wish
to pursue. Some students have a variety of career interests and need time to make
a decision regarding which college and major are most appropriate for them.
Others have narrowed their choices to two or three options but want to explore
these choices in greater depth before they make a firm commitment. Still other
students have no idea what they want to major in or what career opportunities are
available to them.
Understanding the development of matriculating students is essential to
those seeking to make a difference in these student’s lives. Vincent Tinto (1993)
outlined three stages students move through: separation, transition and
incorporation. Students first go through a separation stage in which they move
away from their home environment. Although this can be quite traumatic for
students, most eventually are able to move to the second stage, transition. During
this stage students are torn between their old environment and the new one; they
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
8
may not feel they belong in their old environments but have yet to find their
places in the new one. Finally, student’s move into incorporation when they have
achieved full membership into the social and academic communities of the
institution.
Gardner and Siegel (2001) cite data gathered by American College Testing
(ACT) indicating that 28 percent of students in public four year institutions fail to
continue beyond their first year in college. Because of this, and other factors,
interventions targeted to first year students have become important. Many
institutions have adopted programs designed to provide a “rite of passage” in
which students are welcomed, supported, celebrated, and eventually assimilated
into the campus (Gardner, 1986).
Butner, B., et.al. (2001), qualitatively examined the college choice process
for African American and Hispanic females at a large Southwestern University.
Identified, through the voices of these women three major themes that support
their decision to attend college: familial influences, the Quinter essential
American dream and striving to overcome
Amilbahar (2003) conducted a study to determine: 1) the relative strength
of the students across different intelligences; 2) if the socio-demographic
characteristics of the students are significantly treated to their chosen course; 3) if
the socio-demographic characteristics of the respondents influence their relative
strength in different intelligence and 4) if the relative strength of the respondents
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
9
in different intelligence influence their choice of course enrolled by them. The
study included a total of three hundred ninety three (393) first year college
students of the University of Southern Mindanao. Data were gathered using
questionnaire and the standardized Multiple Intelligence Checklist for Adults
(MICA).
The students were strongest in intrapersonal intelligence, followed by
naturalistic then interpersonal and linguistic intelligence. The last four where the
students showed less competence were bodily-kinesthetic, musical, spatial and
logical mathematical intelligence. Family monthly income has no significant
relation with the relative strengths of the students in different intelligences. Age,
gender, father’s occupation and mother’s occupation were found to have
significant influence on the relative strengths of the students in different
intelligences. Age, gender, family monthly income and parents’ occupation
showed significant relation with the chosen course at 5% level of significance.It
was also found that the relative strength of the students in bodily-kinesthetic,
intrapersonal, logical-mathematical and spatial intelligence influenced the choice
of course in which they enrolled.
Bouse, G. A. and Hossler, D. (1991) conducted five-year study of the
various choices students (n=4923) must make at each grade level in high school to
make solid decisions about which college or university to attend. Findings
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
10
revealed that parents played important role in formation of students’ educational
aspirations and in decisions about saving for a post-secondary education.
According to the study of Dr. Wafaa Guirgis (1998), all of the children of
educated women were enrolled in school; 91 percent of low-literate respondent’s
children in small families were enrolled, and 72 percent of low-literate
respondent’s children in large families were enrolled. In these small families, 5
percent had both boys and girls not enrolled, 3 percent had boys only not enrolled,
and 1 percent had girls only not enrolled. Sixteen percent of the large families
reported boys and girls not enrolled, 5 percent boys not enrolled, and 7 percent
girls not enrolled. Children were kept home because school was too expensive,
they had failed school, or they were needed at home or at work.
Erdmann, D. G. (1983), surveyed high school graduating seniors (N=401)
and guidance counselor, (N=536) concerning influences on college choice.
Results showed that important factors included availability of specific programs,
reputation, location and size, and counselor and parent recommendations.
Broekemier, G. M., (2002) surveyed adult college students regarding their
motivation for attending college and the relative importance of college choice
criteria. Getting better jobs, gaining general knowledge and enhancement of self-
esteem were most frequently mentioned. Availability of desired programs, days /
times of needed classes, locations of courses, cost and faculty reputations are the
most important choice criteria.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
11
A study of 740 first-year university students found a strong positive
correlation between students’ college choice, subsequent expectations of the
institution and intent to stay at or leave the university. It is suggested that process
occurring before matriculation are as significant as the college social and
academic environment or external forces in student retention (Villella, E. F., &
Hu, M., 1990).
Astin’s Involvement theory (1985) emphasized that students learn and
develop when they become active in the collegiate experience. Upcraft (1995)
expanded on this theory when he stated “The greater the quantity and quality of
involvement, the more likely the student will succeed in college”(p.18). College
personnel can help students become active in a number of ways; two successful
strategies are an activities carnival at the beginning of the year that introduces
students to different campus organizations and scheduling freshman planning
conference between students and their advisers.
Upcraft (1995) noted another common student development theory
attributed to Scholssberg, Lynch and Chickering. This theory concentrates on
students’ needs to feel they matter and are appreciated. College personnel must
realize that students need support form peers, faculty, staff and family if they are
to succeed. Support networks must be in place no freshman can begin to make the
important connections that will help them cope.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
12
THEORETICAL FRAMEWORK
Logistic Regression Model
Binomial (binary) logistic regression model is a form of regression that is
used when the dependent (Y) is a dichotomous variable and the independent
variable (Xi) are continuous or categorical variables or both. The data consist of N
vectors of joint observed values of the response variables and the explanatory
variables for n cases.
For example logistic regression model states that
Y = 1 (P, if decided to pursue their enrolled course)
Y = 0 (1 – P, if decided not to pursue their enrolled course)
Cox (1970) introduced logistic regression to describe the dependency of
binary variables on a set of continuous variable. Bioassay is a statistical model
under logistic regression model, which was introduced by Brekson (1953) in
connection with the analysis. The name of the model is derived from the fact that
the log transformation of the probability of Y = 1 is linear. The probability density
function for Y has its form.
L = P (Y) = Piyi ( 1 – Pi )1-Yi
(1)
Where: i = 1, 2, 3, … n
Pi = e ( Bo + ∑BiX )
(2)
1 + e ( Bo + ∑BiXjn )
Where: j = 1, 2, 3, … p
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
13
Wald Statistics
The Wald Statistics is the ratio of the unstandardized logit coefficient to its
standard error. It is used to test the significance of individual logistic regression
coefficients for each independent variable that is to test the null hypothesis in
logistic regression that a particular logit (effect) coefficient is zero.
Wald Statistic (χ2) = b2 / ASE 2
b
(3)
Where:
ASE 2
b = standard error of b
Estimation of Odds Ratio
Logit can be converted easily into a statement simply by using exponential
function (raising the natural “log e” to the b1 power). The original odd is
multiplied by e to the both power, where b is the logistic regression coefficient,
when the given independent variable increases one unit. The ratio of the Odd
ratios of the independent is the ratio of the relative importance of the variables in
terms of effect on the dependent variables odds.
b1=odds (∏ / 1-∏) x exp (b1) for any unit increased in X1
For X= Xj
(4)
∏’ (Xj ) Xj ) bo + b1 ( Xj + 1 ) (b)+ b1 ( Xj + 1 ) (b)
(5)
∏’ = log e ( ∏ / 1 - ∏ )
∏ = 1 – e (bo + b1x)
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
14
Therefore
∏’ = log e ( e (bo + b1x) / 1 – e (bo + b1x) )
(6)
B1 = log e ( ∏ / 1 - ∏ )
= log e ( odds 2) log e ( odds 1)
Where: odds 2 / odds 1 = e b1
Weighted Least Squares Estimation
We use the logistic function;
( Y ) = P = e (bo + b1x) / 1 + e (bo
+
b1x)
(10)
Assuming that when E ( Y ) approach 0 and 1 gradually, and 0 ≤ 1 ; where P is
decided on the enrolled course. This function best describes the odds of the
decision of Benguet State University freshmen college students in their enrolled
course using logistic regression analysis.
Goodness of Fit Test
Goodness of fit also known as Hosmer and Lemeshow’s goodness, is an
alternative to model chi – square ( 2
Χ ) for assessing the significance of a logistic
regression model. Assuming that the test statistics Bo is thus goodness of fit tests
statistics. The regression model is rejected as a description of the data if the level
of significance, P = ( Bo > Bo ). Computed in an x2 distribution with I – (P + 1)
degrees of freedom is sufficiently small. Determining its significance,
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
15
Ho : E (Y) = e (bo + b1x) / 1 + e (bo + b1x)
Chi – square (χ2) is used to test the statistics and it is in the form:
(
χ2 ) = ∑∑ ( Ojx = Ejk )2 / Ejk,
(7)
where:
Ejk
=
∑∏i = ∏ ( 1 - ∏i) = nj – Ej, from these equation we
can test its statistic by
χ2 ~ χ2c2
(8)
where: c = number of classes
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
16
Definition of Terms
Choice of Concentration . It is a particular course chosen by a student.
Degree Offering. Refers to the degree programs offered by the different
colleges. at the Benguet State University. The different degree offerings are as
follows: Bachelor of Science in Agriculture(BSA), Bachelor of Science in
Agriculture Education(BSAE), Bachelor of Science in Agri – Business(BSAB),
Bachelor of Science in Agricultural Engineering(BSAEng), Bachelor of Science
in Forestry(BSF), Bachelor of Science in Nutrition Dietetics(BSND), Bachelor of
Science in Home Economics(BSHE), Bachelor of Science in Elementary
Education(BEE), Bachelor of Science in Secondary Education(BSE), Bachelor of
Science in Nursing(BSN), Bachelor of Science in Applied Statistics(BSAS),
Bachelor of Science in Environmental Science(BSES), Bachelor of Science in
Information Technology(BSIT) and Doctor of Veterinary Medicine(DVM).
Factor. It is consist of a set of variables that are interrelated with one
another.
Freshmen. The students on the first level of college.
Logistic. It is a popular choice for modeling a dichotomous regression
variable
Odds Ratio. It is equal to exponential function of the slope.
Random. It is a process by which sampling is obtained without aiming for
specific individual object or condition.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
17
Regression. It refers to a function that yields the main value of a random
variable under the condition that one or more independent variables have special
values.
Regression Analysis. It is concerned with the study of the dependence of
one variable, the dependent variable, on one or more other variables, the
explanatory variables with a view to estimating and/or predicting the mean or
average value of the former in terms of unknown of the fixed values of the latter.
Forward stepwise regression. It is a regression used in adding or deleting
of independent variables depending on the marginal contribution of the said
variables.
Variables. It is a characteristic of interest measurable and observable in
every subject in the study.
Wald Statistics. It is commonly used to test the significance of individual
logistics coefficients for each independent variable.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
18
METHODOLOGY
Respondents of the Study
The respondents of the study college freshmen college students from the
different degree programs offered at Benguet State University. This study was
conducted at Benguet State University from November 2007 to March 2008.
Data Collection
Data were gathered from the Benguet State University freshmen college
students. A structural survey questionnaire was administered and floated to the
respondents according to the number of enrollees per course. From the economic
and demographic profile of the respondents, 10 variables were considered. The
independent variables were as follows; gender, age, weekly allowance, average
grade in high school, family monthly income, family size, number of siblings in
the family that are college graduate, working vs. non-working student, with
scholarship vs. without scholarship, financially dependent to parents vs.
financially independent to parents. And the dependent variable is the decision of
freshmen college students to pursue their enrolled course.
Data Gathered
Table 1 shows the variables observed and measured in the study. The
dependent variable was the decision of freshmen college students to pursue (1) or
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
19
not to pursue (0) their enrolled course. Independent variables included were the
possible factors that may contribute to the respondents’ decision to pursue or not
to pursue their enrolled course which includes: gender, age, weekly allowance,
average grade in high school, family monthly income, family size, number of
siblings in the family who are college graduates, working vs. non-working
student, with scholarship vs. without scholarship and financially dependent vs.
financially independent to parents.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
20
Table 1. The variables observed and measured in the study
Variable
Description
Observed
Value
Decision
Decided
to
Pursue
1
Decided
not
to
Pursue
0
Gender
Male
1
Female
0
Age
Weekly Allowance
Average Grade in High School
Family Monthly Income
5,000 & below
1
5,000-10,000
2
10,000-20,000
3
20,000
&
above
4
Family Size
Number of siblings who are college graduate
None
0
1-3
1
4 & above
2
Working Student vs. Non-working Student
Working
Student
1
Non-working
Student
2
With Scholarship vs. Without Scholarship
With
Scholarship
1
Without
Scholarship
2
Financially Dependent vs. Financially Independent
Financially
Dependent
1
Financially
Independent
2
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
21
Data Analysis
The collected data were encoded in the computer and analyzed using
Statistical Packages for Social Sciences (SPSS). Logistic regression techniques
was employed on the data gathered from the respondents treating the decision of
freshmen college students to pursue their enrolled course as the dependent
variable and the socio-economic profile as the independent variables which
include gender, age, weekly allowance, average grade in high school, family
monthly income, family size, number of siblings in the family who are college
graduates, working vs. non-working student, with scholarship vs. without
scholarship and financially dependent vs. financially independent to parents.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
22
RESULTS AND DISCUSSION
The population of freshmen college students from the selected courses of
Benguet State University school year 2007-2008 is shown in Table 2. Obviously,
the population of freshmen college students during the first semester decreased
during the second semester. This shows that some students were not able to
continue their studies in the university. Some might have transferred in other
universities, shifted to other course or dropped from school. With this, the study
aims to determine the factors that affect freshmen college students to pursue their
enrolled course.
Decision of Freshmen College Students
to Pursue Their Enrolled Course
Table 3 shows the descriptive statistics of the independent variables by the
dependent variable and total number of respondents. Majority of the age groups
are decided to pursue their enrolled course. Age ranging from 17-18 proves that
half of the respondents are decided to pursue their enrolled course having a
percentage of 45 percent. All male respondents are decided to pursue their
enrolled course while 27.16 percent females are decided not to pursue their
enrolled course. It also shows that more students have low weekly allowance and
they have a higher frequency of deciding not to pursue their enrolled course. For
the family monthly income, more students belong to low income families and
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
23
students from this families have a more frequency of deciding not to pursue their
enrolled course.
Respondents from a family size of 9-11 have the highest
percentage on the decision, not to pursue their enrolled course which is 36.36
percent and respondents having a family size of 3-5 have the lowest percentage on
the decision, not to pursue their enrolled course. Therefore, students from a bigger
family size have a lesser chance not to pursue their enrolled course compared to
those students who belong to a smaller family size.
Half of the respondents have brothers or sisters who are college graduate
and almost all of them are decided to pursue their enrolled course.95 percent of
the respondents are non-working students and 20 of them are decided not to
pursue their enrolled course. 63 percent have no scholarship and 50 of them are
decided to pursue their enrolled course. 81 percent of the students are financially
dependent to their parents and 14 are decided not to pursue their enrolled course,
19 are financially independent and 7 are decided not to pursue their enrolled
course. From the total number of respondents, 78 percent are decided to pursue
their enrolled course as opposed to 22 percent of the population who answered
that they are decided not to pursue their enrolled course.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
24
400
364
350
316
300
S
257
NT
E250
214
TUD
S
F200
1ST SEM
179
O
R
166
2ND SEM
143
BE 150
138
NUM
96
97
100
85
95
77
84
83
73
69
76
76
62
50
62
45
50
50
50
43 32
36
26
25
0
BSA
M
BSABBSDCBSAS BSIT BSES BSAE BSF BSHEBSNDBSET BEE BSE BSN DV
COURSES
Figure 1. Enrollment of Freshmen College Students from Selected Courses of
Benguet State University SY: 2007-2008
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
25
The first year of college is the most trying time for many students; new
responsibilities and expectations can be overwhelming. For this reason, a large
percentage of students do not make it to their sophomore year. Gardner and Siegel
(2001) cite data gathered by American College Testing (ACT) indicating that 28
percent of students in public four year institutions fail to continue beyond their
first year in college. Because of this, and other factors, interventions targeted to
first year students have become important. Many institutions have adopted
programs designed to provide a “rite of passage” in which students are welcomed,
supported, celebrated, and eventually assimilated into the campus (Gardner,
1986).
Table 2. Summary of the Independent Variable by the Dependent Variable
Variables Decided
Percentage
Decided
Percentage Total
to pursue
(%)
not to (%)
Frequenc
pursue
y
Gender
Male
19
100
0
0
19
Female
59
72.84
22
27.16
81
Age
15-16
28
75.68
9
24.32
37
17-18
42
79.25
11
20.75
53
19-20
8
80
2
20
10
Weekly
Allowance
200 & below
25
64.10
14
35.90
39
201-300
18
90
2
10
20
301-400
14
77.78
4
22.22
18
401-500
15
93.75
1
6.25
16
501 & above
6
85.71
1
14.29
7
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
26
Continuation…
Average Grade 11
61.11
5
27.78
18
in High School 34
82.93
7
17.07
41
80 & below
24
70.59
10
29.41
34
81-85
7
100
0
0
7
86-90
91 & above
Family
Monthly
Income
1
26
72.22
10
27.78
36
2
35
77.78
10
22.22
45
3
14
93.33
1
6
15
4
3
75
1
25
4
Family Size
3-5
30
83.33
6
17.65
36
6-8
29
82.86
6
17.14
35
9-11
14
63.64
8
36.36
22
12 & above
5
71.43
2
28.57
7
No. of Sibling
in the Family
who are
college
graduate
0
38
82.61
8
17.39
46
1
52
98.11
1
1.89
53
2
0
0
1
100
1
Working vs.
Non-working
Student
1
3
60
2
40
5
2
75
78.95
20
21.05
95
With
Scholarship vs.
Without
Scholarship
1
26
70.27
9
24.32
37
2
50
79.37
13
20.63
63
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
27
Continuation…
Financially
Dependent vs.
Financially
67
79.01
14
17.28
81
Independent
12
63.16
7
36.84
19
1
2
Total Number
of Respondents 78
22
100
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
28
Forward Stepwise Logistic Regression Result
In identifying the factors that contribute to the decision of freshmen
college students to pursue their enrolled course, the data were analyzed prior to
the application of forward stepwise logistic regression analysis. The response on
the question, “are you decided to pursue your enrolled course” was used as the
dependent variable, wherein 78 percent of the respondents answered “yes”.
The results were summarized in Table 3 in which there are two models
identified.
Model 1 contains the variable gender having a Wald statistics of 0.129,
which was significant at 5 percent level of significance. The model has a
goodness of fit with step chi-square as well as the model chi-square and block chi-
square of 10.632. This model accounted for 78 percent correct percentage
prediction with Nagelkerke (R2) of 15.5 percent of the explained variation on
students’ decision to pursue their enrolled course. The odds ratio of 0.000
indicates the probability of students who are decided to pursue their enrolled
course over the probability of students who are decided not to pursue their
enrolled course.
Butner, B., et.al. (2001), qualitatively examined the college choice process
for African American and Hispanic females at a large Southwestern University.
Identified, through the voices of these women three major themes that support
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
29
their decision to attend college: familial influences, the Quinter essential
American dream and striving to overcome.
Broekemier, G. M., & Seshadri, S. (1999), compared the importance
ratings of various college choice criteria by 395 college found high school seniors
and their parents. Identified significant variables are student/ parent and student
gender differences. Safety and academic issues were more important to females
than males. Students identified social life, friends attending and athletic programs
as more important than did parents.
Model 2 includes both gender and family size of respondents. It had a
Wald statistics of 0.142, which was significant at 5 percent level of significance.
The goodness of fit of this model slightly increased, the step chi-square decreased
to 6.231 with model and block chi-square of 16.864. This model accounts 79
percent correct prediction by Nagelkerke (R2) of 23.8 percent explained variation
on the students’ decision in pursuing their enrolled course. The odds ratio of
0.786 indicates the proportionate change in the odds after the first step and before
the second step. It means that the odds of students who are decided to pursue their
enrolled course is 0.786 times greater than for students who are decided not to
pursue their enrolled course.
According to the study of Dr. Wafaa Guirgis (1998), all of the children of
educated women were enrolled in school; 91 percent of low-literate respondent’s
children in small families were enrolled, and 72 percent of low-literate
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
30
respondent’s children in large families were enrolled. In these small families, 5
percent had both boys and girls not enrolled, 3 percent had boys only not enrolled,
and 1 percent had girls only not enrolled. Sixteen percent of the large families
reported boys and girls not enrolled, 5 percent boys not enrolled, and 7 percent
girls not enrolled. Children were kept home because school was too expensive,
they had failed school, or they were needed at home or at work.
Table 3. Factors affecting the decision of freshmen college students
MODEL
1
MODEL2
VARIABLES
Coef Wald Odds Coef Wald Odds
(S.E.) Stat Ratio (S.E.) Stat Ratio
Sig.
Sig.
Constant
9.203 22.857
11.046 0.251
Gender
-8.216 0.129 0.000 -8.300 0.142 0.000
Family size
-0.241 5.784 0.786
Step- Chi-square[df] 10.632 [1]
6.231 [1]
Model Chi-square
10.632 [1]
16.864 [2]
Block Chi-square
10.632 [1]
16.864 [2]
% Correct Percentage
78.0
79.0
Nagelkerke (R2)
0.155
0.238
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
31
Classification plot-results
Figure 2 displays the histogram of the model 1 which contains the
variable, gender. This shows the predicted probability of a students’ decision to
pursue or not to pursue their enrolled course. Gender is dichotomous and so there
are only two columns of cases on the plot. All the students who are decided to
pursue their enrolled course are on the right side and there are only two
misclassified cases.
Figure2: Step number: 1
Observed Groups and Predicted Probabilities
160
F
R 120
E
Q
U
E 80 1
N 1
C 1
Y 1
40 1
1
0 1
0 1
Predicted
Prob: 0 .25 .5 .75 1
Group: 000000000000000000000000000000111111111111111111111111111111
Predicted Probability is of Membership for decided to pursue
The Cut Value is .50
Symbols: 0 - decided not to pursue
1 - decided to pursue
Each Symbol Represents 10 Cases.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
32
Figure 3 shows the histogram of the second model which is the family
size. The cases are spread out across many columns but more of the cases cluster
at the right side of the plot. It shows that the outcome did actually occur (the
students are decided to pursue their enrolled course) and that the probability of the
event is high. For the misclassified cases, there are two 1’s on the left side and 15
0’s in the right side.
Figure 3: Step number: 2
Observed Groups and Predicted Probabilities
20
1
1
F 1
R 15 1 1
E 1 1
Q 1 1 1
U 1 1 1
E 10 1 1 1 1 1
N 1 1 1 1 1 1
C 1 1 1 1 1 1 1
Y 1 1 1 1 1 1 1
5 1 0 1 1 1 1 1 1
1 0 1 1 1 1 11 1
1 1 0 1 0 0 1 1 0 01 1
0 0 0 0 1 0 0 0 0 0 00 1
Predicted
Prob: 0 .25 .5 .75 1
Group: 000000000000000000000000000000111111111111111111111111111111
Predicted Probability is of Membership for decided to pursue
The Cut Value is .50
Symbols: 0 - decided not to pursue
1 - decided to pursue
Each Symbol Represents 1.25 Cases.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
33
SUMMARY, CONCLUSION AND RECOMMENDATION
Summary
Some students don’t have a clear vision regarding their reason for being in
college. Perhaps they are here at their parent’s insistence while not feeling that
they are doing what they really want to do with their life. Sometimes courses and
majors are chosen to please others, but have little or no relationship to the
student’s true interests. Many students just aren’t sure about what they really want
to do in their future career. It takes a fairly clear purpose to motivate a student to
successfully engage in the lengthy and difficult process of higher education. The
objective of this study was to determine the odds on the decision of freshmen
college students to pursue their enrolled course, and to determine the factors that
affect freshmen college students to pursue their enrolled course using logistic
regression analysis. Questionnaires had been floated to the respondents for the
data gathering.
Conclusions
Based on the result of the study, using the statistical tool called logistic
regression, the researchers found out that 78 percent of the students are decided to
pursue their enrolled course and 22 percent are decided not to pursue their
enrolled course. The variables considered that were assumed to affect the decision
of freshmen college students to pursue their enrolled course are: age, gender,
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
34
weekly allowance, average grade, family monthly income, family size, number of
siblings in the family who are college graduate, present residence, working
student or nonworking student, with or without scholarship and financially
dependent or financially dependent to parents. From the 10 independent variables,
the result showed that only the variables gender, family size and scholarship have
a significant effect on the decision of freshmen college students to pursue their
enrolled course.
Recommendations
It is therefore recommended that;
1. Parents must guide and support their children in every action or
decision they make.
2. The Benguet State University should implement lectures on college and
course choice to incoming freshmen college students who are interested in
studying at the institution before or after the Entrance Qualifying Examination for
freshmen students.
3. Other organizations of the Benguet State University should be united
and have a common stand of students’ welfare especially in their decision to
pursue a course.
4. Students must evaluate or examine their selves on what field of
specialization they are interested in before entering college.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
35
5. The result of this study maybe a great help to Benguet State University
guidance and counseling unit and other government and non-government
organization. It maybe of help in counseling and fostering students awareness on
having a direction to achieve success in the future. This would guide them in
structuring activities or programs to divert students’ attention to things that would
enhance their personalities and their decision making for their future. With all the
results obtained, further study on the decision of freshmen college students is
recommended.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
36
LITERATURE CITED
Amilbahar, R. G. (2003). The Relation of Multiple Intelligence and the Chosen
Course of University of Southern Mindanao Students. 84 pp.
Bers, T. H. & Galowich, P. M. 2003. Parents and the College Choice Process for
Community College, 10 (1) 24-40.
Butner, B., et. al. (2001). The college choice process of African American and
Hispanic women: Implications for college transitions. Journal of College
Orrientation and Transition, 9(1) 24-32.
Bouse, G. A. & Hossler, D. (1991). Studying college Choice. A Progress Report.
Journal of College Admission, (130) 11-16.
Broekemier, G. M. (2002). A comparison of two-year and four year adult
students: Motivations to attend college and the importance of choice
criteria. Journal of Marketing for Higher Education, 12(1) 31-48.
Erdmann, D. G. (1983). An examination of factors influencing student choice in
college selection process October 1, 2004, from ERIC database.
Villella, E. F., & Hu, M. (1990). College choice as a thinking variable between
recruitment and retention. Journal of Marketing for Higher Education, 3(1)
79-88.
Ragheb, S., & Guirgis, W. (1998). Family size and gender equity and child
bearing. Faculty of Nursing Alexandria University. Egypt.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
37
Apola, W. (2004). “Logistic Regression Analysis on the Number Coding Scheme
as a Solution to the Traffic Problems of Baguio City.” Unpublished
Undergraduate Thesis, CAS Benguet State University, La Trinidad, Benguet.
Begnaden, J. S. & Tickchap, J. C. (2006). “Logistic Regression Analysis on
the Willingness of College Student Smokers to Stop from Smoking.”
Unpublished Under Graduate Thesis, CAS Benguet State University, La
Trinidad Benguet.
Costes, G. & Emuslan, M. (2004). “Factor Analysis and Logistic Regression
Analysis on the Economic Crisis Management of College Students of
Benguet State University.” Unpublished Under Graduate Thesis, CAS
Benguet State University, La Trinidad Benguet.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
38
Appendix A: The Questionnaire
Benguet State University
College of Arts and Sciences
Math - Physics – Statistics Department
La Trinidad, Benguet
Dear Respondent,
As a requirement of the course, Bachelor of Science in Applied Statistics,
we are conducting our thesis; a research entitled “LOGISTIC REGRESSION
ANALYSIS ON THE DECISION OF FRESHMEN COLLEGE STUDENTS
TO PURSUE THEIR ENROLLED COURSE”. In this connection, we would
like to solicit your cooperation & assistance by answering the questionnaires
completely and honestly. The researchers assure that your answers will be kept
and treated with outmost confidentiality.
Thank you very much and more power.
Truly
yours,
Juliet N. Carias
Researcher
Meldrid A. Luisito
Researcher
Noted by:
CRISTINA B. OCDEN
Adviser
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
39
QUESTIONNAIRE
Direction: Please check or write on the blank that corresponds to your answers.
Name: (Optional) Gender: Civil Status:
Age:
Average Grade in High school:
Number of Family members living (Family size):
Number of brothers/sisters who graduated in college:
Weekly allowance (please indicate):
Income of Both Parents (monthly):
1. Less than Php5, 000.00
3. Php10, 000-20,000
2. Php5, 000-10,000
4. Above Php20, 000
Present Residence:
1. Dormitory
3. With family/parents
2. Boarding house
4.With relatives
Working student (1):
Non-working student (2):
With scholarship (1): Without scholarship (2):
Financially dependent to parents (1):
Financially independent to parents (2):
Are you determined to pursue your enrolled course?
( ) YES
( ) NO
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
CARIAS, JULIET N. and LUISITO, MELDRID A., April 2008. Logistic
Regression Analysis on the Decision of Freshmen College Students to Pursue Their
Enrolled Course. Benguet State University, La Trinidad, Benguet.
ADVISER: CRISTINA B. OCDEN
ABSTRACT
This study was conducted to determine the odds on the decision of freshmen
college students to pursue their enrolled course and to determine the variables that affect
the decision of freshmen college students to pursue their enrolled course.
The data were obtained from randomly selected freshmen college students
enrolled in degree programs of the Benguet State University with the aid of a
questionnaire as guide. The Logistic Regression Analysis was used to determine the
factors that affect freshmen college students whether to pursue or not to pursue their
enrolled course.
Based on the result of the study, 78 percent of the students are decided to pursue
their enrolled course as opposed to 22 percent students who are decided not to pursue
their enrolled course. The variables that affect significantly in pursuing their course are:
gender, family size, and scholarship.
With all the results obtained, further study on the decision of freshmen college
students is recommended to identify other factors that influence freshmen college
students to pursue or not to pursue their enrolled course.
TABLE OF CONTENTS
Pages
Bibliography……………………………………………………………… i
Abstract……………………………………………………………………. i
Table of Contents…………………………………………………………. ii
INTRODUCTION
Background of the Study …………………………………………. 1
Objectives of the Study …………………………………………... 3
Importance of the Problem ……………………………………… 3
Scope and Delimitation of the Study ……………………………. 4
REVIEW OF RELATED LITERATURE……………………………….. 5
THEORETICAL FRAMEWORK ……………………………………….
12
Definition of Terms ………………………………………………. 16
METHODOLOGY
Respondents of the study………………………………………… 18
Data Collection…………. ………………………………............ 18
Data Gathered………………………………………………........ 18
Data Analysis……………………………………………………. 21
RESULTS AND DISCUSSION……………………………………....... 22
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS............ 33
LITERATURE CITED………………………………………………….. 36
APPENDICES…………………………………………………………... 38
ii
1
INTRODUCTION
Background of the Study
Going to college is one of the most important decisions a high school
student can make. Contrary to popular belief, preparing for college is even more
important than simply going to college. How a student prepares for college has a
great influence whether or not the student will actually go to college. Many
students and parents underestimate the importance of college preparation. Due to
this miscalculation; students go to college without a clear choice of concentration
and the right choice of college.
Choice of concentration or course required inquiry and reasoned
judgement and some creative research on the part of students. Because choice of
concentration is a complex decision, it is not surprising that a number of students
change concentrations after first semester or after the first year of college.
Aside from choosing a course, choosing a college is one of the most
difficult and important decisions/choices a student will ever have to make. It is
important to select a college that first, has the major that is interests you and
second, has a good program. Even if students are not sure what course to take,
they must visit different colleges to have information on the courses offered.
According to a survey conducted at University of California, Los Angeles
(UCLA), the percentage of freshmen choosing their college because of low tuition
fee rose sharply to its second highest point ever, 31.3 percent compared to 27.7
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
2
percent in 1995. And those going elsewhere because they were not offered
financial aid by their first college reached a high 5.7 percent up from 5.2 percent
in 1989, when this question last was asked.
Benguet State University is one of the 81 State Universities and Colleges
(SUC’s) in the Philippines located at La Trinidad, Benguet where students,
Filipinos or foreigners choose to study. It is known for its popularity as the center
of excellence in agricultural education, offering high quality education and
assessing low cost tuition.
The courses offered are Bachelor of Science in Agriculture (BSA),
Bachelor of Science in Agricultural Engineering (BSAE), Bachelor of Science in
Applied Statistics (BSAS), Bachelor of Science in Environmental Science
(BSES), Bachelor of Science in Information Technology (BSIT), Bachelor of
Science in Nursing (BSN), Doctor of Veterinary Medicine (DVM), Bachelor of
Science in Elementary Education (BEE), Bachelor of Science in Home
Economics (BSHE), Bachelor of Science in Nutrition Dietetics (BSND), Bachelor
of Science in Forestry (BSF), Bachelor of Science in Agri-business (BSAB),
Diploma in Agro-Forestry (DAF), and a 6 months Automotive Technology.
Choice of college and choice of concentration or course go together.
Changing concentration and changing of college is a waste of time and money. To
avoid this, it is important to study the factors that affect the decision of freshmen
college students to pursue their enrolled course.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
3
Objectives of the Study
The main objective of this study is to determine the odds on the decision
of freshmen college students to pursue their enrolled course. Specifically, it aims
to determine the variables that affect freshmen college students in their decision to
pursue their enrolled course.
Importance of the Study
The result of this study would serve as a guideline to identify the factors
that affect the decision of freshmen college students to pursue their enrolled
course and provide insight on the importance of self - evaluation in making
decisions in life. For high school students, it would serve as an enlightment for
them to take their studies seriously. It could also give information to parents in
order for them to guide and support their children in their studies.
Furthermore, the result of the study could guide teachers, school
administrators and guidance counselors in structuring programs or activities for
the incoming freshmen college students to help them in choosing their course as
well as choosing the college. It could also serve as basis or reference for future
related studies.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
4
Scope and Delimitation of the Study
The study was conducted at the Benguet State University campus. The
respondents were the freshmen college students and they were selected randomly
from the different courses of Benguet State University.
The number of respondents was determined using stratified random
sampling technique. The study was conducted from November 2007 to March
2008.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
5
REVIEW OF RELATED LITERATURE
Application of Logistic Regression Analysis
Logistic regression analysis is a type of predictive model that can be used
when the target variable is a categorical variable with two categories for example:
live / die, has disease/doesn’t have disease, purchase product / doesn’t purchase,
wins race / doesn’t win, etc. A logistic regression model does not involve decision
trees and is more a kin to nonlinear regression such as fitting a polynomial to a set
of data values. Logistic regression can be used only with two types of target
variables; 1) a categorical target variable that has exactly two categories (i.e. a
binary or dichotomous variable) 2) a continuous target variable that has values in
the range 0.0 to 1.0 representing probability values or proportions
(Http:www.dtreg.com/logistic.htm)
Logistic regression is a part of a categorical model called generalized
linear models. This broad class of models includes ordinary regression and
ANOVA, as well as Multivariate Statistical Technique like MANOVA, ANOVA
and Loglinear regression. An excellent generalized model is presented by Agresti
(1996).
The logistic regression model can be extended to more factorial designs.
Graphical models can be used in checking the logistic regression and the local
mean deviant plots are used for detecting the over all lack of fit mentioned by
Elizon Rogel (1998).
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
6
Comparing logistic regression model and discriminant analysis, it was
found that logistic regression model is more preferred because the simple model
for P (X/Y) is ambiguous when the number of explanatory variables are moderate
or large and that the parameters to be estimated are only in the conditional
contributions of y given x in logistic regression as cited by Rubin (1984).
Logistic regression allows one to predict a discrete outcome such as group
membership, from one set of variable that may be continuous, or a mix of any of
these. Generally, the dependent or response variable is dichotomous, such as
presence/absence or success/ failure. Discriminant Analysis is also used to predict
group membership with only two groups. However, Discriminant Analysis can
only be used with continuous independent variables. Thus, in instances where the
independent variables are categorical or a mix of continuous and categorical,
Logistic Regression is preferred.
Logistic Regression aims to correctly predict the category of outcome for
individual cases using the most parsimonious model.
To accomplish this goal, a model is created that includes all predictor
variables that are useful in predicting the response variable. Several different
options are available during the model creation. Variables can be entered into the
model specified by the researcher. The fitted model can be tested after each
coefficient is added or deleted in the procedure called stepwise regression.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
7
For the uses of logistic regression, this is used for the prediction of group
membership. Since logistic regression calculates the probability or success over
the probability of failure, the results of the analysis are in the form of odds ratio.
The logistic regression also provides knowledge of the relationship and
strengths among the variables.
Studies on Choosing and Enrolling a Course
Many students come to school uncertain about the field of study they wish
to pursue. Some students have a variety of career interests and need time to make
a decision regarding which college and major are most appropriate for them.
Others have narrowed their choices to two or three options but want to explore
these choices in greater depth before they make a firm commitment. Still other
students have no idea what they want to major in or what career opportunities are
available to them.
Understanding the development of matriculating students is essential to
those seeking to make a difference in these student’s lives. Vincent Tinto (1993)
outlined three stages students move through: separation, transition and
incorporation. Students first go through a separation stage in which they move
away from their home environment. Although this can be quite traumatic for
students, most eventually are able to move to the second stage, transition. During
this stage students are torn between their old environment and the new one; they
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
8
may not feel they belong in their old environments but have yet to find their
places in the new one. Finally, student’s move into incorporation when they have
achieved full membership into the social and academic communities of the
institution.
Gardner and Siegel (2001) cite data gathered by American College Testing
(ACT) indicating that 28 percent of students in public four year institutions fail to
continue beyond their first year in college. Because of this, and other factors,
interventions targeted to first year students have become important. Many
institutions have adopted programs designed to provide a “rite of passage” in
which students are welcomed, supported, celebrated, and eventually assimilated
into the campus (Gardner, 1986).
Butner, B., et.al. (2001), qualitatively examined the college choice process
for African American and Hispanic females at a large Southwestern University.
Identified, through the voices of these women three major themes that support
their decision to attend college: familial influences, the Quinter essential
American dream and striving to overcome
Amilbahar (2003) conducted a study to determine: 1) the relative strength
of the students across different intelligences; 2) if the socio-demographic
characteristics of the students are significantly treated to their chosen course; 3) if
the socio-demographic characteristics of the respondents influence their relative
strength in different intelligence and 4) if the relative strength of the respondents
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
9
in different intelligence influence their choice of course enrolled by them. The
study included a total of three hundred ninety three (393) first year college
students of the University of Southern Mindanao. Data were gathered using
questionnaire and the standardized Multiple Intelligence Checklist for Adults
(MICA).
The students were strongest in intrapersonal intelligence, followed by
naturalistic then interpersonal and linguistic intelligence. The last four where the
students showed less competence were bodily-kinesthetic, musical, spatial and
logical mathematical intelligence. Family monthly income has no significant
relation with the relative strengths of the students in different intelligences. Age,
gender, father’s occupation and mother’s occupation were found to have
significant influence on the relative strengths of the students in different
intelligences. Age, gender, family monthly income and parents’ occupation
showed significant relation with the chosen course at 5% level of significance.It
was also found that the relative strength of the students in bodily-kinesthetic,
intrapersonal, logical-mathematical and spatial intelligence influenced the choice
of course in which they enrolled.
Bouse, G. A. and Hossler, D. (1991) conducted five-year study of the
various choices students (n=4923) must make at each grade level in high school to
make solid decisions about which college or university to attend. Findings
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
10
revealed that parents played important role in formation of students’ educational
aspirations and in decisions about saving for a post-secondary education.
According to the study of Dr. Wafaa Guirgis (1998), all of the children of
educated women were enrolled in school; 91 percent of low-literate respondent’s
children in small families were enrolled, and 72 percent of low-literate
respondent’s children in large families were enrolled. In these small families, 5
percent had both boys and girls not enrolled, 3 percent had boys only not enrolled,
and 1 percent had girls only not enrolled. Sixteen percent of the large families
reported boys and girls not enrolled, 5 percent boys not enrolled, and 7 percent
girls not enrolled. Children were kept home because school was too expensive,
they had failed school, or they were needed at home or at work.
Erdmann, D. G. (1983), surveyed high school graduating seniors (N=401)
and guidance counselor, (N=536) concerning influences on college choice.
Results showed that important factors included availability of specific programs,
reputation, location and size, and counselor and parent recommendations.
Broekemier, G. M., (2002) surveyed adult college students regarding their
motivation for attending college and the relative importance of college choice
criteria. Getting better jobs, gaining general knowledge and enhancement of self-
esteem were most frequently mentioned. Availability of desired programs, days /
times of needed classes, locations of courses, cost and faculty reputations are the
most important choice criteria.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
11
A study of 740 first-year university students found a strong positive
correlation between students’ college choice, subsequent expectations of the
institution and intent to stay at or leave the university. It is suggested that process
occurring before matriculation are as significant as the college social and
academic environment or external forces in student retention (Villella, E. F., &
Hu, M., 1990).
Astin’s Involvement theory (1985) emphasized that students learn and
develop when they become active in the collegiate experience. Upcraft (1995)
expanded on this theory when he stated “The greater the quantity and quality of
involvement, the more likely the student will succeed in college”(p.18). College
personnel can help students become active in a number of ways; two successful
strategies are an activities carnival at the beginning of the year that introduces
students to different campus organizations and scheduling freshman planning
conference between students and their advisers.
Upcraft (1995) noted another common student development theory
attributed to Scholssberg, Lynch and Chickering. This theory concentrates on
students’ needs to feel they matter and are appreciated. College personnel must
realize that students need support form peers, faculty, staff and family if they are
to succeed. Support networks must be in place no freshman can begin to make the
important connections that will help them cope.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
12
THEORETICAL FRAMEWORK
Logistic Regression Model
Binomial (binary) logistic regression model is a form of regression that is
used when the dependent (Y) is a dichotomous variable and the independent
variable (Xi) are continuous or categorical variables or both. The data consist of N
vectors of joint observed values of the response variables and the explanatory
variables for n cases.
For example logistic regression model states that
Y = 1 (P, if decided to pursue their enrolled course)
Y = 0 (1 – P, if decided not to pursue their enrolled course)
Cox (1970) introduced logistic regression to describe the dependency of
binary variables on a set of continuous variable. Bioassay is a statistical model
under logistic regression model, which was introduced by Brekson (1953) in
connection with the analysis. The name of the model is derived from the fact that
the log transformation of the probability of Y = 1 is linear. The probability density
function for Y has its form.
L = P (Y) = Piyi ( 1 – Pi )1-Yi
(1)
Where: i = 1, 2, 3, … n
Pi = e ( Bo + ∑BiX )
(2)
1 + e ( Bo + ∑BiXjn )
Where: j = 1, 2, 3, … p
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
13
Wald Statistics
The Wald Statistics is the ratio of the unstandardized logit coefficient to its
standard error. It is used to test the significance of individual logistic regression
coefficients for each independent variable that is to test the null hypothesis in
logistic regression that a particular logit (effect) coefficient is zero.
Wald Statistic (χ2) = b2 / ASE 2
b
(3)
Where:
ASE 2
b = standard error of b
Estimation of Odds Ratio
Logit can be converted easily into a statement simply by using exponential
function (raising the natural “log e” to the b1 power). The original odd is
multiplied by e to the both power, where b is the logistic regression coefficient,
when the given independent variable increases one unit. The ratio of the Odd
ratios of the independent is the ratio of the relative importance of the variables in
terms of effect on the dependent variables odds.
b1=odds (∏ / 1-∏) x exp (b1) for any unit increased in X1
For X= Xj
(4)
∏’ (Xj ) Xj ) bo + b1 ( Xj + 1 ) (b)+ b1 ( Xj + 1 ) (b)
(5)
∏’ = log e ( ∏ / 1 - ∏ )
∏ = 1 – e (bo + b1x)
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
14
Therefore
∏’ = log e ( e (bo + b1x) / 1 – e (bo + b1x) )
(6)
B1 = log e ( ∏ / 1 - ∏ )
= log e ( odds 2) log e ( odds 1)
Where: odds 2 / odds 1 = e b1
Weighted Least Squares Estimation
We use the logistic function;
( Y ) = P = e (bo + b1x) / 1 + e (bo
+
b1x)
(10)
Assuming that when E ( Y ) approach 0 and 1 gradually, and 0 ≤ 1 ; where P is
decided on the enrolled course. This function best describes the odds of the
decision of Benguet State University freshmen college students in their enrolled
course using logistic regression analysis.
Goodness of Fit Test
Goodness of fit also known as Hosmer and Lemeshow’s goodness, is an
alternative to model chi – square ( 2
Χ ) for assessing the significance of a logistic
regression model. Assuming that the test statistics Bo is thus goodness of fit tests
statistics. The regression model is rejected as a description of the data if the level
of significance, P = ( Bo > Bo ). Computed in an x2 distribution with I – (P + 1)
degrees of freedom is sufficiently small. Determining its significance,
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
15
Ho : E (Y) = e (bo + b1x) / 1 + e (bo + b1x)
Chi – square (χ2) is used to test the statistics and it is in the form:
(
χ2 ) = ∑∑ ( Ojx = Ejk )2 / Ejk,
(7)
where:
Ejk
=
∑∏i = ∏ ( 1 - ∏i) = nj – Ej, from these equation we
can test its statistic by
χ2 ~ χ2c2
(8)
where: c = number of classes
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
16
Definition of Terms
Choice of Concentration . It is a particular course chosen by a student.
Degree Offering. Refers to the degree programs offered by the different
colleges. at the Benguet State University. The different degree offerings are as
follows: Bachelor of Science in Agriculture(BSA), Bachelor of Science in
Agriculture Education(BSAE), Bachelor of Science in Agri – Business(BSAB),
Bachelor of Science in Agricultural Engineering(BSAEng), Bachelor of Science
in Forestry(BSF), Bachelor of Science in Nutrition Dietetics(BSND), Bachelor of
Science in Home Economics(BSHE), Bachelor of Science in Elementary
Education(BEE), Bachelor of Science in Secondary Education(BSE), Bachelor of
Science in Nursing(BSN), Bachelor of Science in Applied Statistics(BSAS),
Bachelor of Science in Environmental Science(BSES), Bachelor of Science in
Information Technology(BSIT) and Doctor of Veterinary Medicine(DVM).
Factor. It is consist of a set of variables that are interrelated with one
another.
Freshmen. The students on the first level of college.
Logistic. It is a popular choice for modeling a dichotomous regression
variable
Odds Ratio. It is equal to exponential function of the slope.
Random. It is a process by which sampling is obtained without aiming for
specific individual object or condition.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
17
Regression. It refers to a function that yields the main value of a random
variable under the condition that one or more independent variables have special
values.
Regression Analysis. It is concerned with the study of the dependence of
one variable, the dependent variable, on one or more other variables, the
explanatory variables with a view to estimating and/or predicting the mean or
average value of the former in terms of unknown of the fixed values of the latter.
Forward stepwise regression. It is a regression used in adding or deleting
of independent variables depending on the marginal contribution of the said
variables.
Variables. It is a characteristic of interest measurable and observable in
every subject in the study.
Wald Statistics. It is commonly used to test the significance of individual
logistics coefficients for each independent variable.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
18
METHODOLOGY
Respondents of the Study
The respondents of the study college freshmen college students from the
different degree programs offered at Benguet State University. This study was
conducted at Benguet State University from November 2007 to March 2008.
Data Collection
Data were gathered from the Benguet State University freshmen college
students. A structural survey questionnaire was administered and floated to the
respondents according to the number of enrollees per course. From the economic
and demographic profile of the respondents, 10 variables were considered. The
independent variables were as follows; gender, age, weekly allowance, average
grade in high school, family monthly income, family size, number of siblings in
the family that are college graduate, working vs. non-working student, with
scholarship vs. without scholarship, financially dependent to parents vs.
financially independent to parents. And the dependent variable is the decision of
freshmen college students to pursue their enrolled course.
Data Gathered
Table 1 shows the variables observed and measured in the study. The
dependent variable was the decision of freshmen college students to pursue (1) or
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
19
not to pursue (0) their enrolled course. Independent variables included were the
possible factors that may contribute to the respondents’ decision to pursue or not
to pursue their enrolled course which includes: gender, age, weekly allowance,
average grade in high school, family monthly income, family size, number of
siblings in the family who are college graduates, working vs. non-working
student, with scholarship vs. without scholarship and financially dependent vs.
financially independent to parents.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
20
Table 1. The variables observed and measured in the study
Variable
Description
Observed
Value
Decision
Decided
to
Pursue
1
Decided
not
to
Pursue
0
Gender
Male
1
Female
0
Age
Weekly Allowance
Average Grade in High School
Family Monthly Income
5,000 & below
1
5,000-10,000
2
10,000-20,000
3
20,000
&
above
4
Family Size
Number of siblings who are college graduate
None
0
1-3
1
4 & above
2
Working Student vs. Non-working Student
Working
Student
1
Non-working
Student
2
With Scholarship vs. Without Scholarship
With
Scholarship
1
Without
Scholarship
2
Financially Dependent vs. Financially Independent
Financially
Dependent
1
Financially
Independent
2
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
21
Data Analysis
The collected data were encoded in the computer and analyzed using
Statistical Packages for Social Sciences (SPSS). Logistic regression techniques
was employed on the data gathered from the respondents treating the decision of
freshmen college students to pursue their enrolled course as the dependent
variable and the socio-economic profile as the independent variables which
include gender, age, weekly allowance, average grade in high school, family
monthly income, family size, number of siblings in the family who are college
graduates, working vs. non-working student, with scholarship vs. without
scholarship and financially dependent vs. financially independent to parents.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
22
RESULTS AND DISCUSSION
The population of freshmen college students from the selected courses of
Benguet State University school year 2007-2008 is shown in Table 2. Obviously,
the population of freshmen college students during the first semester decreased
during the second semester. This shows that some students were not able to
continue their studies in the university. Some might have transferred in other
universities, shifted to other course or dropped from school. With this, the study
aims to determine the factors that affect freshmen college students to pursue their
enrolled course.
Decision of Freshmen College Students
to Pursue Their Enrolled Course
Table 3 shows the descriptive statistics of the independent variables by the
dependent variable and total number of respondents. Majority of the age groups
are decided to pursue their enrolled course. Age ranging from 17-18 proves that
half of the respondents are decided to pursue their enrolled course having a
percentage of 45 percent. All male respondents are decided to pursue their
enrolled course while 27.16 percent females are decided not to pursue their
enrolled course. It also shows that more students have low weekly allowance and
they have a higher frequency of deciding not to pursue their enrolled course. For
the family monthly income, more students belong to low income families and
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
23
students from this families have a more frequency of deciding not to pursue their
enrolled course.
Respondents from a family size of 9-11 have the highest
percentage on the decision, not to pursue their enrolled course which is 36.36
percent and respondents having a family size of 3-5 have the lowest percentage on
the decision, not to pursue their enrolled course. Therefore, students from a bigger
family size have a lesser chance not to pursue their enrolled course compared to
those students who belong to a smaller family size.
Half of the respondents have brothers or sisters who are college graduate
and almost all of them are decided to pursue their enrolled course.95 percent of
the respondents are non-working students and 20 of them are decided not to
pursue their enrolled course. 63 percent have no scholarship and 50 of them are
decided to pursue their enrolled course. 81 percent of the students are financially
dependent to their parents and 14 are decided not to pursue their enrolled course,
19 are financially independent and 7 are decided not to pursue their enrolled
course. From the total number of respondents, 78 percent are decided to pursue
their enrolled course as opposed to 22 percent of the population who answered
that they are decided not to pursue their enrolled course.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
24
400
364
350
316
300
S
257
NT
E250
214
TUD
S
F200
1ST SEM
179
O
R
166
2ND SEM
143
BE 150
138
NUM
96
97
100
85
95
77
84
83
73
69
76
76
62
50
62
45
50
50
50
43 32
36
26
25
0
BSA
M
BSABBSDCBSAS BSIT BSES BSAE BSF BSHEBSNDBSET BEE BSE BSN DV
COURSES
Figure 1. Enrollment of Freshmen College Students from Selected Courses of
Benguet State University SY: 2007-2008
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
25
The first year of college is the most trying time for many students; new
responsibilities and expectations can be overwhelming. For this reason, a large
percentage of students do not make it to their sophomore year. Gardner and Siegel
(2001) cite data gathered by American College Testing (ACT) indicating that 28
percent of students in public four year institutions fail to continue beyond their
first year in college. Because of this, and other factors, interventions targeted to
first year students have become important. Many institutions have adopted
programs designed to provide a “rite of passage” in which students are welcomed,
supported, celebrated, and eventually assimilated into the campus (Gardner,
1986).
Table 2. Summary of the Independent Variable by the Dependent Variable
Variables Decided
Percentage
Decided
Percentage Total
to pursue
(%)
not to (%)
Frequenc
pursue
y
Gender
Male
19
100
0
0
19
Female
59
72.84
22
27.16
81
Age
15-16
28
75.68
9
24.32
37
17-18
42
79.25
11
20.75
53
19-20
8
80
2
20
10
Weekly
Allowance
200 & below
25
64.10
14
35.90
39
201-300
18
90
2
10
20
301-400
14
77.78
4
22.22
18
401-500
15
93.75
1
6.25
16
501 & above
6
85.71
1
14.29
7
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
26
Continuation…
Average Grade 11
61.11
5
27.78
18
in High School 34
82.93
7
17.07
41
80 & below
24
70.59
10
29.41
34
81-85
7
100
0
0
7
86-90
91 & above
Family
Monthly
Income
1
26
72.22
10
27.78
36
2
35
77.78
10
22.22
45
3
14
93.33
1
6
15
4
3
75
1
25
4
Family Size
3-5
30
83.33
6
17.65
36
6-8
29
82.86
6
17.14
35
9-11
14
63.64
8
36.36
22
12 & above
5
71.43
2
28.57
7
No. of Sibling
in the Family
who are
college
graduate
0
38
82.61
8
17.39
46
1
52
98.11
1
1.89
53
2
0
0
1
100
1
Working vs.
Non-working
Student
1
3
60
2
40
5
2
75
78.95
20
21.05
95
With
Scholarship vs.
Without
Scholarship
1
26
70.27
9
24.32
37
2
50
79.37
13
20.63
63
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
27
Continuation…
Financially
Dependent vs.
Financially
67
79.01
14
17.28
81
Independent
12
63.16
7
36.84
19
1
2
Total Number
of Respondents 78
22
100
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
28
Forward Stepwise Logistic Regression Result
In identifying the factors that contribute to the decision of freshmen
college students to pursue their enrolled course, the data were analyzed prior to
the application of forward stepwise logistic regression analysis. The response on
the question, “are you decided to pursue your enrolled course” was used as the
dependent variable, wherein 78 percent of the respondents answered “yes”.
The results were summarized in Table 3 in which there are two models
identified.
Model 1 contains the variable gender having a Wald statistics of 0.129,
which was significant at 5 percent level of significance. The model has a
goodness of fit with step chi-square as well as the model chi-square and block chi-
square of 10.632. This model accounted for 78 percent correct percentage
prediction with Nagelkerke (R2) of 15.5 percent of the explained variation on
students’ decision to pursue their enrolled course. The odds ratio of 0.000
indicates the probability of students who are decided to pursue their enrolled
course over the probability of students who are decided not to pursue their
enrolled course.
Butner, B., et.al. (2001), qualitatively examined the college choice process
for African American and Hispanic females at a large Southwestern University.
Identified, through the voices of these women three major themes that support
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
29
their decision to attend college: familial influences, the Quinter essential
American dream and striving to overcome.
Broekemier, G. M., & Seshadri, S. (1999), compared the importance
ratings of various college choice criteria by 395 college found high school seniors
and their parents. Identified significant variables are student/ parent and student
gender differences. Safety and academic issues were more important to females
than males. Students identified social life, friends attending and athletic programs
as more important than did parents.
Model 2 includes both gender and family size of respondents. It had a
Wald statistics of 0.142, which was significant at 5 percent level of significance.
The goodness of fit of this model slightly increased, the step chi-square decreased
to 6.231 with model and block chi-square of 16.864. This model accounts 79
percent correct prediction by Nagelkerke (R2) of 23.8 percent explained variation
on the students’ decision in pursuing their enrolled course. The odds ratio of
0.786 indicates the proportionate change in the odds after the first step and before
the second step. It means that the odds of students who are decided to pursue their
enrolled course is 0.786 times greater than for students who are decided not to
pursue their enrolled course.
According to the study of Dr. Wafaa Guirgis (1998), all of the children of
educated women were enrolled in school; 91 percent of low-literate respondent’s
children in small families were enrolled, and 72 percent of low-literate
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
30
respondent’s children in large families were enrolled. In these small families, 5
percent had both boys and girls not enrolled, 3 percent had boys only not enrolled,
and 1 percent had girls only not enrolled. Sixteen percent of the large families
reported boys and girls not enrolled, 5 percent boys not enrolled, and 7 percent
girls not enrolled. Children were kept home because school was too expensive,
they had failed school, or they were needed at home or at work.
Table 3. Factors affecting the decision of freshmen college students
MODEL
1
MODEL2
VARIABLES
Coef Wald Odds Coef Wald Odds
(S.E.) Stat Ratio (S.E.) Stat Ratio
Sig.
Sig.
Constant
9.203 22.857
11.046 0.251
Gender
-8.216 0.129 0.000 -8.300 0.142 0.000
Family size
-0.241 5.784 0.786
Step- Chi-square[df] 10.632 [1]
6.231 [1]
Model Chi-square
10.632 [1]
16.864 [2]
Block Chi-square
10.632 [1]
16.864 [2]
% Correct Percentage
78.0
79.0
Nagelkerke (R2)
0.155
0.238
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
31
Classification plot-results
Figure 2 displays the histogram of the model 1 which contains the
variable, gender. This shows the predicted probability of a students’ decision to
pursue or not to pursue their enrolled course. Gender is dichotomous and so there
are only two columns of cases on the plot. All the students who are decided to
pursue their enrolled course are on the right side and there are only two
misclassified cases.
Figure2: Step number: 1
Observed Groups and Predicted Probabilities
160
F
R 120
E
Q
U
E 80 1
N 1
C 1
Y 1
40 1
1
0 1
0 1
Predicted
Prob: 0 .25 .5 .75 1
Group: 000000000000000000000000000000111111111111111111111111111111
Predicted Probability is of Membership for decided to pursue
The Cut Value is .50
Symbols: 0 - decided not to pursue
1 - decided to pursue
Each Symbol Represents 10 Cases.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
32
Figure 3 shows the histogram of the second model which is the family
size. The cases are spread out across many columns but more of the cases cluster
at the right side of the plot. It shows that the outcome did actually occur (the
students are decided to pursue their enrolled course) and that the probability of the
event is high. For the misclassified cases, there are two 1’s on the left side and 15
0’s in the right side.
Figure 3: Step number: 2
Observed Groups and Predicted Probabilities
20
1
1
F 1
R 15 1 1
E 1 1
Q 1 1 1
U 1 1 1
E 10 1 1 1 1 1
N 1 1 1 1 1 1
C 1 1 1 1 1 1 1
Y 1 1 1 1 1 1 1
5 1 0 1 1 1 1 1 1
1 0 1 1 1 1 11 1
1 1 0 1 0 0 1 1 0 01 1
0 0 0 0 1 0 0 0 0 0 00 1
Predicted
Prob: 0 .25 .5 .75 1
Group: 000000000000000000000000000000111111111111111111111111111111
Predicted Probability is of Membership for decided to pursue
The Cut Value is .50
Symbols: 0 - decided not to pursue
1 - decided to pursue
Each Symbol Represents 1.25 Cases.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
33
SUMMARY, CONCLUSION AND RECOMMENDATION
Summary
Some students don’t have a clear vision regarding their reason for being in
college. Perhaps they are here at their parent’s insistence while not feeling that
they are doing what they really want to do with their life. Sometimes courses and
majors are chosen to please others, but have little or no relationship to the
student’s true interests. Many students just aren’t sure about what they really want
to do in their future career. It takes a fairly clear purpose to motivate a student to
successfully engage in the lengthy and difficult process of higher education. The
objective of this study was to determine the odds on the decision of freshmen
college students to pursue their enrolled course, and to determine the factors that
affect freshmen college students to pursue their enrolled course using logistic
regression analysis. Questionnaires had been floated to the respondents for the
data gathering.
Conclusions
Based on the result of the study, using the statistical tool called logistic
regression, the researchers found out that 78 percent of the students are decided to
pursue their enrolled course and 22 percent are decided not to pursue their
enrolled course. The variables considered that were assumed to affect the decision
of freshmen college students to pursue their enrolled course are: age, gender,
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
34
weekly allowance, average grade, family monthly income, family size, number of
siblings in the family who are college graduate, present residence, working
student or nonworking student, with or without scholarship and financially
dependent or financially dependent to parents. From the 10 independent variables,
the result showed that only the variables gender, family size and scholarship have
a significant effect on the decision of freshmen college students to pursue their
enrolled course.
Recommendations
It is therefore recommended that;
1. Parents must guide and support their children in every action or
decision they make.
2. The Benguet State University should implement lectures on college and
course choice to incoming freshmen college students who are interested in
studying at the institution before or after the Entrance Qualifying Examination for
freshmen students.
3. Other organizations of the Benguet State University should be united
and have a common stand of students’ welfare especially in their decision to
pursue a course.
4. Students must evaluate or examine their selves on what field of
specialization they are interested in before entering college.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
35
5. The result of this study maybe a great help to Benguet State University
guidance and counseling unit and other government and non-government
organization. It maybe of help in counseling and fostering students awareness on
having a direction to achieve success in the future. This would guide them in
structuring activities or programs to divert students’ attention to things that would
enhance their personalities and their decision making for their future. With all the
results obtained, further study on the decision of freshmen college students is
recommended.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
36
LITERATURE CITED
Amilbahar, R. G. (2003). The Relation of Multiple Intelligence and the Chosen
Course of University of Southern Mindanao Students. 84 pp.
Bers, T. H. & Galowich, P. M. 2003. Parents and the College Choice Process for
Community College, 10 (1) 24-40.
Butner, B., et. al. (2001). The college choice process of African American and
Hispanic women: Implications for college transitions. Journal of College
Orrientation and Transition, 9(1) 24-32.
Bouse, G. A. & Hossler, D. (1991). Studying college Choice. A Progress Report.
Journal of College Admission, (130) 11-16.
Broekemier, G. M. (2002). A comparison of two-year and four year adult
students: Motivations to attend college and the importance of choice
criteria. Journal of Marketing for Higher Education, 12(1) 31-48.
Erdmann, D. G. (1983). An examination of factors influencing student choice in
college selection process October 1, 2004, from ERIC database.
Villella, E. F., & Hu, M. (1990). College choice as a thinking variable between
recruitment and retention. Journal of Marketing for Higher Education, 3(1)
79-88.
Ragheb, S., & Guirgis, W. (1998). Family size and gender equity and child
bearing. Faculty of Nursing Alexandria University. Egypt.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
37
Apola, W. (2004). “Logistic Regression Analysis on the Number Coding Scheme
as a Solution to the Traffic Problems of Baguio City.” Unpublished
Undergraduate Thesis, CAS Benguet State University, La Trinidad, Benguet.
Begnaden, J. S. & Tickchap, J. C. (2006). “Logistic Regression Analysis on
the Willingness of College Student Smokers to Stop from Smoking.”
Unpublished Under Graduate Thesis, CAS Benguet State University, La
Trinidad Benguet.
Costes, G. & Emuslan, M. (2004). “Factor Analysis and Logistic Regression
Analysis on the Economic Crisis Management of College Students of
Benguet State University.” Unpublished Under Graduate Thesis, CAS
Benguet State University, La Trinidad Benguet.
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
38
Appendix A: The Questionnaire
Benguet State University
College of Arts and Sciences
Math - Physics – Statistics Department
La Trinidad, Benguet
Dear Respondent,
As a requirement of the course, Bachelor of Science in Applied Statistics,
we are conducting our thesis; a research entitled “LOGISTIC REGRESSION
ANALYSIS ON THE DECISION OF FRESHMEN COLLEGE STUDENTS
TO PURSUE THEIR ENROLLED COURSE”. In this connection, we would
like to solicit your cooperation & assistance by answering the questionnaires
completely and honestly. The researchers assure that your answers will be kept
and treated with outmost confidentiality.
Thank you very much and more power.
Truly
yours,
Juliet N. Carias
Researcher
Meldrid A. Luisito
Researcher
Noted by:
CRISTINA B. OCDEN
Adviser
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
39
QUESTIONNAIRE
Direction: Please check or write on the blank that corresponds to your answers.
Name: (Optional) Gender: Civil Status:
Age:
Average Grade in High school:
Number of Family members living (Family size):
Number of brothers/sisters who graduated in college:
Weekly allowance (please indicate):
Income of Both Parents (monthly):
1. Less than Php5, 000.00
3. Php10, 000-20,000
2. Php5, 000-10,000
4. Above Php20, 000
Present Residence:
1. Dormitory
3. With family/parents
2. Boarding house
4.With relatives
Working student (1):
Non-working student (2):
With scholarship (1): Without scholarship (2):
Financially dependent to parents (1):
Financially independent to parents (2):
Are you determined to pursue your enrolled course?
( ) YES
( ) NO
Logistic Regression Analysis on the Decision of Freshmen College Students
to Pursue Their Enrolled Course /Juliet N. Carias & Meldrid A. Luisito. 2008
Document Outline
- Logistic
Regression Analysis on the Decision of Freshmen College Students to Pursue Their
Enrolled Course.
- BIBLIOGRAPHY
- ABSTRACT
- TABLE OF CONTENTS
- INTRODUCTION
- REVIEW OF RELATED LITERATURE
- THEORETICAL FRAMEWORK
- METHODOLOGY
- RESULTS AND DISCUSSION
- SUMMARY, CONCLUSION AND RECOMMENDATION
- LITERATURE CITED
- Appendix