.1
. . l. ”I 5:. :1.
I .

: A . 1 5...“... ... ‘ . . l I‘\‘4‘
afifi . r a, r5 _ .53;

.3 yawn ..

G: afizx

1151'! .1.

95...... ,1...
x...

l .

7:: .330:

2 3|:

3:55):
t: a

s! 1. nae!

a

21.52;}:

 

 

 

a... 4:...» ‘

I... I. .5...‘

i 1:3;3‘1
1:.

 

’1

20!!

An lntegrative Psychological Model of Student Withdrawal

This is to certify that the
thesis entitled

presented by

Mark C. Zorzie

has been accepted towards fulfillment
of the requirements for the

degree in _ Psycholng

 

Major Professor’s Signature

05/04/201 0

Date

MSU is an Affirmative ActiorVEqual Opportunity Employer

LIBRARY
Michigan State
University

 

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIProi/Acc8PrelelRC/DateDueJndd

 

 

AN INTEGRATIVE PSYCHOLOGICAL MODEL OF STUDENT WITHDRAWAL
By

Mark C. Zorzie

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS
Psychology

2010

UMI Number: 1487146

All rights reserved

INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI

Dissertation Publishing

UMI 1487146
Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.

Pro uest'

 

ProQuest LLC
789 East Eisenhower Parkway
PO. Box 1346
Ann Arbor, MI 48106-1346

ABSTRACT
AN INTEGRATIVE PSYCHOLOGICAL MODEL OF STUDENT WITHDRAWAL
By
Mark C. Zorzie

College student withdrawal rates have remained relatively consistent despite
decades of attention to the issue. While turnover in the organizational context and
student withdrawal in the university context are typically considered conceptually
distinct, the two contexts have much in common. In this study, findings from these
literatures are integrated to create a model of college student withdrawal. In general, it is
hypothesized that students come to college with varying background characteristics that
affect integration into the university environment. This integration affects satisfaction
and commitment, which in turn affect withdrawal cognitions and behaviors. These
antecedents ultimately lead to a decision to withdraw. Findings from this longitudinal
study demonstrate that the general progression outlined in the model is supported.

Theoretical and practical implications are discussed.

ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Neal Schmitt, for his thoughtful and timely
feedback and for the support he has given me throughout this process. I would also like
to thank my committee members, Dr. Ann Marie Ryan and Dr. Kevin Ford, for their

guidance, which undoubtedly improved this thesis.

Finally, I would like to thank my parents, Paul and Barb Zorzie, for the support
they have given me over the years and continue to provide, and to Allison Shaw, whose

encouragement was steadfast and much appreciated.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES .......................................................................................................... vii
INTRODUCTION .............................................................................................................. 1
Entry Characteristics ............................................................................................... 9
Personality ................................................................................................... 9

Goal Orientation ........................................................................................ 12

Career Orientation ..................................................................................... 16

Perseverance ............................................................................................. 18

Knowledge Acquisition ............................................................................ 20
Social/Cultural Involvement ..................................................................... 21

Academic Self-Efficacy ............................................................................ 22

Other Variables ......................................................................................... 24

Entry Characteristics’ Role in the Model .................................................. 25

The Effects of Entry Characteristics on Integration ............................................. 26

The Effects of Integration on Psychological Outcomes ....................................... 28
Family Support .......................................................................................... 32

Academic Self-Efficacy ............................................................................ 34

Shocks ....................................................................................................... 35

Psychological Outcomes ....................................................................................... 36
Commitment ............................................................................................. 36

Satisfaction ................................................................................................ 39

College Grade Point Average ................................................................... 42

The Effects of Psychological Outcomes on Withdrawal Cognitions and Behaviors
................................................................................................................... 42

Other Variables ......................................................................................... 44

THE PRESENT STUDY .................................................................................................. 45
METHOD ................................................................................................................... 47
Procedure .............................................................................................................. 47
Measures ............................................................................................................... 48
Analyses ................................................................................................................ 53
RESULTS ................................................................................................................... 55
Preliminary Analysis of the Voluntary Nature of Withdrawal ............................. 55
Descriptive Statistics ............................................................................................. 56
Confirmatory Factor Analyses .............................................................................. 56

Stage 1 Variables ...................................................................................... 58

Stage 2 Variables ...................................................................................... 59

iv

Stage 3 Variables ...................................................................................... 61

Stage 4 Variables ...................................................................................... 62

Revised Descriptive Statistics ............................................................................... 64

Early Attrition Model Correlations ........................................................... 65

Longitudinal Model Correlations .............................................................. 66

Structural Equation Models .................................................................................. 67
Support for Hypotheses ........................................................................................ 75

Model Modification .............................................................................................. 79

Early Attrition Model Modification .......................................................... 81

Longitudinal Model Modification ............................................................. 87
DISCUSSION ................................................................................................................... 90
Main Findings ....................................................................................................... 90
Theoretical Implications ....................................................................................... 95
Practical Implications ............................................................................................ 98
Strengths and Limitations ..................................................................................... 99

Future Work ........................................................................................................ 101
CONCLUSION ............................................................................................................... 103
APPENDICES ................................................................................................................ 105
REFERENCES ............................................................................................................... 198

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 1 1.

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

LIST OF TABLES

Demographics for Entry, Semester 1, Semester 2, Semester 3, and

GPA/Graduation Samples ........................................................................... 129
Timeline of the Variables Collected ............................................................ 131
Early Attrition Model Scale Intercorrelations ............................................. 132
Longitudinal Model Scale Intercorrelations ................................................ 135
Revised Early Attrition Model Scale Intercorrelations ............................... 141
Revised Longitudinal Model Scale Intercorrelations .................................. 144
Correlations of the Exogenous Latent Variables in the Early Attrition Model.
..................................................................................................................... 147
Correlations of the Exogenous Latent Variables in the Longitudinal Model...
..................................................................................................................... 148
Intercorrelations Among the Observed Variables in the Early Attrition Model
..................................................................................................................... 149
Standardized Regression Coefficients in the Early Attrition and Longitudinal
Models ......................................................................................................... 167
Squared Multiple Correlations in the Models ............................................. 168
Intercorrelations Among the Observed Variables in the Longitudinal Model ..
..................................................................................................................... 1 69
Summary ofModel Support ......................................... 184
Standardized Regression Coefficients in the Modified Early Attrition Model.
..................................................................................................................... 186
Squared Multiple Correlations for the Revised Early Attrition Model ....... 187
Standardized Regression Coefficients in the Modified Longitudinal Model
..................................................................................................................... 188
Fit Indices Across the Tested Models ......................................................... 189

vi

Figure 1.

Figure 2.
Figure 3.
Figure 4.
Figure 5.

Figure 6.

LIST OF FIGURES

The hypothesized model of student persistence ............................................. 190
The measurement model ................................................................................. 191
The structural model ....................................................................................... 193
The full model ................................................................................................ 194
Standardized regression coefficients in the early attrition model .................. 195
Standardized regression coefficients in the longitudinal mode] ..................... 196

vii

INTRODUCTION

Turnover is a widespread problem among organizations of all types, and increased
turnover has been found to relate to a variety of negative outcomes for organizations
(Zimmerman, 2008). Universities are no exception. Attrition rates of college students
have been of interest for at least 80 years (Summerskill, 1962). Despite decades of
research on the issue, however, these rates have remained relatively consistent in the
educational setting (Tinto, 2006; Kalsner, 1991). In the last 25 years, persistence rates
have remained around 40% for four-year public institutions and around 56% for four-year
private institutions within five years after entry (ACT, 2008). This is a concern to
universities because of the potential revenue lost, but also in the attraction of new
students; a low retention rate weighs against a university as a prospective student views
quality rankings in deciding which institution to attend (Tinto, 2005). Many universities
have implemented programs and developed strategies to reduce withdrawal, but these

have met differing levels of success (Tinto, 2006).

While “turnover” is the term commonly used in the organizational literature to
refer to the departure of an individual from his or her job (e.g. Youngblood, Mobley, &
Meglino, 1983; Lee, Mitchell, Wise, & Fireman, 1996; Harman, Lee, Mitchell, Felps, &
Owens, 2007), a variety of terms have been used in the past few decades to describe the
same phenomenon in the university setting. Consistent with Berger & Lyon’s (2005)
description of the terminology, in this article I will refer to the departure of students as
“withdrawal” or “dropout” and the overall failure of reenrollment from one semester to

the next as “attrition.” Withdrawal intentions represent a planned course of action to

drop out, while actual withdrawal refers to whether or not a student reenrolled at the
university he or she had been attending. Although Berger and Lyon (2005) use
“persistence” to refer to simply continuing one’s education somewhere, including
transferring from one institution to another, I will use this term to refer to the continuing
one’s education within the university at which a particular student began; that is, I will

use persistence to mean the opposite of withdrawal.

Throughout this paper I will attempt to integrate findings from the organizational
literature and findings from the literature on the academic arena; this may seem like a
stretch at first. Munson and Rubenstein (1992), however, argued that the two are actually
quite alike, with similar individual functions and tasks, situations, environmental
conditions, and assessment characteristics. Types of schoolwork can be categorized
along the same dimensions of data, people and things on which other work is categorized
(e.g. O*NET). Students are attracted to and are successful in different areas of
schoolwork just as workers outside the classroom are in various careers. As Munson and
Rubenstein note, both students and these workers must operate under stressful conditions,
perform tasks they would rather not, follow directions, and work with others. While I
agree that there are important differences between these two contexts, I believe there are

also relevant similarities and that these corollaries are worth discussing.

In the organizational turnover literature, the first significant developments came
from March and Simon (1958). They theorized that dissatisfaction resulted from
organizational variables such as inadequate pay and few advancement opportunities, and
that this dissatisfaction led to a desire to voluntarily leave. Mobley (1977) adapted this

idea of dissatisfaction into a larger model. In his conceptualization of this process,

2

dissatisfaction led to thoughts about quitting, which in turn triggered search and
evaluation of the available alternatives to one’s current job. If the alternatives were
judged as better than the current job, the individual would intend to quit and then follow

through with that intention.

In the 16 years following the publication of Mobley’s (1977) model, many others
were proposed, but these mainly focused on satisfaction, commitment, and turnover
intentions (Peterson, 2004). While these variables consistently demonstrated significant
relationships with turnover (Tett & Meyer, 1993), Lee and Mitchell (1994) noted that
these models typically explained less than 15% of the variance in voluntary turnover.
These authors proposed an unfolding model that, in contrast to previous models, was
based on significant life events or “shocks”. These will be discussed in greater detail
later in this paper, but the general idea is that these shocks cause one to leave a job,

sometimes with and sometimes without deliberation and a search for alternatives.

Lee and Mitchell’s (1994) model remains influential, but more recent work by
Horn, Griffeth and colleagues (e. g. lHom & Griffeth, 1995; Griffith, Horn, & Gaertner,
2000; Hour & Kinicki, 2001) has helped identify and summarize the effects of a number
of variables on turnover. The roles of stress, leadership, absenteeism, withdrawal
cognitions and other factors have been better elucidated. While there does not appear to
be consensus on a single model of organizational turnover, many of the elements of this
process and the relationships between them are better understood than they were in

March and Simon’s (1958) day.

In the academic context, two theoretical models of college student departure have
garnered the most interest: Tinto’s (1975) student integration model and Bean’s (1980)
model of attrition. Initially these models had many differences, but over time criticism
and research have led to revision and some convergence in the models. Many of the
propositions in each model have been tested over the years, and a clearer picture of the

causes of student withdrawal is emerging.

The most widely cited framework for student departure is Tinto’s (1975; 1993)
model (Metz, 2002). This model was based on works by Durkheim (1951) and Spady
(1971). Durkheim posited that suicide was largely a result of a lack of integration
between an individual and society. Spady adapted this theory into the first theoretical
model of student departure. He theorized that students who had trouble integrating into
the social fabric of a university would be more likely to drop out. Tinto (1975) built upon
this model. He theorized that students come to college with differing attributes, including
family background, skills, abilities, and some level of initial commitment. At college,
students experience academic and social interactions with their faculty and peers as well
as differing levels of academic performance. Successful performance and high-quality
interactions lead to academic and social integration, which in turn leads to greater

commitment and persistence.

Though Tinto’s model has received the most attention, Bean’s (1980; 1981; 1985)
model has also received a great deal of interest. Bean’s (1980) initial Causal Model of
Student Attrition incorporated background variables and a large number of individual
determinants such as goal commitment, GPA, one’s major, and the opportunity to get a

job or transfer. The interaction between these elements was theorized to lead to

4

satisfaction, which in turn led to commitment. This ultimately resulted in a decision to
withdraw or persist. Bean (1981) later attempted to synthesize elements from different
theoretical models into a more unified model; the subsequent model also included
background variables’ effect on interactions with an organization, but the role of external
factors was more pronounced. Intent was also included in this model as a direct

precursor to withdrawal.

Just as Bean has attempted to synthesize aspects of several models into his own,
Tinto (1993) modified his model based on criticism it had received. He included external
factors affecting one’s goals and commitments, and also included one’s intentions as a
direct precursor to dropping out. Tinto’s and Bean’s models are similar in that they both
account for the effects of precollege characteristics, commitment, and the match or fit
between students and colleges on degree attainment. They are different in that Tinto’s
model places greater emphasis on academic and social integration and institutional and
goal commitment, whereas Bean’s model focuses on the effects of persistence intentions,

attitudes, and factors external to the institution (Cabrera, Nora, & Castaneda, 1992).

Empirical research exploring these models has supported many of the
propositions in each. Among the findings in support of Tinto’s model, student entry
characteristics have been found to affect withdrawal (Pascarella & Terenzini, 1980;
Braxton, Sullivan, & Johnson, 1997) and academic integration has been found to affect
commitment to graduating. Among the findings regarding Bean’s model, institutional fit
has been found to affect intent to persist (Cabrera, Castaneda, Nora, & Hengstler, 1992)
and personal and organizational variables have been found to affect attitudes and intents

(Bean & Vesper, 1990).

Given the interest in both models, it is not surprising that attempts have been
made to combine them into a single model. Cabrera, Nora, & Castaneda (1992) used
structural equation modeling to test a model that combined the two. In the resulting
model, external factors were found to affect both academic and social integration. This
integration led to institutional and goal commitment, which led to intent to persist and
ultimately to persistence. Academic integration also affected GPA, which had a direct
effect on persistence. Subsequent models in this line of research have incorporated such
findings into a more unified theory on persistence. Bean and Eaton (2000) developed a
psychological model of student retention that includes entry characteristics, elements of
the institutional environment including academic and social interactions, integration, fit,

commitment, and persistence intentions.

To my knowledge, only one author has attempted to bridge the research from both
the collegiate and organizational contexts into a unified model. Peterson (2004)
combined elements from Tinto’s model and from various turnover models into the
Organizational Model of Employee Persistence. In this model, pre-entry attributes affect
employee characteristics such as goals, commitment, and satisfaction. These latter
elements affect a host of organizational experiences, which are framed as different facets
of integration. These integrational characteristics affect subsequent goals, commitment
and satisfaction (all of which have appeared previously in the model), which eventually
lead to a decision to stay with or leave an organization. In a subsequent test of this
model, Peterson (2007) found support for the hypotheses that initial goals, commitment,
satisfaction, and career decision-making self-efficacy affect employee integration. This

integration explained 25% of the variance in subsequent goals, commitment, and

satisfaction, though actual turnover was only weakly predicted by the variables in this
model. Though this model combined elements from these two areas of research, it was

designed to describe turnover and has not been applied to an undergraduate sample.

One of the common and generally supported elements of these models is
background or entry characteristics (e.g., Braxton, Sullivan, & Johnson, 1997). Students
come from unique backgrounds, so it should be expected that they bring with them
different experiences, skills, abilities, and personalities. These factors will influence the

various interactions and outcomes of the college experience.

Though this broad category of precollege characteristics has garnered support as a
whole, the specific elements that compose it have been varied and inconsistently
employed. Braxton, Sullivan and Johnson (1997) found support for the inclusion of
prematriculation characteristics in Tinto’s model, but they looked at effects across
multiple studies that included different characteristics. Some elements that have been
mentioned by Tinto and Bean have not received much attention. For example, Tinto
(1993) specifically discussed the hypothesized but largely untested effects of personality
on attrition, and Bean (1981) incorporated Pascarella’s (1980) model (which includes
personality orientations) into a synthesized model of student departure. Despite this
theorization, personality has been largely ignored by researchers. Cabrera, Nora, and
Castaneda (1992) attempted to integrate Tinto’s and Bean’s models into a unified model
of attrition, but they only included Financial Attitudes and Approval by Family and
Friends as precollege characteristics. An attempt to bring consistency and to integrate

these background characteristics is needed.

Through a review of the literature, I identified 10 prematriculation characteristics
that could affect student attrition: Motivation, Social/cultural involvement, Support and
Encouragement, Initial Self-Efficacy, Career Orientation, Locus of Control, Coping
Strategies, Knowledge, Perseverance, and Personality. Seven of these variables were
present in the archival data set I plan to use, and as will be discussed, six were deemed
appropriate for inclusion in the first stage of a new model of student withdrawal. In an
effort to both investigate the effects of these variables and to provide an integration of the
organizational and educational literature on tumover/withdrawal, I developed Figure 1,
which represents a general heuristic model of the factors that may lead to withdrawal. In
the following pages, I provide theoretical arguments for each of the stages of withdrawal
that are implicit in this figure. Also provided is rationale for hypotheses related to each

of these stages.

A few basic ideas outline one’s progression through this model. First, a student
enters a university with the aforementioned entry characteristics. These entry
characteristics affect the manner in which a student integrates with the university
environment. Greater integration leads to satisfaction and commitment. Increased
satisfaction leads to reduced thoughts and behaviors related to withdrawal, and these

reduced thoughts and behaviors ultimately reduce actual withdrawal.

This general progression is represented by the variables in this model, which are
divided into stages. The entry characteristics appear in the first stage. These entry
characteristics lead to the second stage variables of academic and social integration, and
academic integration leads to GPA; financial integration and family support also appear

as exogenous variables in this stage. The variables in stage 2 lead to the intermediate

8

outcomes of academic, social, and financial satisfaction and institutional and degree
commitment in stage 3. The stage 3 variables reduce one’s activities regarding
withdrawal cognitions and behaviors (represented by the stage 4 variables of thoughts of
leaving, intent to withdraw, and search behaviors), which ultimately result in reduced

likelihood of withdrawing.

Entry Characteristics

Personality.

Personality is one of the entry characteristics that has been included in theoretical
models of student withdrawal. Early investigation of the role of personality in student
withdrawal lacked both a model of withdrawal and a commonly accepted framework for
personality. Grace (1957) found that independent and responsible students were less
likely to withdraw and that anxious students were more likely to withdraw. Dropouts
have been shown to be lower on abasement, achievement, order, and endurance, higher
on autonomy, exhibition, and aggression (Heilbrun, 1965), higher on impulsivity and
change (Maudal et. al., 1974), and higher on assertiveness, stubbornness, and

independence (Pandey, 1973).

Though Tinto (1993) recognized the difficulty of utilizing such findings without a
good framework for personality, he nevertheless theorized that personality affected
dropout decisions, saying, “. .. though we sense that personality must play a part in
student departure, we are thus far unable to say just how different elements of personality

affect student leaving in different institutional settings” (p. 45). The Big 5 framework of

personality has provided a useful structure for studying withdrawal. The Big 5 consists
of openness to experience, conscientiousness, extraversion, agreeableness, and emotional
stability. Barrick and Mount (1991) describe openness to experience as “being
imaginative, cultured, curious... [and] broad minded,” conscientiousness as “being
careful, thorough, responsible, organized... hardworking, [and] achievement-oriented,”
extraversion as “being sociable, gregarious, assertive, talkative, and active,”
agreeableness as “being courteous, flexible,. .. cooperative... and tolerant,” and
emotional stability as a lack of “being anxious, depressed, angry,. .. worried, and

insecure.”

Since the advent of the Big 5 personality factors, some researchers have begun
investigating the relationship between personality and student withdrawal. Tross, Harper,
Osher, and Kneidinger (2002) studied conscientiousness (along with achievement and
resiliency) and found that it predicted persistence among college students both directly
and indirectly through GPA. Lounsbury, Saudargas, and Gibson (2004) examined the
Big Five as a predictor of withdrawal intention in first-year students and found that
conscientiousness, extraversion, agreeableness, and emotional stability were significantly
related to withdrawal intention, with conscientiousness and emotional stability

accounting for 17% of the variance.

In the organizational context, there have been three studies that have aided in
understanding the personality-tumover relationship. In Barrick and Mount’s (1991)
meta-analysis, they found significant but fairly weak relationships between
conscientiousness and openness to experience and a variable similar to turnover. This

variable was a combination of turnover and tenure, however, so it is hard to draw

10

conclusions from these relationships. Salgado (2002) conducted a meta-analysis where
he found that persistence in an organization was affected by conscientiousness,
agreeableness, and openness to experience. Lastly, Zimmerman (2008) found that
emotional stability had a strong association with intentions to quit, while agreeableness

and conscientiousness had strong relationships with actual turnover.

These findings are encouraging in the sense that they support a relationship
between personality and withdrawal. Despite this, to my knowledge only one study has
investigated the role of the Big' 5 personality characteristics on college student
withdrawal in the context of a larger model. Okun and Finch (1998) incorporated the Big
5 into Tinto’s framework and found that conscientiousness had the largest direct effect on
student departure (-.29), while neuroticism had the largest effect on social integration (-
.48), which led to institutional commitment and resulted in decreased withdrawal. A
limitation of the study, however, was that only 14 students out of 240 dropped out. This
rate of less than 6% withdrawal is much lower than the near-30% dropout rate at a typical
4-year public institution from the first year of college to the second (ACT, 2008), so the
university in the study may not have a student body representative of others. In addition,
the low withdrawal rate likely attenuated the observed relationships between personality

factors and the outcomes of withdrawal and integration.

These studies have helped to clarify the nature of the relationships between the
Big 5 factors and withdrawal. All five studies that included conscientiousness and either
turnover/withdrawal or intent to turnover/withdraw found significant relationships
(Salgado, 2002; Zimmerman, 2008; Tross, Harper, Osher, and Kneidinger, 2002;
Lounsbury, Saudargas, & Gibson, 2004; Okun & Finch, 1998). All four that included

emotional stability and these outcome variables found direct or indirect significant
relationships. Three of these four supported relationships between agreeableness and
turnover/withdrawal. Only one of the four found such a relationship regarding
extraversion, and none found such a relationship regarding openness to experience.
Thus, conscientiousness, emotional stability, and agreeableness have the most consistent
relationships with turnover and withdrawal and intent to turnover/withdraw when both

the employment and academic settings are considered.

While much of the research has assessed direct relationships between personality
and withdrawal or withdrawal intent, I believe it is more likely that these relationships are
mediated by academic and social integration. For instance, students who are more
conscientious students should have higher quality social relationships with other students
due to increased dependability, and should also be more meticulous with academic work
due to the increased organization, thoroughness, and will to achieve (Barrick & Mount,
1991). Thus, conscientiousness should increase integration into the social and academic
environments. The relationships of these personality factors to academic and social

integration are hypothesized in the next section of the proposal.

Goal Orientation.

In college, not everything a student must do to succeed is directed and prescribed.
Students must practice self-regulation with regard to their learning objectives. Self-
regulation is useful as a general skill because it allows one to engage in activities that

facilitate the attainment of some sort of goal (Porath & Bateman, 2006). When applied

12

to the context of academics, this self-regulated learning facilitates success through control
of motivation, regulation of affect, and modification of behavior (Schunk & Zimmerman,

2008).

One facet of motivation on which individuals can regulate themselves is goal
orientation. Goal orientation (Dweck & Legget, 1988) is a “relatively stable dispositional
trait that co-varies with the individual’s implicit theory of ability,” (p. 26, Button &
Matthieu, 1998). There are three independent dimensions of goal orientation: mastery
goal orientation, performance-approach goal orientation, and performance-avoid goal
orientation. According to Ames (1992), mastery-oriented individuals believe the effort
they expend will lead to success. Such learners have increased intrinsic motivation and
excitement about learning because they desire to understand and internalize information
(Elliot & Harackiewicz, 1996). Mastery-orientation has generally been positively related
to learning-related outcomes. Individuals high on this motivational factor have been
shown to persist when facing challenges (Elliot & Dweck, 1988; Vollmeyer &
Rheinberg, 2000), succeed when faced with complex situations (Erez, 2005), have greater
persistence in college (Edens, 2006), have greater commitment in high school
(Tuominen-Soini, Salmela—Aro, & Niemivirta, 2008), and demonstrate greater motivation

to learn (Colquitt & Simmering, 1998).

Performance orientation was originally considered a unitary dimension (Elliot &
Harackiewicz, 1996). Individuals with a high performance orientation, it was believed,
desire to succeed because they want to outperform others. They see learning as a
mechanism for achieving their goals, but they are not too concerned with a deep

understanding of the topics they “learn.” Under this original conceptualization of

13

performance orientation, it was found that individuals with such an orientation generally
demonstrate more undesirable behaviors, including a lack of progress in high school and
lower self-esteem (Tuominen-Soini, 2008), excessive daytime sleepiness (Edens, 2006),
lower motivation to learn (Colquitt & Simmering, 1998) and a “helpless” pattern of

response to challenges (Button & Matthieu, 1998).

Despite these findings, there was evidence that performance orientation was not
maladaptive under some circmnstances. Recognizing inconsistency in the results of a
number of studies using a simple performance/mastery dichotomy, Elliot and
Harackiewicz (1996) proposed that the dimension of performance orientation be divided
into performance-approach and performance-avoid dimensions. Individuals with a
performance-approach orientation are said to be motivated to learn by the challenge of a
task and adaptive competition with others, and this is believed to increase concentration
and excitement. On the other hand, individuals with a high performance-avoid
orientation are motivated by a desire to avoid demonstrating incompetence; this is said to

decrease intrinsic motivation.

Subsequent findings using this trichotomous framework have found that” the main
effect of mastery goal orientation holds, and the roles of the two types of performance
goals are becoming better understood. Performance approach goals have generally been
found to lead to positive outcomes, while performance avoid goals have generally found
to lead to negative outcomes. Harackiewicz, Barron, Tauer, Elliot, & Thrash (2002)
found that mastery goals lead to higher interest and enjoyment, while performance
approach goals lead to higher GPAs both in the course of interest and overall. The

authors concluded that “both mastery and performance-approach goals have positive and

I4

complementary consequences for motivation and performance in college courses over the
course of students’ academic careers” (p. 574). Similarly, Barron and Harackiewicz
(2001) found that mastery goals led to higher interest while performance goals led to
better performance for college students working on a math task. A meta-analysis on the
goal-orientation literature by Payne, Youngcourt & Beaubien (2007) found that both
mastery and performance approach goal orientations had positive relationships with

learning, while a performance avoid orientation had a negative relationship with learning.

In the organizational context, results have been very similar. This is not
surprising, as both students and workers in organizations must set many goals as they
progress. While perhaps not shared by workers, the overarching goal of graduation for
students is a product of many smaller goals that factor into the success of a student. In
their aforementioned meta-analysis, Payne, Youngcourt & Beaubien (2007) found
positive relationships between both mastery and performance approach goal orientations
and job performance. Turnover, however, has not been a variable of interest in the goal
orientation literature when using this three-factor model. To my knowledge, there have
also not been any studies examining the direct or indirect effects of these three facets of
goal orientation on college student persistence. Despite the lack of research in these
specific domains, the findings relating to each of these facets provide reason to believe
they might aid in our understanding of the student withdrawal process. Being high on the
mastery and performance-approach goal orientations should lead students to engage in
behaviors that facilitate persistence. The consequences of these goal orientations,
including greater interest, learning, and performance should also lead to greater academic

integration.

15

There has been some evidence that the positive effects of performance approach
goals are dependent on mastery goal orientation level; for example Pintrich (2000) found
that high performance-approach goals led to negative motivational outcomes when they
were not accompanied by corresponding mastery goals. However, other studies have
found direct, positive effects of performance-approach goals on various outcomes (e. g.
Barron & Harackiewicz, 2001). Although the verdict is still out on this matter, in this

study I will investigate the independent effects of each.

While the outcomes of goal orientation have largely been related to performance
or success, I believe goal orientation should have an effect on academic integration.
Mastery-oriented students are intrinsically motivated to learn and believe that their effort
can result in positive outcomes. Such students would likely appreciate their classes more
and be more genuinely interested in their major, increasing academic integration.
Students high on performance-approach goal orientation should be motivated to learn
material in order to demonstrate knowledge, also increasing appreciation of classes and
their choice of major, which in turn should increase academic integration. Students high
on performance-avoid goal orientation, on the other hand, would be less motivated to
deeply engage in the academic environment and may only do what is needed to “get by”.
This approach would lower academic integration. The effects of these dimensions on

academic integration are hypothesized in the next section.

Career Orientation.

l6

The formal purpose of college is to prepare students for life after graduation—
mainly, to prepare students for some sort of career. Some students may choose to pursue
further schooling to this end, and some may choose alternative routes such as service
work, but for the majority of students, a career is the logical next step after graduation.
Though this may be a common progression in individuals’ lives, there is variability in the
importance of career-related goals and the belief that one can attain them. Some students
may be at a university due to familial or social pressures rather than a strong desire to
pursue a career. Others may appreciate the importance of secondary education in
pursuing a career, but they may not expend much effort in deciding upon a particular
career to pursue. In either case, career goals are not strong. When a student does not
have strong career goals or does not believe he or she can attain them, it is likely that less
time will be spent on gathering career-related information. Similarly, degree attainment
becomes less important when it does not have a place in one’s larger goal framework. A
student can still want to obtain a degree for other reasons, but if career goals are absent,

many students will have fewer reasons for continuing in college.

Indeed, career goal identification has been shown to lead to persistence (Peterson,
1993; Sprandel, 1986). Kahn, Nauta, Gailbreath, Tipps, & Chartrand (2002) showed that
students who were more anxious about choosing a career were more likely to dropout.
Hull-Blanks and colleagues (2005) showed that freshmen with well-defined career goals
made better persistence decisions than freshmen lacking such goals. Furthermore, there
has been a great deal of interest in students’ beliefs that they can identify a career. This
has been termed career decision-making self-efficacy, and it has demonstrated

relationships with persistence (Peterson, 1993; Sandler, 2000). Together, the

17

identification of career goals and the belief that one can attain them are important
elements of student persistence. In this paper, career orientation reflects a combination of
these elements and is defined as one’s ability to identify career-related goals and one’s
belief he or she can attain these goals. Accordingly, the items in the career orientation
measure refer to both identification of career goals and confidence in attaining them.
While career goal identification and career decision-making self-efficacy have
demonstrated direct relationships with withdrawal, I believe career orientation is more
appropriately conceptualized like the other entry characteristics as being related to
academic integration. Students who can identify career goals and believe they can attain
them should concentrate more on courses related to those goals, and they should also
select a major that is congruent with those goals and have a strong desire to earn a degree
in that major. The relationship between career orientation and academic integration is

hypothesized in the next section.

Perseverance.

Perseverance is defined as “steady persistence in a course of action, a purpose, a
state, etc., esp. in spite of difficulties, obstacles, or discouragement” (Random House,
2009). Essentially, this is what college students do in their pursuit of a degree.
Intuitively, it follows that students with a greater propensity for perseverance would be
more likely to overcome obstacles and ultimately obtain a degree. The problem with the
current literature is that perseverance is most ofien treated only as an outcome.

Perseverance, like persistence, is simply treated as continued progress toward a degree; it

18

is a product of many other factors such as the prematriculation characteristics I have
already discussed. It is generally accepted that past behavior is a strong predictor of
future behavior (Owens & Schoenfeldt, 1979), so I believe there is reason to believe that
past perseverance behaviors should predict positive outcomes even on complex tasks

such as college persistence and degree attainment.

In the organizational literature, it seems that a tendency toward perseverance has
not been treated as a precursor to turnover. To my knowledge, the only published study
that has treated perseverance as an exogenous variable in the collegiate context was a
study dealing with racial differences in the effects of noncognitive factors on degree
attainment (Tracey & Sedlacek, 1986). In that study, items were factor analyzed and
clusters were given labels (such as perseverance). The authors found that this
perseverance scale significantly predicted graduation of Black students, but that it did not
have any effect among White students. While the research in this area has been limited, I
believe there is enough of a theoretical justification to contend that an a-priori
perseverance scale may have a more consistent effect for college students on degree
completion. Reflecting the definition presented above, the items in the perseverance
scale used in this study gauge one’s ability to persist in a course of action, generally
toward some goal, in spite of difficulties. Students who have done this previously in life
should be likely to do this in their courses and in pursuit of a major. Thus, students with
greater perseverance should perceive greater academic integration. This is hypothesized

in the next section.

19

Knowledge Acquisition.

It is generally assumed that grade-point average (GPA) is a measure of a student’s
ability to acquire and demonstrate knowledge, though this assumption has been
questioned (Pemberton, 1970). Course selection by students, instructor grading practices,
and other between-institution differences can have a large impact on a high school
student’s GPA (Lei, Bassiri, & Schulz, 2001). Despite this, high school grades (along
with SAT scores) continue to demonstrate among the strongest relationships with college
success, and thus are a widely used metric in admissions decisions (e.g. Geiser &

Santelices, 2007).

If the variables of interest are really those relating to knowledge acquisition and
demonstration of abilities, it is possible these can be captured in a manner other than
cumulating GPAs across classes. Biographical items can be used to gauge the amount
one typically learns in a class, use of effective knowledge-acquiring techniques, and
learning and past performance relative to other students—all elements that should
theoretically also affect GPA. These should affect success in college and directly or
indirectly impact persistence. Other studies have used biographical inventories in
predicting college outcomes (e. g. Richards, Holland & Lutz, 1967), but there is only one
study potentially relevant to this factor dealing with the biodata constructs and student
attrition. Owens & Shoenfeldt (1976) clustered students on factors derived from a
biographical inventory, and the results showed that a group labeled “dependent, poorly
adjusted dropouts” was lower on past academic achievement (though it is not clear
exactly what academic achievement entailed). The dearth of literature in this area only

increases the importance and utility of including such a factor in the current study.

20

In the present study, knowledge acquisition represents a student’s ability and
desire to learn. The items in this scale reflect past tendencies to learn, learning ability
compared to other students, and determination to learn concepts and information in
school. Greater knowledge acquisition should result in similar behaviors in college

courses, increasing academic integration. This is hypothesized in the next section.

Social/Cultural Involvement.

Just as students participate in social and cultural activities in college, they also
participate to varying degrees in these types of activities in high school. Such activities
may be formal extracurriculars, such as belonging to a sports team or a debate club, or
informal activities, such as hosting social gatherings, volunteering in the community, or
visiting museums and festivals. Participating in these activities in either the high school
or college context contributes to integration into the social fabric of one’s surroundings.
As I will later discuss, this social integration is important for success and persistence in

college.

Hossler and Stage (1992) found that participation in high school activities does
have an impact on the student’s aspirations in college. Other researchers have found that
involvement in these sorts of activities have a positive impact on educational attainment
(Spady, 1975; Hearn, 1984), and both the time a student spends in clubs or groups and
the time he or she plans to spend volunteering in college have been found to relate

directly to student persistence (Astin & Oseguera, 2005). Essentially, the integration a

21

student experiences in high school reflects an ability to achieve similar integration in a

different surrounding.

While these studies have mainly been concerned with social rather than cultural
activities, cultural activities are theoretically relevant. The items in the social fit scale
refer to shared ideas of firn with other students, as well as a campus having social
activities one enjoys. These could easily be cultural activities. Students in universities
come from a variety of backgrounds and cultures, and the university environment
commonly provide opportunities for students to attend art museums, ethnically-related
events, or other cultural events. Increased participation in these events in high school,
then, should facilitate integration into the social environment of college. In the present
study, the social/cultural involvement scale reflects involvement in both social and
cultural activities, including involvement in clubs, volunteering, and attending museums
and theatrical performances. Because of the nature of social/cultural involvement, it is

hypothesized (in the next section) that it will affect social integration.

Academic Self-Efficacy.

Another component of self-regulated learning is self-efficacy, which Zimmerman
and Shunk (2008) describe as “judgments of personal capabilities to organize and execute
courses of action required to attain designed types of goals.” When a student has success
in academic pursuits, it increases aspirations in future academic pursuits (Bean & Eaton,
2000). If an individual believes he or she can attain a goal, he or she will be more likely

to engage in the activities that facilitate success. For instance, in theory a student who is

22

getting a C or considering dropping out but believes he or she is actually capable of
succeeding if they make some changes and perhaps increase effort might seek out a TA’s
office hours or end up studying more, whereas a student low in self-efficacy may not
engage in these behaviors. These behaviors would lead to success and would increase the

likelihood that a student would complete a degree.

The performance of both students and workers in organizations relies on self-
regulation. Self-efficacy is important for both groups in directing behaviors that
accomplish one’s goals. The extant literature on self-efficacy in the organizational
context provides encouraging results. Staj kovic & Luthans (1998) conducted a meta-
analysis in which self-efficacy was found to significantly relate to work performance.
Saks (1995) found a positive relationship between initial self-efficacy when workers
entered a company and eventual turnover. McNatt and Judge (2008) found that a self-
effrcacy-raising intervention increased job attitudes and reduced turnover among

employees who underwent it.

Findings pertaining to the relationship between self-efficacy and student
persistence have been mixed. Haines, McGrath & Pirot (1980) found that self-efficacy
does not predict persistence on a lab task. Kahn & Nauta (2001) found that precollege
self—efficacy does not predict persistence, but that several social-cognitive factors
measured in the second semester including self-efficacy do predict persistence from a
student’s first year to their second. Multon, Brown & Lent’s (1991) meta-analysis
revealed that self-efficacy accounts for 14% of the variance in academic performance and

12% of the variance in academic persistence. Robbins, Allen, Casilla, & Peterson (2006)

23

found that academic self-efficacy predicts GPA and persistence above traditional

cognitive predictors.

A study by Gore (2006) may help reconcile these findings. He found that the
timing of the measurement of self-efficacy beliefs may be important. Self-efficacy
measured at the beginning of college was not a significant predictor of persistence, but
self-efficacy measured at the end of the first semester did predict persistence. Students
may need feedback in order to develop accurate efficacy beliefs. In light of these
findings, I believe that self-efficacy should be included later in the hypothesized model

and will redirect attention to this construct in subsequent discussion.

Other Variables.

As was mentioned earlier in this paper, there are other variables that could be
relevant prematriculation characteristics. While family support and encouragement was
not measured at the beginning of college, it was measured at later time points. It is likely
that such support is stable across time points—this can be investigated based on the data
in the later time points, and if true, would lend credence to the belief that it would be
similar if it had been measured at the beginning of college. Therefore, I do not think its

absence at the initial time point will detract from this study.

Locus of control and coping strategies, on the other hand, are variables that were
not measured in this data collection. Even so, I do not believe they are completely absent
from our data; these constructs have ties to motivation and self-efficacy. Individuals with
an internal locus of control believe they have the power to control the events that occur in

24

their lives, while individuals with an external locus of control believe external forces
have a stronger influence on their lives. This has much to do with motivation and self-
efficacy; if one does not believe he or she can control the events in his or her life, there
will be little belief that one’s actions will matter and less motivation to perform those
actions. Coping Strategies can be viewed in a similar fashion. The road through college
is an often bumpy one, and students need to know how to deal with the problems they
encounter. If they believe these problems are solvable and also that they have the ability
to deal with the problems, they will be more likely to solve them. Because of these
connections to variables that were measured, I do not believe the omission of locus of

control and coping strategies will significantly impact the study.

Entry Characteristics’ Role in the Model.

Thus far, I have described the nature of a set of pre-entry characteristics. As
Figure 1 illustrates, and as was mentioned in the description of each of these variables, it
is likely that the relationships between these constructs and withdrawal are mediated by
other constructs. The prematriculation characteristics that have been described affect
how one integrates into his or her environment. This integration affects the outcomes of
commitment, satisfaction, GPA and self-efficacy. These in turn affect intentions and
behaviors related to withdrawal, which ultimately lead to a decision to withdraw or

continue in college. In the following sections, these relationships are described in detail.

25

The Effects of Entry Characteristics on Integration

A central component of both Tinto and Bean’s models is fit or integration. The
interactions that one has in college affect how he or she perceives a “match” with the
university environment. As Tinto (1993) explainsit, this occurs within two systems: the
academic and the social. The academic system is the formal educational structure at a
university. Within this system, individuals engage in a number of activities, such as
going to class, completing assignments, speaking with teachers, and so on. The social
system consists of more informal elements, such as interactions not specifically serving

purely academic functions between friends, faculty members, and RAs.

Students can integrate into either of these systems independently of the other.
One could enjoy a rich social experience while not really appreciating the academic
aspects of a particular university. Conversely, one could thrive academically while
failing to engage in adequate social interactions. In addition to these previously-
conceptualized facets of fit, I also believe that financial fit could have an impact on
attrition. Finances have consistently been shown to be an issue in student dropout
decisions (Kalsner, 1991). Cabrera, Nora and Castaneda (1993) found that financial
attitudes affect persistence indirectly through GPA. Insufficient finances become a
problem when they affect what an individual can or cannot do in the university
environment. For instance, if a student’s friends are going out to a bar or restaurant but
he or she does not have the money to participate in this activity, this might affect later
variables in the model such as satisfaction. If this individual had gone to a university that
matched his or her financial situation better, this person may not be restricted in the

activities in which he or she chooses to engage. This would indicate a better financial fit.

26

Thus, in addition to the academic and social aspects of fit, financial fit may be an

important factor in the process of student withdrawal.

As was mentioned, Tinto (1993) and Bean (1980) assert that the pre-entry
variables discussed in the previous section may lead to increased fit. These variables
affect how a student interacts with others in his or her environment. For instance, a
student who is low on the personality characteristics of agreeableness and emotional
stability may have difficulty engaging in productive or successful academic or social
interactions because they could be perceived as unfriendly and discourteous (fiiendliness
and courteousness being two components of agreeableness; Barrick & Mount, 1991),
leading to lower academic and social integration. A student higher on knowledge
acquisition and perseverance may seek out and maintain better academic interactions,
leading to higher academic integration. A student who is not socially involved may have
worse social interactions, leading to lower social integration. Thus, the following

relationships between entry characteristics and integration are hypothesized.

H] a. Conscientiousness will have positive effects on academic integration.
H I b. Agreeableness will have positive eflects on academic integration.

H1 c. Emotional Stability will have positive eflects on academic integration.

H2a. Conscientiousness will have positive effects on social integration.

H2b. Agreeableness will have positive eflects on social integration.

27

H2c. Emotional Stability will have positive effects on social integration.

H3a. Mastery goal orientation will have a positive effect on academic integration.

H3 b. Performance approach goal orientation will have a positive eflect on academic

integration.

H3c. Performance avoid goal orientation will have a negative eflect on academic

integration.

H4. Career orientation will have a positive effect on academic integration.
H5. Knowledge acquisition will have a positive eflect on academic integration.
H6. Perseverance will have a positive efifact on academic integration.

H7. Social/cultural involvement will have a positive eflect on social integration.

The Effects of Integration on Psychological Outcomes

The path from integration to withdrawal is not hypothesized to be direct. Both the
[/0 literature and the educational literature have supported mediational effects of the
variables of satisfaction, commitment, self-efficacy and GPA. While each of these
variables is discussed in greater detail later, the relationships between integration and

these variables are discussed below.

28

Theoretically, integration should affect how satisfied a student is with the
university environment (Bean, 1981) and how committed he or she is to the university
and graduation (Tinto, 1993). A student who perceives a match with this environment
will enjoy being in it and will have a stronger loyalty to the institution and getting a
degree from that institution. Indeed, in a review of the literature testing aspects of
Tinto’s (1993) model, Braxton, Sullivan and Johnson (1997) found moderate support
among multi-institutional studies for the propositions that academic integration and social
integration affect commitment to graduation and commitment to the institution,
respectively. Subsequently, Liu and Liu (2000) found that academic integration affects
student retention directly and that academic and social integration influence retention

indirectly through satisfaction.

In the organizational context, the concept of person-organization fit (P-O fit) has
been thoroughly investigated. This term refers to the compatibility between a person and
the organization at which he or she works (Kristof, 1996). It is similar to academic
integration in that they both describe the extent to which an individual is compatible with
a particular aspect of his or her environment (i.e., school or work). Though P—O fit has
been operationalized in different ways, it shares much with the operationalization of
integration in the present study. The academic and social integration items address two
areas: whether one perceives similar goals and interests with and generally feels
comfortable with other students, and whether one feels his or her selected major and
courses are suited to academic goals and interests. P-O fit has been operationalized as
goal congruence with coworkers and as the extent to which organizational systems are

congruent with one’s preferences or needs (Sekiguchi, 2004). If coworkers in the

29

organizational context fulfill a similar role as peers in the university context, and if
majors and courses are considered organizational systems that can match a student’s

preferences or needs, then these conceptualizations are similar.

Theoretically, the antecedents and consequences of P-O fit should be similar to
those in this model. Individuals enter an organization with varying individual
characteristics that affect how an individual interacts with his or her environment.
Indeed, entry characteristics including goals and career decision-making self-efficacy
have been found to affect fit with the organizational environment (Peterson, 2007).
Increased fit should make one more satisfied with the organization in which he or she
works, and also more committed to that organization. This satisfaction and commitment
should make one think less about quitting. Results have indicated that P-O fit is related
to organizational commitment and turnover intentions (Sekiguchi, 2004) and to
satisfaction and turnover (V erquer, Beehr & Wagner, 2003; Kristof—Brown, Zimmerman,
& Johnson, 2005). Though they recognized the consistency in the findings regarding
satisfaction as an outcome of fit, Wheeler, Coleman, Gallagher, Brouer, and Sablynski
(2007) also recognized the problem of only looking at simple relationships between fit
and this outcome. They found that satisfaction mediated the relationship between fit and
turnover. The current model is similar in that fit is leading to satisfaction, but different in
that there are withdrawal cognitions and behaviors mediating the relationship of
satisfaction and withdrawal. Though the exact progression of fit to satisfaction,
satisfaction to turnover cognitions and behaviors, and turnover cognitions and behaviors
to turnover has not been studied in the organizational context, the relationships found

among these variables across studies support the placement of fit in the current model.

30

The evidence presented above supports the contention that academic and social
integration should lead to satisfaction and commitment as part of the process leading to
withdrawal. More specifically, academic integration is hypothesized to lead to academic
satisfaction and both degree and institutional commitment, while social fit is
hypothesized to lead to social integration and both degree and institutional commitment.
The facets of satisfaction and commitment are explained in greater detail in the next

section.

Satisfaction and commitment are not the only variables mediating the relationship
between integration and thoughts of withdrawal. When students experience this
integration it is likely they will be more efficacious with regard to academics. Gore’s
(2006) aforementioned study on self-efficacy leads me to believe it can be a product of
positive academic interactions that would give a student perceptions of integration.
Though he did not test the effect of academic integration on self-efficacy, this theoretical
reasoning suggests it exists. Based on the findings and theory pertaining to self-efficacy,
I believe academic integration will affect persistence via self-efficacy in the manner

specified in the model.

H8a. Academic integration will have positive effects on degree commitment.
H8b. Academic integration will have positive effects on institutional commitment.
H8c. Academic integration will have positive eflects on academic satisfaction.

H8d. Academic integration will have positive effects on academic self-efficacy.

31

H9a. Social integration will have a positive effect on social satisfaction.
H9b. Social integration will have a positive eflect on degree commitment.

H9c. Social integration will have a positive effect on institutional commitment.

H 1 0a. Financial integration will have a positive effect onfinancial satisfaction.
H10b. Financial integration will have a positive eflect on degree commitment.

H10c. Financial integration will have a positive efi'ect on institutional commitment.

Family Support.

The support fiom a student’s family is an important factor contributing to success.
Students encounter problems as they work toward their degree, and having family
members to speak with and receive encouragement from can help them deal with these
problems. Family approval, which is similar though not identical to support, was

included in Bean’s (1981) synthesis of attrition models.

Family support is defined as the motivational, emotional, and financial assistance
a student’s family provides to encourage success in college. When one’s family supports
attending college, a student can reaffirm his or her decision to attend college and receive

encouragement when obstacles are encountered. The student may feel more satisfied

32

with academics, social relations, and finances, and feel more committed to his or her
institution and obtaining a degree. Support can improve academic satisfaction by
improving how one views work and gets through tough assignments. Support can also
improve social satisfaction by providing emotional aid and encouraging the development
of relationships with peers, or can improve financial satisfaction by providing monetary
assistance. Degree and institutional commitment are improved because of

encouragement to continue working toward a goal and to remain at an institution.

The findings in this area have been very consistent. Bean & Vesper (1990) found
support for the inclusion of family support in a model of persistence. Hossler and Stage
(1992) found that parental encouragement and expectations positively affected student
aspirations. Sandler (2000) found that family encouragement had both an indirect effect
on intentions to persist and direct and indirect effects on persistence. Consistent with.
these findings, Cabrera, Nora, and Castaneda (1993) found that encouragement from
family and fiiends was the single best predictor of institutional commitment, and that it
correlated with academic and social integration. Simply put, family support and
encouragement is beneficial for students as they work toward a degree. I have no reason
to believe that family support would result from the entry characteristics in this model,
but the findings pertaining to the relationship between this support and both commitment
and integration lead me to believe it should be entered into the model in the same stage as
the integration variables. While direct relationships have been found between family
support and the outcomes of persistence intentions and actual persistence, this could be
because they were not studied in the context of a larger model with the mediating

variables of satisfaction and commitment. I hypothesize that family support will be

33

positively related to the academic, social, and financial facets of satisfaction and to the

degree and institutional facets of commitment.

H1 1 a. Family support will have a positive effect on academic satisfaction.

H I 1 b. Family support will have a positive effect on social satisfaction.

H11c. Family support will have a positive eflect on financial satisfaction.

H1 Id. Family support will have a positive effect on institutional commitment.

HI I e. Family support will have a positive efi‘ect on degree commitment.

Academic Self-Efficacy.

As was previously discussed, self-efficacy has demonstrated a stronger
relationship with both performance and persistence (Multon, Brown, & Lent, 1991;
Robbins, Allen, Casilla, & Peterson, 2006) when it is measured after a student has been
in college for some period of time (Gore, 2006). Consistent with these findings, I
hypothesize that academic self-efficacy measured after a student has been in college for
some period of time will affect GPA. Self-efficacy was measured at the end of students’
first and third semesters in this study, so by the third semester students should have had
time to integrate and have received feedback on performance. This should help them

develop more accurate self-efficacy beliefs.

H12. Academic self-eflicacy will have a positive effect on GPA.

34

Shocks.

Drawing on Beach’s (1990) image theory, Lee and Mitchell (1994) posited that
employees make turnover decisions in accordance with four distinct paths. In one of the
paths, turnover results from accumulating dissatisfaction rather than a shock. This

dissatisfaction leads to a search for alternatives or an automatic decision to quit.

The other three paths involve shocks. Lee and Mitchell (1994) describe shocks as
“jarring event[s] that [force] people to notice readily available opportunities” (p.71). The
events that fall under this label are quite varied, from random, unexpected, or unlikely
events (such as winning the lottery or getting cancer) to more common or expected events
(such as getting other job offers or getting married). When one of these shocks occurs,
individuals may act upon a pre-set script that results in an automatic decision to quit. For
instance, becoming pregnant, whether planned or not, could result in a decision to quit
with no deliberation simply because it is the plan one has always had if such an event
occurred. Alternatively, a shock could lead to dissatisfaction with one’s current job. In
this case, either an automatic decision to quit could be made or a search for alternatives
could be initiated. Researchers have generally found support for the proposed pathways
(Lee, Mitchell, Wise, & Fireman, 1996; Lee, Mitchell, Holtom, McDaniel & Hill, 1999),
though the percentage of people who quit who use each path differs by job type (Harman,

Lee, Mitchell, Felps, & Owens, 2007).

It is possible that shocks play a similar role in the student withdrawal process.
Students encounter similar events (e.g. pregnancy, job offers, disease acquisition), so it is

quite possible they would react in a similar manner. To my knowledge, this is a

35

proposition that has not yet been tested. In the present model, three of the four paths in
Lee and Mitchell’s model are accounted for. Accumulating dissatisfaction is already
present in the model, accounting for the first path. Shocks are hypothesized to lead to
decreased satisfaction, but are also directly related to lesser withdrawal, accounting for
two other paths. The only path not accounted for is that in which dissatisfaction leads
directly to a quitting decision. One of the contributions of this model is the
conceptualization of withdrawal cognitions and behaviors; it is expected that these
mediate the relationship between satisfaction and withdrawal, so a direct path from
satisfaction to withdrawal was not included. Thus, the “shocks” variable, which reflects
whether or not a student has experienced at least one of 21 life events, leads both directly
to withdrawal and indirectly to withdrawal through the satisfaction-thoughts of quitting

pathway.
H13a. Experiencing shocks will negatively affect academic satisfaction.

H 13b. Experiencing shocks will positively relate to withdrawal.

Psychological Outcomes

Commitment.

When students integrate into their environment and are efficacious in their ability
to get a degree, they are theorized to have increased commitment. There are two kinds of

commitment relevant to the college context: goal and institutional. According to Tinto

36

(1993), goal commitment is a “person’s willingness to work toward the attainment of
[personal education and occupational goals]” (p. 43), while institutional commitment is
loyalty and a sense of attachment to the particular institution an individual is attending.
When a student’s goal is obtaining a degree, goal commitment takes on the more specific
form of working toward a degree. This can be termed degree commitment. Degree
bommitment may arise from various sources, such as an understanding of the impact a
degree will have on one’s future earnings and quality of life, or from an intrinsic desire to
possess certain knowledge. Similarly, institutional commitment may come about in
different ways, such as stemming from parental attendance at a university or a desire to
have a recognizable university’s name on one’s degree. These are distinct aspects of
commitment; a student with high goal commitment but low institutional commitment

may drop out of one university but continue at another.

In the I/O psychology literature, organizational commitment is a construct that is
similar to institutional commitment. Organizational commitment is divided into the
facets of affective, continuance, and normative commitment (Allen & Meyer, 1990). As
Allen & Meyer (1990) explain, affective commitment refers to the extent to which and
employee “identifies with, is involved in, and enjoys membership in, the organization”
(p. 2). Continuance commitment refers to an employee’s assessment of the costs
associated with quitting. Normative commitment refers to perceived obligations an
employee has to an organization. Employees are theoretically committed to the

organization in which they work by a combination of these three aspects.

In a meta-analysis investigating the antecedents and consequences of these three

facets of organizational commitment, Meyer, Stanley, Herscovitch and Topolnytsky

37

(2002) found that these three facets correlate with both withdrawal cognitions and
turnover. Affective commitment demonstrated the strongest relationships with these
outcomes, followed by normative and then continuance commitment. When
organizational commitment has been investigated as a singular construct, this variable has
consistently demonstrated a negative relationship with intent to leave (Williams & Hazer,
1985) and withdrawal cognitions (Cohen & Freund, 2004). In a meta-analysis on
organizational commitment and turnover, Cohen (1993) found an average correlation of -
.16 to -.33 between the two variables depending on the instrument used to measure
commitment. In a meta-analysis on employee turnover, Griffeth, Hom & Gaertner
(2000) found a -.23 average correlation across 67 studies between organizational
commitment and employee turnover. These relationships lend support for the inclusion

of commitment in the current model.

The effects of both degree and institutional commitment are relatively well-
established in the student persistence literature. Cope and Hannah (1975) reviewed
research in the area and concluded that commitment to graduating is the best predictor of
persistence. Cabrera, Nora and Castaneda (1993) found that institutional commitment
had the largest effect on intent to persist, while degree commitment was the 3rd highest
(after encouragement from fiiends and family). Braxton, Sullivan and Johnson’s (1997)
review found strong support for the influence of degree commitment on persistence and
moderate support for the effect of institutional commitment on graduation. While direct
relationships have been found between commitment and withdrawal, in this study this
relationship is hypothesized to be mediated by withdrawal cognitions and behaviors

(consistent with Cabrera, Nora, and Castaneda, 1993). That is, it appears more likely that

38

students think less about leaving when they are committed to their institution and
obtaining a degree, rather than making an automatic decision to drop out based upon low

commitment.

It is possible to conceptualize institutional commitment along the three
dimensions of organizational commitment. Students can theoretically feel committed to
their university because of feelings of identification with the college, costs associated
with leaving, or a sense of obligation to stay with the university. The original
institutional commitment instrument was developed to reflect these three facets, but the
subscales did not actually reflect these distinct dimensions in the data collected for the
current study. Because of this, institutional commitment is conceptualized as a
unidimensional construct for the purposes of this study. See the “Measures” section for

more information regarding this scale.
H14a. Goal (degree) commitment will have a negative effect on thoughts of leaving.

H14b. Institutional commitment will have a negative effect on thoughts of leaving.

Satisfaction.

One of the components that Bean (1980;1981) has included in his models is
satisfaction. Bean and Bradley (1986) explain that exogenous variables, such as fit,
social life, and extracurricular involvement affect how one perceives their environment.

People appraise and attach emotion to such perceptions. Satisfaction is one of these

39

emotions. When one enjoys or appreciates the various elements of his or her

surroundings and the college experience in general, satisfaction will follow.

Satisfaction has a long history in the turnover literature, dating back to March and
Simon (1958). Subsequent popular models have also included satisfaction (e. g. Mobley,
1977; Lee & Mitchell, 1994; Horn & Griffeth, 1995). Dissatisfaction can result from a
number of elements in the organizational setting, including a lack of opportunity for
advancement, poor relations with coworkers, or simply growing tired of one’s job. Job
satisfaction measures have traditionally been represented by single-faceted measures
assessing satisfaction with various job characteristics; for example, Horn & Kinicki’s
(2001) measure assessed satisfaction with hours worked, team relations, and work duties.
The relationship between job satisfaction and turnover has long been hypothesized to be
mediated withdrawal cognitions. Mobley (1977) proposed that this dissatisfaction leads
to a series of withdrawal cognitions as employees search for and evaluate alternatives,
compare options to the current job, and intend to quit or stay. Ultimately a behavioral
decision to quit or remain in the current job results from these cognitions. Hom,
Caranika-Walker, Prussia, & Griffeth (1992) conducted a meta-analysis on research into
Mobley’s model, finding support for the idea that dissatisfaction leads to withdrawal
cognitions and eventually to turnover. Williams & Hazer (1986) also found that job
satisfaction leads to turnover intent. Griffeth, Horn & Gaertner (2000) found in their
aforementioned meta-analysis a -.19 sample-size weighted average correlation between
satisfaction and turnover. Again, the importance of satisfaction in the current model is

enhanced by this literature.

40

Liu & Liu (2000) found that low satisfaction led to withdrawal intentions.
Mashbum (2000) found that student dissatisfaction led to greater dropout intentions,
which in turn led to actual dropout. Because of findings such as these, I believe
satisfaction is an important psychological consequence of the aforementioned preceding
variables and a significant predecessor of withdrawal, and inclusion in a larger model is

warranted.

It is possible to conceptualize satisfaction along the same dimensions of fit.
Students can be satisfied with some elements of college life and not with others. Great
teachers and classroom experiences can positively influence satisfaction in the formal
academic area while a lack of quality social experiences can negatively influence
satisfaction in the social area, and vice versa. Again, financial situations can lead to
financial satisfaction or dissatisfaction. In addition to these facets, one can be happy with
life in general. Thus, a measure of satisfaction along academic, social, and financial
dimensions was created and validated. The research does not necessarily support
relationships with these facets, but it does support a hypothesized relationship between
satisfaction and thoughts of leaving. By investigating this established relationship with
the addition of these facets, a greater understanding of the nature of satisfaction and its

effects can be attained.
H15a. Academic satisfaction will have a negative effect on thoughts of leaving.
H1 5 b. Social satisfaction will have a negative effect on thoughts of leaving.

H15c. Financial satisfaction will have a negative eflect on thoughts of leaving.

41

College Grade Point Average.

As has been discussed, GPA is a combination of knowledge acquisition and
performance on tests, projects, papers, etc. While not a purely psychological construct
like commitment or satisfaction, college GPA has demonstrated consistently strong
relationships with persistence (Cabrera, Nora & Castaneda; Pascarella & Terenzini, 2005;
Robbins, Allen, Casillas, Peterson, & Le, 2006). When students do not perform well,
they may think they are not “cutout” for college or that getting a job or transferring to
another institution are better options than continuing in school. As is the case with
commitment and satisfaction, it is likely that students with low GPAs think about leaving
prior to making this decision, rather than automatically dropping out. GPA is therefore
placed in the model in accordance with this theoretical relationship with thoughts of

leaving.

H16. College GPA will have a negative effect on thoughts of leaving.

The Effects of Psychological Outcomes on Withdrawal Cognitions and Behaviors

Two key components of Mobley’s (1977) model were the withdrawal cognitions
and associated behaviors that employees possess before they make a turnover decision.
Mobley (1977) argued that workers do not move directly from dissatisfaction to quitting,
but instead go through intermediary stages before this resultant behavior. When a worker
becomes dissatisfied, thoughts of quitting are stimulated and a search for alternatives may
begin. Recognizing that different researchers conceptualized these elements a little

differently and put them in different orders, Sager, Griffith, & Horn (1998) sought out the

42

best way to arrange them. These authors found support for a revised model of Mobley’s
progression in which thoughts of quitting led to intentions to quit. These intentions led to

searching intentions, which led to turnover.

This is the basis for the progression of withdrawal cognitions and behaviors in the
current model. Thoughts of leaving are at the beginning of stage 4 in the model; when
one has these thoughts, he or she simply ponders the idea of dropping out. These
thoughts are divided into two items. One reflects thoughts about transferring to another
school, while another reflects thoughts about leaving school to obtain employment.
Thoughts of leaving lead to withdrawal intentions; with these intentions, one has moved
beyond simply thinking about the possibility of dropping out and actually intends to do
so. These intentions lead to searching behaviors, in which a student gathers information
on possible jobs or possibly transferring to another institution. Like thoughts of leaving,
search behaviors are divided into two items. The first reflects a student’s search
behaviors regarding transfer to another school, while the second reflects search behaviors

regarding leaving school to obtain employment. Ultimately, these behaviors result in

withdrawal.

These cognitions and behaviors are not as well defined in the student persistence
literature. Typically only withdrawal intentions and actual withdrawal are considered as
outcomes, and many studies have used only either withdrawal intentions (e. g. Jaros,
1997; Braxton, Milem & Sullivan, 2000) or actual withdrawal (e.g. Kahn & Nauta, 2001;
Gore, 2006) as outcome variables. This is somewhat defensible, as there is a strong
relationship between these variables. Tett and Meyer (1993) documented the strong

relationship between intent to turnover and actual turnover. Many studies (e. g. Bean,

43

1981; Cabrera, Nora & Castaneda, 1993) have found that withdrawal intentions are the

best predictors of student withdrawal.

Even so, it is advantageous to employ all of the relevant cognition, intention and
behavior variables in this model. Viewing withdrawal this way is theoretically more
consistent with how people act (Sager, Griffith & Horn, 1998). Dropout decisions can be
spontaneous, but if they are resulting from factors such as poor integration and low
commitment, the thoughts and intentions to perform this behavior will precede the actual
behavior. In addition, withdrawal intent items can give an idea why students decided to
leave. Without a measure that addresses intention it is more difficult to tell if a student
left for a job, to transfer to another institution, or for another reason. Consistent with past
student persistence literature, I am conceptualizing withdrawal intent as a direct precursor
to withdrawal, but consistent with the turnover literature, I am including other potentially

relevant variables in this model.
H17. Thoughts of leaving will positively aflect intent to withdraw.
H18. Intent to withdraw will positively affect search behaviors.

H19. Search behaviors will positively aflect withdrawal.

Other Variables.

While the variables in the later stages of this model are more complete than they
have been in previous research, it was not possible to include everything. For instance,

Bean & Eaton (2000) actually placed in their model academic and social interactions that

44

Tinto only implied existed and occurred. In their model, these interactions lead to
psychological processes such as coping strategies, stress management, and one’s locus of
control. Ideally, every one of these could be measured and included in the current model.
Unfortunately, these were not included in the data collection, at least partially due to the
problems associated with expanding surveys that already include 200-400 items. Just as
Tinto inferred certain interactions and processes existed between prematriculation
characteristics and integration, we can make similar inferences, though further

investigation is warranted to better identify and explain these elements.

This study is also not intended to be exhaustive of the variables included in
organizational turnover models—such an intent would likely lead to a fruitless endeavor.
Even though there are similarities between the collegiate and organizational contexts,
there remain many differences. Griffeth, Horn, & Gaertner (2000) assessed the effects of
variables such as compensation, supervisory relations, promotional opportunities, and job
scope on turnover. Variable such as these are inapplicable simply due to differences
between the contexts (i.e. students are not compensated for schoolwork, cannot be
promoted, etc.). In sum, the hypothesized model is not perfect, but as is explained in the
following section, a test of this model should provide significant contributions to the

student withdrawal literature.

The Present Study

This study adds to the extant literature on student persistence in several ways.

First, as was mentioned, the entry characteristics that have been examined in previous

45

models have been varied and inconsistent. The literature search I undertook leads me to
believe that the entry characteristics included in this model are broader and more
complete than they have been in previous research. Second, these prematriculation
variables can be investigated in a full model of withdrawal that contains intermediary
factors that are drawn from both models of student departure and models of employee
turnover. Cabrera, Nora & Castaneda (1993) demonstrated the importance of combining
elements from Bean’s and Tinto’s models, so doing the same in my model could help
explain more of the variance in attrition. Peterson (2004) demonstrated that it may be
useful to combine elements from these two streams of research, so doing the same again
helps better explain the variance in attrition. In addition, the longitudinal nature of the
data collection allows for investigation of dropout in the first, second, third, and fourth
years; many studies have lacked the ability to do this (e.g Terenzini, Pascarella,
Theophilides, & Lorang, 1983; Cabrera, Nora & Castaneda, 1993; Kahn & Nauta, 2001 ).
Lastly, this study is multi-institutional, whereas many previous studies dealing with
persistence have focused on only one institution (e. g. Mashbum, 2000/2001; Braxton,
Milem, & Sullivan, 2000; Edens, 2006). All together, this study advances knowledge of

models of attrition and helps researchers understand the nature of student withdrawal.

Another reason this study is valuable and adds to the literature in the area is that
many of the measures used were developed specifically for this project by the College
Board research team. It is possible that these measures are better suited for use with
undergraduate students or even better predictors of the assorted elements of the
hypothesized model than previous measures have been. All of the measures used in this

study are described in the Method section below.

46

Method

Procedure

Participants were initially contacted in 2004 before the start of their first year of
college at each of ten universities. They provided responses to paper-and-pencil
measures in the first few days or weeks of their college career by participating in group
sessions supervised by admissions officers or other staff members at the university. We
also collected outcome measures at the end of the students’ first, second, and third
semesters via a web-based survey of all student participants in the original survey. With
the students’ permission, seven of the original 10 universities provided GPA and
graduation data for each of the original participants following the students’ eighth

semester for each of the years they attended the university.

Of the seven universities that provided follow-up data, one is a historically Black
college in the Southeast, five are Big Ten Midwestern universities, and one is a highly
selective private mid-westem school. Approximately 2800 students participated in at
least one data collection, and approximately 2000 attended universities that provided
follow up data. Forty-nine percent of the students in the full sample were White, 22%
were Black, 7% were Asian, and 5% were Hispanic. Fifty-seven percent of the
participants were female and 42% were male. Demographics for each of the data

collections are presented in Table l.

47

Measures

The time at which each of these variables was measured is in Table 2.

Careers, Perseverance, and Knowledge scales. These three biographical data scales were
developed and validated as described in Oswald, Schmitt, Kim, Ramsay, & Gillespie
(2004). Items in these scales pertained to experiences students encountered in high
school and life in general. Each consisted of 10 multiple-choice items with varying
response scales. The careers (or = .79) and perseverance (a = .75) scales demonstrated
moderate reliability, though the knowledge scale (a = .67) was below the traditional .70

criterion level. See Appendices A, B, and C, respectively.

Social/Cultural Involvement scale. To develop this scale, I searched the complete
biographical data inventory developed by Oswald, Schmitt, Kim, Ramsay, & Gillespie
(2004) for items asking about activities involving socializing with others in any way—
volunteering in the community, organizing social gatherings, attending plays or art
exhibitions, etc. Based on item content, I originally identified 18 items. When those
with low item-total correlations were removed, this resulted in a lO-item scale (a = .79).
Four items in this scale were “artistic,” three dealt with “citizenship,” two involved

“cultural appreciation,” and one was designed to measure “leadership.” See Appendix D.

Goal Orientation. Goal orientation was assessed using a new situational judgment
measure created by the College Board team. This measure reflected the three goal
orientation dimensions of mastery orientation, performance-approach orientation, and
performance-avoid orientation. Eight scenarios were given, and each of these scenarios

was associated with 3 or 4 items reflecting a particular behavior or rationale associated

48

with one of the goal orientation dimensions. Participants were asked to rate their
likelihood of responding to each of these items on a 4-point Likert-type scale ranging
from I (definitely not react this way) to 4 (definitely react this way). The 26 total items
in the measure broke down into an 8-item mastery scale (a = .80), a 9-item performance-
approach scale (a = .77), and a 9-item performance-avoid scale (a = .84). See Appendix

E.

Personality. The Big 5 personality traits were assessed using the scales available from
the International Personality Item Pool (Goldberg, 1999). The 10-item scales were used
to measure emotional stability, agreeableness, openness to experience, extraversion, and
conscientiousness. Each item represented a phrase descriptive of a person, such as,
“Make people feel at ease.” Participants rated the extent to which they believed each
phrase reflected their personalities on a 5-point Likert-type scale ranging from 1 (very
inaccurate) to 5 (very accurate). The openness (a = .85), conscientiousness (o. = .90),
extraversion (a = .93), agreeableness (a = .89), and emotional stability ((1 = .92) scale all

demonstrated good reliability. See Appendix F.

F it/Integration. The academic (a = .80) and social (or = .79) subscales are based on scales
found in the literature (Dowaliby, Garrison, & Dagel, 1993; Pascarella & Chapman,
1983; Pascerella & Terenzini, 1980). No existing scale was found for financial fit, so
original items for a financial fit/integration scale were generated; unfortunately, this scale

demonstrated low reliability ((1 = .48). See Appendix G.

Family Support. An 8-item measure of family support (a = .68) was created by the

College Board research team. An example of a family support item is, “My family

49

supports my decision to attend this school.” Response options are on a 5-point Likert

scale ranging from Strongly Disagree to Strongly Agree. See Appendix H.

Degree and Institutional Commitment. Hollenbeck, Williams, and Klein’s (1989)
measure of goal commitment was adapted for the degree commitment scale (a = .77).
Items were modified to reflect the goal of graduating from college specifically. For
example, the item, “It is hard to take this goal seriously” was re-worded to say, “It is hard
for me to take the goal of graduating from college seriously.” The institutional
commitment scale was adapted from Allen and Meyer’s (1990) organizational
commitment scale, which reflected the dimensions of affective, continuance, and
normative commitment. Items were adapted to reflect commitment to one’s university.
This scale was factor-analyzed to discern whether or not the items reflected the three
intended dimensions. It was observed that all but two items loaded well on a single
dimension. When these two items were omitted, the resulting 14-item scale demonstrated
better reliability ((1 = .84) than the continuance (a = .60) and normative (a = .74)
subscales and comparable reliability to the affective subscale (a = .84). Thus,
institutional commitment will be treated as a unitary construct for the purposes of this
project. Response options for both the degree and institutional commitment scales ranged

from Strongly Disagree to Strongly Agree on a 5-point Likert scale. See Appendix I.

Satisfaction. The items in the academic (or = .81), social ((1 = .91), and financial (o. = .83)
subscales were based on scales found in the literature (Smith, Kendall, & Hulin, 1969;
Betz, Klingensmith, and Menne, 1970; Reed, Lahey, and Downey, 1984). The original
41-item scale was pilot tested on a group of 109 students and factor-analyzed. Based

upon these analyses and subsequent sorting tasks, the scales were refined in order to

50

create the resultant l8-item measure. Participants responded to these items on a 5-point

Likert-type scale ranging from “strongly disagree” to “strongly agree.” See Appendix J.

Self-efficacy. Academic self-efficacy was measured with a 4-item scale (a = .82) created
by the researchers. Each item was answered on a 5-point Likert-type scale ranging from

1 (“strongly disagree”) to 5 (“strongly agree”). See Appendix K.

Thoughts of leaving. These two items were adapted from the interview questions used by
Lee, Mitchell, Wise, and Fireman (1996). The two items used were: “I am considering
transferring to another school” and “I am considering other job options instead of

continuing in school.” These two items were correlated r = .35. See Appendix L.

Intent to Withdraw. Intentions to withdraw from the university were collected after
participants’ first, second, and third semesters. Students’ intentions to drop out or
transfer were assessed using three self-report items on a 5-point Likert-type scale ranging
from 1 (strongly disagree) to 5 (strongly agree). The intent to withdraw scale (a = .76)
was adapted from the intent to withdraw/turnover scales by Eaton and Bean (1995) and

Griffeth and Hom (1988). See Appendix M.

Search behaviors. As was the case with the items assessing thoughts of leaving, the
search behavior items were adapted from the interview questions used by Lee, Mitchell,
Wise, and Fireman (1996). These two items were: “I am gathering lots of information
about other schools I could transfer to” and “I am gathering lots of information about job
options as opposed to continuing in school. These two items were correlated r = .43. See

Appendix N.

51

Shocks. The shock scale consists of 21 items, each representing a shock that could affect
an average college student. This list was generated by the College Board team from team
members’ own experiences and observations of college life, from an interview with a
university counselor, and from a focus group conducted with undergraduate students.
Mitchell and Lee suggest that a shock need not be experienced by the individual him or
herself, but that the experience of a shock by a close “other” can be enough to have an
impact. Because of this, the shock measure contained an option in which individuals
indicated whether or not a shock had happened to someone close to them. However, it
could not be theoretically justified why most of the shocks happening to a friend or
relative would affect persistence or satisfaction in college. Therefore, for each shock,
only the item that asked students whether the shock has happened to them or not was
used. The shock variable is a binary value; that is, a person who experienced one or more
shocks will receive a value of “1 ,” while a person who did not experience a shock will

receive a “0.” See Appendix 0.

GPA. GPA was collected via two methods. First, at waves 2, 3 and 4 of data collection,
GPA was assessed with a single-item self-report measure. Self-reported GPA has been
found to correlate highly with actual GPA (Cassady, 2001; Gray and Watson, 2002); in a
meta-analysis, Kuncel, Crede and Thomas (2005) found a .90 correlation between self-
reported and actual GPA. The item assessing GPA was, “What is your current
cumulative GPA?” Second, following the participants’ eighth semester, school-reported
GPA was provided by the universities for each of the four years. Both self-reported GPA

and school-reported GPA were utilized for reasons described in the Analyses section.

52

The GPA data likely differ across institutions, so these data were corrected for this to

enhance interpretability.

Withdrawal. As was noted, following the participants’ eighth semester, the universities
were asked to provide GPA and graduation data for the students for each of the four
years. These data were used to compute withdrawal variables for each of the four years.
For example, students are considered dropouts of Year 3 if the university indicated that
they had not graduated, they had valid GPA data for Years 1 and 2, and they were
missing GPA data for Years 3 and 4. In addition, an overall withdrawal variable was

computed to indicate students who dropped out versus persisted after any year.

Analyses

First, confirmatory factor analyses were conducted to better understand the
measurement properties of the data. Next, structural equation modeling in AMOS 17 was
used to test the hypothesized model. Two tests of this general model were conducted.
The first year of college has been identified as a critical time period in student
development (Hull-Blanks et al., 2005) and as the year when most students drop out
(ACT, 2008). Nearly 12% of students in this sample dropped out in this first year, which
is almost half of the total percentage of students who dropped out. Thus, the first analysis
focused on attrition in the first year, and is hence termed the “early attrition model”. In
the first analysis, withdrawal at the end of student’s first year was predicted using the

entry characteristic data (collected at the beginning of college), the mediating variable

53

data collected toward the end of students’ first semester, and withdrawal data for the first

year.

While nearly all of the variables (besides entry characteristics) used in the early
attrition model were measured in the second wave, academic self-efficacy was not.
Because of this, academic self-efficacy data from the first wave were used. Also,
because withdrawal was calculated using first-year GPA data, individuals who dropped
out did not have school-reported GPAs for this year. Thus, for this model test the self-

reported GPA data collected in the second wave were used.

The second test was designed to test attrition in a more longitudinal manner and is
termed the “longitudinal model.” In this model, the entry characteristics, the integration
data, and family support were the same as in the early attrition model. That is, the entry
characteristics were measured at wave 1 and the integration and family support data were
measured at wave 2. School-reported GPA from the first academic year was used.
Academic self-efficacy, commitment, and satisfaction were measured during wave 3, and
withdrawal cognitions and behaviors were measured during wave 4. The second, third,
and fourth years had dropout rates of 4.7%, 2.9%, and 3.8% respectively. Because of
these low attrition rates, the test of the longitudinal model was conducted by combining
withdrawal data from these three years. By using data collected across four different
time points, it is possible to better assess how the process leading up to withdrawal occurs

over a longer span of time in one’s undergraduate career.

54

Results

Preliminary Analysis of the Voluntary Nature of Withdrawal

The aim of this study was to investigate voluntary withdrawal. If a student is
required to drop out by his or her university, the psychological process leading to
withdrawal would not occur. Universities typically have a probationary system set up for
students with low GPAs. For instance, at one of the universities in this study, students
with a GPA lower than 2.00 for a semester are placed on probation (Michigan State
University, 2009). If the student has a cumulative GPA below 2.00 in the probationary
semester, the student is placed on final probation. If the student does not raise his or her
GPA above 2.00 in this final probationary semester, dismissal may result. If the student
is making progress toward an acceptable GPA, however, a decision can be made to retain

the student.

Because the process varies between universities, and because decisions can be
made on a case-by-case basis despite general guidelines for administrators involved in the
probation and dismissal process, it is not possible to know whether or not students in this
sample were forced to withdraw. However, of the students who dropped out during the
first year, only 14 had first-semester GPAs low enough to be placed on academic
probation (assuming a barrier of 2.0 for this to occur). Only seven students had a GPA
below 2.00 during the second semester. Only two students had these low GPAs in
successive semesters. Twenty-eight students who dropped out in the second year had
GPAs below 2.0 for the second and third semesters. .While it is not possible to know if

these students were dismissed, the low number of students with GPAs likely to result in

55

probation and dismissal indicate that it is unlikely students who were dismissed had a

large impact on the results.

Descriptive Statistics

Means, standard deviations, alpha reliabilities, and intercorrelations among the
scales are presented in Tables 3 and 4. Table 3 presents these statistics for the early
attrition model, while Table 4 presents them for the longitudinal model. Prior to delving
into a complete analysis of these statistics, a few issues were noticed. Some variables of
a similar nature were strongly correlated. For instance, mastery goal orientation was
correlated r=.67 with performance approach goal orientation and r=-.57 with performance
avoid goal orientation. Intent to turn over was correlated r=.68 with information
gathering behaviors and r=.70 with thoughts of leaving; information gathering behaviors
and thoughts of leaving were correlated r=.84. Such correlations could indicate that some
variables are not represented with the appropriate structure, be it due simply to low
discriminant validity among similar constructs or a product of the common method used
to measure such variables. In order to increase confidence in the structure of the

variables in this model, a number of confirmatory factor analyses were undertaken.

Confirmatory Factor Analyses

In deciding what input to supply for these confirmatory factor analyses, it was
realized that the sheer number of parameters in the model could be problematic. Models

with a large number of parameters to be estimated can make “specification and empirical

56

evaluation unwieldy” (Liang, Lawrence, Bennett, & Whitelaw, 1990). To prevent this
from being an issue, item parcels were used as opposed to single-item indicators. For
scales with more than five items, sets of three items were grouped together to form a
parcel. The cutoff number of five items was chosen with the goal of maintaining
information for smaller scales (i.e., those with five items or fewer), but reducing scales
with more items to a manageable size. As an example, the mastery, performance
approach, and performance avoid goal orientation scales each consisted of nine items.
After item parceling, each of these latent variables was represented with three indicators
that were averages of three items each. For scales that had an odd number of items, the
extra item was assigned to the last existing parcel. For instance, the scales measuring
conscientiousness, agreeableness, and emotional stability were each comprised of 10
items. The first and second parcels for these constructs represented three items, while the
third parcel represented four items. Items were not ordered in any systematic manner in
the original scales, so parcels were composed of items next to each other in the survey
instrument to approximate the RAND method (Landis, Beal, & Tesluk, 2000). In the
end, item parcels ended up being used for all constructs except family support, academic
self-efficacy, and withdrawal cognitions and behaviors. The final collection of latent
variables (after revisions based upon the CF As below) had between three and eight
indicators; the reasons for some of these variables having sucha large number of

indicators will be explained later.

A series of confirmatory factor analyses (CFAs) were then conducted to assess the
structure of the variables in this model. These were conducted using data from the first

(for entry characteristics and academic self-efficacy) and second (for all other variables)

57

waves of data collection. These data were used because they represented the largest
samples obtained for their respective variables. For some of the CFAs, variables of a
similar nature were analyzed together because, though conceptually distinct, they may
not have demonstrated uniqueness in practice. For instance, academic integration and
social integration are both forms of integration. In theory they should not be highly
correlated because they tap different, independent life domains. However, when these
measures are administered the perception of one may affect an individual’s perception of
the other. They also were measured using similar self-report scales. If it appears that
individuals were not able to discriminate between the two measures, combining them will
prevent multicollinearity issues in the structural equation model. Also of note, multiple
researchers (e. g. McDonald & Ho, 2002; Garson, 2009) have recommended reporting
three fit indices to assess CFAs or SEMs: the chi-square, the CF I (or other baseline fit
measure), and RMSEA. The results of these CFAs as well as the results for the structural

equation models in the next section follow such recommendations.

Stage 1 Variables.

The first set of CFAs assessed the structure of the background (i.e. stage 1)
variables. First, the mastery, performance approach, and performance avoid goal
orientation variables were represented as three separate constructs in an initial test with
the parcels measuring each as their indicators. Results indicated that fit was slightly less-
than-desirable (12(24)=1029.40, p<.01, CFI=.91, RMSEA=.12), with indicator loadings

ranging from 59-89. Alternate configurations of these facets (e.g. with mastery goal-

58

orientation and performance-approach goal orientation loading on one latent variable, and
with all three facets loading on one variable) were tested but did not result in better fit.

For this reason, the original representation with three facets was retained.

Next, the personality variables of conscientiousness, agreeableness, and emotional
stability were tested. These were represented as separate latent variables with the parcels
measuring each as their indicators. The results indicated reasonable fit (12(24)=368.15,
p<.01, CFI=.96, RMSEA=.07). Indicator loadings ranged from .62-.84 and the three
latent variables were correlated with each other r=.33 or below. This structure was

retained.

The remaining entry variables of knowledge, career orientation, perseverance, and
social/cultural involvement were subjected to another CFA. All of these were measured
using biographical data scales but, as has been discussed, are distinct constructs. They
were treated as separate constructs in a single analysis with the parcels measuring each as
their indicators. Indicator loadings ranged from .58 to .86, and the results indicated good
fit (x2(48)=354.48, p<.01, CFI=.97, RMSEA=.05). These structures were used in further

analyses.

Stage 2 Variables.

The first analysis conducted involving second stage variables assessed the
structure of the fit/integration data. In this test, academic, social, and financial
integration were represented as three separate constructs with the parcels (for academic
integration) and items (for social and financial integration) as indicators. While this

59

structure demonstrated good fit (x2(32)=2o3.22, p<.01, CFI=.94, RMSEA=.04), there
appeared to be issues with the financial integration scale. The standardized loadings of
two of the items were extremely low (b=~.30), and the loading of the third was
excessively high (b>l ). Removing either of the first two items made the structure
unidentified, while removing the third item lowered the loadings of the other two items to
.28 and .12. This information was considered in conjunction with the previously reported
poor reliability, and it was decided that what this scale was measuring was unclear.
Because of this, it was dropped from further analyses. The remaining social and
academic integration scales were tested as separate constructs in a subsequent analysis
and indicated good fit (x2(13)=99.88. p<.Ol, CFI=.97, RMSEA=.05) with indicator
loadings ranging from 59-85. Because of the moderate correlation between the two
constructs (r=.46), an alternative model was tested in which all indicators loaded on a
single fit/integration construct. This resulted in poorer indicator loadings which ranged
from .45-.68 and much worse fit than the original (x2(l4)=826.34, p<.01, CFI=.70,
RMSEA=.14). Because of this, the original model with separate social and academic

integration constructs was retained.

The second CFA in this stage assessed the structure of the academic self-efficacy
data. The analysis indicated reasonable fit (x2(2)=37.39. p<.01, CFI=.99, RMSEA=.08)

and items loaded 54-88 on the latent variable, so this structure was accepted.

A CFA regarding the family support variable revealed that the third indicator had
a very poor loading (.17). A subsequent CF A using the items as indicators revealed that
the three items that composed this third parcel had loadings ranging from .10-.40. In

reviewing the content of the items, the reasons for these poor loadings became clear. The

60

first five items were more directly related to the support a student received from his or
her family, each beginning with “My family supports/supported. . .” The last three items
assessed how much one’s family wanted him or her to attend the school one actually
attends, how much the family would care if the student transferred, and whether going to
the school was a family tradition. While these items may aid in understanding the nature
of the relationship a student has with his or her family, they are qualitatively different
than basic support a family can provide. These three items were removed and the
structure was tested again. The resulting S-item structure demonstrated reasonable fit
(x2(5)=69.94, p<.01, CF I=.95, RMSEA=.07), while item loadings ranged from .59-.73.

This structure was retained for further analyses.

Stage 3 Variables.

A CF A was conducted to assess the structure of the satisfaction and commitment
data. Testing a model in which the latent variables have only two indicators each can
result in underidentification and unreliable error estimates (Garson, 2009). Because the
facets of satisfaction had only two indicators each, they were tested along with
institutional and goal/degree commitment. The facets of satisfaction and commitment
were treated as separate constructs in a single analysis with the parcels for each serving
as indicators (i.e. academic, social, and financial satisfaction were treated as separate
constructs, as were goal and institutional commitment). Indicator loadings ranged from
.68 to .93 for the satisfaction facets and from .50 to .81 for the commitment facets. The

results indicated good fit (x2(67)=219.96, p<.01; CF I=.97; RMSEA=.03). Two

61

exploratory analyses were conducted to assess the possibility that all three facets of
satisfaction were better represented by one construct or that both commitment facets were
better represented by one construct. In the satisfaction analysis the commitment data
were removed, and in the commitment analysis the satisfaction data were removed. Both
the test involving satisfaction (x2(9)=1233.84, p<.01, CFI=.59; RMSEA=.22) and the test
involving commitment (x2(21)=780.27, p<.01, CFI=.69, RMSEA=.11) resulted in poorer
fit. The original structures for these variables were intended to be retained in firrther
analyses, but out of necessity the facets for each were combined into single satisfaction
and commitment variables in the full model; the reasons for this are described later in this

section.

Stage 4 Variables.

A CF A was conducted to assess the structure of the withdrawal cognition data.
Information gathering, withdrawal thoughts, and withdrawal intent were treated as
separate constructs in a single analysis with the items for each as indicators. The results
indicated poor fit (x2(11)=1170.76, p<.Ol , CF l=.75; RMSEA=.20). Grouping all of the
items together as indicators of a single construct did not result in a better fitting model
(x2(14)=1377.80, p<.01, CFI=.71; RMSEA=. 19). However, it was observed in the
correlation matrix of these items that the job-related items from each of these scales were
related to each other, and that the transfer-related items were related to each other as well.
A third model was tested in which the job-related items represented one construct and the

transfer-related items represented another. Creating this structure involved removing the

62

item about general intent to leave the university, as this item was related to both job-
related withdrawal and transfer-related withdrawal. This model resulted in greatly
improved fit (x2(8)=188.54, p<.01, CFI=.95, RMSEA=.09). This structure was retained
for use in subsequent analyses, but like the satisfaction and commitment data, had to be
modified for use in the full model. The reasons for this arose in the CFA of the full

model, which is described next.

Lastly, a CF A was conducted with all of the variables in the model to determine
the overall fit of the data to the measurement model. Significant issues arose in
conducting this analysis. The satisfaction, commitment, and job and transfer constructs
were unidentified, despite several “rules of thumb” (i.e. the model being recursive and
having positive degrees of freedom, each of the latent variables having a scale) indicating
the model should have been identified. While these general guidelines are useful, they are
not always sufficient for empirical identification (Kline, 2005, p. 107). One possibility
for this is Heywood cases (i.e. negative error variance estimates), but there was no
evidence of this. Another possible reason for this underidentification is multicollinearity.
When the significant correlations among these multiple unidentified variables were
considered together, multicollinearity appeared to be problematic. That is, academic,
social, and financial satisfaction were correlated with each other to a degree that
prevented convergence of the measurement model, as were institutional and degree

commitment with each other and job and transfer thoughts and behaviors with each other.

To deal with this problem it was necessary to combine facets of constructs. At
first, only the satisfaction facets were combined, but this did not result in convergence of

the model. Combining only the commitment facets or only the withdrawal cognitions

63

and behaviors produced the same unsatisfying result. Combining any two of the three
problematic groups of variables also did not result in model convergence. In the end, it
was necessary to combine the facets of each of these problematic groups of variables.
This resulted in composite constructs of satisfaction, commitment, and withdrawal
cognitions/behaviors. Item parceling resulted in a latent variable representing satisfaction
with six indicators: two academic, two social, and two financial. The latent variable
representing commitment had three goal commitment parcels and five institutional
commitment parcels. The latent variable representing withdrawal cognitions and
behaviors was comprised of the three items related to transfer cognitions and behavior
and three items related to job cognitions and behavior. After combining these facets, the
CFA for the measurement model was identified and demonstrated a good RMSEA but a
poor CFI (x2(1817)=10385.29, p<.01, CFI=.84, RMSEA=.04). Interpretation of this
difference in fit indices is provided later when the full model is discussed. While the
resulting model is not optimal and decreases fit to some extent, it has greater parsimony
and is not at odds with the theoretical bases for the various paths that were affected. This

revised model structure was used in the subsequent analyses.

Revised Descriptive Statistics

Means, standard deviation, a reliabilities, and intercorrelations among the revised
' variables are presented in Tables 5 and 6. Table 5 provides these statistics for the early
attrition model and Table 6 provides them for the longitudinal model. Reliability was
generally acceptable for variables in both models, though the knowledge (a=.63) and

satisfaction ((r=.68 in the early attrition model and .63 in the longitudinal model)

54

variables demonstrated low reliability. It is not surprising that the reliability of the
satisfaction scales were poor, given the need to combine facets to facilitate model
convergence. Even after combining facets, however, the commitment and withdrawal

cognitions and behaviors scales demonstrated acceptable reliability.

Early Attrition Model Correlations.

As was discussed, the early attrition model utilizes background variables and
academic self-efficacy data from the first data collection, but all other variables in the
model are from the second data collection. The zero order correlations for this model
provide preliminary support for many of the hypotheses. Hypotheses were supported by

a statistically significant r (p<.05).

Mastery goal orientation, performance approach goal orientation, knowledge,
perseverance, career orientation, agreeableness, conscientiousness, and emotional
stability were all positively related to academic integration. These correlations ranged
fi'om .09 to .30. Performance avoid goal orientation was negatively related to academic
integration (r=-.09.) Agreeableness and emotional stability were positive related to social
integration (r=.17 and r=.23 respectively), but contrary to hypotheses, conscientiousness
and social/cultural involvement were not related to social integration. In total, 12 of the
14 hypothesized links between stage 1 and stage 2 were supported by the zero-order

correlations.

Academic integration, social integration, and family support demonstrated
relatively strong relationships with satisfaction, with correlations of .44, .62, and .31

65

respectively. These variables also demonstrated similarly strong relationships with
commitment (academic integration r=.44; social integration r=.53; family support r=.28).
Academic integration was significantly related to academic self-efficacy (r=.21), but
academic self-efficacy was not related to GPA. The shocks variable was related to
satisfaction (F-. l 7) but not to actual withdrawal (r=-.05). In total, eight of the 10 links in

this section of the hypothesized model were supported by the zero-order correlations.

In the next stage of the model, satisfaction, commitment, and GPA all
demonstrated significant relationships with the withdrawal cognition and behavior
variables (r=-.48, r=-.59, r=-.21 respectively). Thus, all three of the links in this section
of the model were supported at the bivariate level. Lastly, withdrawal cognitions and
behaviors were not related to actual withdrawal. Across the entire model, the correlations
provided support for 23 of the 28 hypothesized relationships. The exceptions were the
relationships between conscientiousness and social integration, social/cultural
involvement and social integration, academic self-efficacy and GPA, shocks and
withdrawal, and withdrawal cognitions and behaviors and actual withdrawal. The SEMs

described below consider support for the model as a whole.

Longitudinal Model Correlations.

Each stage in the longitudinal model corresponded to a different data collection
time point (i.e. stage 1 utilized data from the first data collection, stage 2 utilized data
from the second, stage 3 utilized data from the third, and stage 4 utilized data from the

fourth). As a result of this strategy, variables from stage 1 and stage 2 were the same in

66

the longitudinal model as in the early attrition model. The resulting intercorrelations
between variables were the same in both models, so the previous section should be

consulted for discussion of these correlations.

The next stage is where this model begins to differ from the early attrition model.
Academic integration, social integration, and family support were again related to
satisfaction (r=.3 8, r=.48, and r=.27 respectively) and to commitment (r=.29, r=.44,
r=.25). Academic integration was related to self-efficacy (r=.26), but unlike in the
previous model, self-efficacy was in turn related to GPA (r=.17). Shocks were related to
satisfaction (r=-. 14) and withdrawal (r=.09). All of the 10 links in this section of the

model were supported by these correlations.

Satisfaction, commitment, and GPA again demonstrated significant relationships
with withdrawal cognitions and behaviors (r=-.35, r=-.3 7, and r=-.21 respectively).
Lastly, withdrawal cognitions were positively related to actual withdrawal (r=.23). Thus,
26 of the 28 hypothesized relationships in the longitudinal model were supported by the
zero-order correlations. The exceptions were the relationships between conscientiousness

and social integration and between social/cultural involvement and social integration.

Structural Equation Models

To test the hypothesized models, structural equation model (SEM) analyses were
conducted in AMOS 17. Figure 2 displays the measurement model that was identified
through the CF As in the previous section. Due to the size of the model, the portion of the
measurement model concerning the entry characteristics is displayed on one page, while

the rest of the variables are displayed on the next page with the entry characteristics

67

represented only by their latent variables. Figure 3 displays the structural model. In this
model, the paths between constructs are represented by one-headed arrows. In addition,
disturbances have been added to the endogenous variables to account for unmodeled

determinants of these variables.

It should also be noted that in the actual tests of these models, the exogenous
variables (i.e. all of the entry characteristics, family support, and shocks) were free to
covary. However, due to the high number of exogenous variables in the general model,
graphically representing the correlations between these variables resulted in this portion
of the figure being unreadable, so these correlations are not presented in the figure.
Figure 4 represents the full model, which is a combination of the measurement and
structural models. As was the case in the measurement and structural models, the
structural portion representing the entry characteristics and the correlations among
exogenous latent variables are not displayed in this figure despite being included in tests

of the model. The correlations among these exogenous variables are in Tables 7 and 8.

In the SEM analyses that were conducted, one factor loading on each latent
variable was fixed to a value of l for sealing purposes. Full-information maximum
likelihood was used to handle missing data, as this technique has demonstrated better
versatility and performance than other methods of handling missing data (Carter, 2006;
Kline, 2005). The use of full-information maximum likelihood is especially
advantageous in comparison to other methods when subjects drop out of a longitudinal
study, both when this nonresponse is random and when it is due to a low score on a
measured variable (Wothke, 2000). In this study, the original sample of students who

completed at least a portion of the wave 1 survey consisted of 2,716 students, meaning

68

this was the number available for analysis. In waves 2, 3, and 4, 1149, 996, and 863
students responded, respectively. The return rate of 42% from wave 1 to wave 2 was
clearly the lowest, but the return rate from wave 2 to wave 4 of 75% was much better.
Because these data were collected during the first two years, it is possible that up to
approximately 16% of the nonresponse rate from wave 1 to wave 4 was a result of

withdrawal.

The first SEM analysis was conducted to assess the fit of the data to the early
attrition model. In this model, as was previously discussed, withdrawal at the end of the
students’ first year was predicted using the entry characteristic data (collected at the
beginning of college), the mediating variable data collected toward the end of students’
first semester (except for the variables of academic self-efficacy and GPA as noted

earlier), and withdrawal data for the first year.

Intercorrelations among the indicators in the early attrition model are presented in
Table 9, while the results for this analysis (as well as for the analysis of the longitudinal
model) are in Table 10. Figure 5 also displays the regression coefficients obtained in this
analysis. The correlations among the exogenous latent variables (in Table 7 for the early
attrition model) indicated high collinearity between some of the exogenous variables.
For instance, perseverance was correlated r=.41 or greater with all of the entry
characteristics except emotional stability. Knowledge was highly correlated with

conscientiousness (r=.56), and many other correlations were above r=.30.

For the paths between stage 1 and stage 2 variables, career orientation,

agreeableness, and emotional stability were all positively and significantly (p<.05) related

69

to academic integration. The standardized regression coefficients were weak to
moderate, ranging from .09 to .35. Performance avoid goal orientation was negatively
related (B=-.15, p<.05) to academic integration. Agreeableness and emotional stability
were also related (p<.05) to social integration, with regression coefficients of .22 and .27
respectively. In the next stage of the model, academic integration was related to both
academic self-efficacy and commitment, and social integration was related to both
satisfaction and commitment. These relationships were moderate to strong, ranging fiom
.23 to .88. Family support was related to commitment and shocks were related to
satisfaction, but these were weaker than the other relationships in this stage, with
standardized regression coefficients of .13 and -.06 respectively. In the stage 3-stage 4
relationships, GPA, satisfaction and commitment were related to withdrawal cognitions
and behaviors, with standardized regression coefficients of -.24, -.09 and -.62
respectively. Lastly, withdrawal cognitions and behaviors were weakly related to actual
dropout (B=-.07). Fit indices for this model and for subsequent model tests are presented
in Table 15. The overall model demonstrated curious fit, with a significant chi-square
(x2(2188)= 13788.13, p<.01) and poor CF I (.80) but with a good RMSEA (.04). Possible

explanations for the difference in these fit indices are provided later.

The squared multiple correlations for the endogenous variables in this model are
presented in Table 11. Academic integration (R2=.31), social integration (R2=.l3), and
academic self-efficacy (R2=.22) were predicted moderately, while satisfaction (R2=.78),
commitment (R2=.44), and withdrawal cognitions and behaviors (R2=.51) were predicted
more strongly. GPA (R2=.00) and withdrawal (R2=.01), each having only one associated

correlate, were not predicted at a significant level (p<.05).

70

The second analysis was designed to test relationships among these variables over
a more extended time frame. In this analysis, the entry characteristics used data from the
initial data collection, the integration-stage variables used data from the second data
collection, the satisfaction/commitment-stage variables used data from the third data
collection, and the withdrawal cognitions and behaviors were from the fourth data
collection. Because of low attrition rates in the second, third and fourth years,

withdrawal data from these years were combined to form a composite.

Intercorrelations among variables in the longitudinal model are presented in Table
12. As was mentioned, results from this analysis are presented in Table 10. Figure 6 also
displays the regression coefficients found in this analysis. Collinearity among exogenous
latent variables (displayed. in Table 8) was similar to that in the early attrition model. In
the first stage of the longitudinal model, knowledge, career orientation, agreeableness,
and emotional stability were positively related (p<.05) to academic integration. These
relationships were weak to moderate, ranging from .12 to .41. Mastery goal orientation
and perseverance were negatively related (p<.05) to academic integration, with betas of -
.26 and -.19, respectively. Neither of these negative relationships was consistent with
expectations, but the positive zero-order correlations associated with each suggested that
collinearity affected these regression coefficients. In addition, agreeableness and
emotional stability were related to social integration, with betas of .20 and .25

respectively.

In the next stage of the model, academic integration was related to academic self-
efficacy, satisfaction, and commitment, social integration was related to satisfaction and

commitment, academic self-efficacy was related to GPA, and family support was related

71

to commitment and satisfaction. These relationships ranged from fairly weak (e.g., both
of the relationships involving family support were below .13) to strong (e. g., both of the
relationships involving social integration were above .55). In the third stage of the
model, GPA, satisfaction, and commitment were all related to withdrawal cognitions and
behaviors. Satisfaction was weakly related (b=-.10), but GPA and commitment
demonstrated stronger relationships ([3 =-.45 and B =-.41, respectively). Each of these
negative coefficients are in the hypothesized direction. Finally, withdrawal cognitions
and behaviors demonstrated a strong relationship with actual dropout ([3 =.45). The
overall model displayed a similar pattern of fit to the early attrition model (x2(2188)=

10786.96, p<.01, CFI=.83, RMSEA=.04).

The squared multiple correlations for the longitudinal model are also presented in
Table 11. Academic integration (R2=.21), social integration (R2=.11), academic self-
efficacy (R2=.10), satisfaction (R2=.49), commitment (R2==.39), and withdrawal
cognitions and behaviors (R2=.42) displayed a similar pattern to the squared multiple
correlations in the early attrition model, but they were all slightly lower. The two
multiple correlations that were not significant in the early attrition model, however, were

significant in the longitudinal model (GPA R2=.05, withdrawal 18:20).

There were some notable differences between these models. Despite having
variables from the same waves of data collection in stages 1 and 2 of the models, the
patterns of significance were slightly different. The relationship between performance
avoid goal orientation and academic integration was significant in the early attrition
model but not in the longitudinal model, while the relationships between mastery goal

orientation and academic integration and between knowledge and academic integration

72

were significant in the longitudinal mode] but not in the early attrition model. This
indicates that the additional variables in later stages of the model correlated with the
variables in stages 1 and 2 in ways that either enhanced or diminished relationships
between variables at these initial two stages. Aside fi'om these changes in statistical
significance, the magnitude of the significant relationships between these stages was

similar, differing at most by .06.

The relationships between stage 2 and stage 3 were also fairly consistent across
the models but there were a few differences. The shocks variable was weakly related to
satisfaction in the early attrition model but not in the longitudinal model. Family support
and academic integration were related to satisfaction and GPA was related to academic
self-efficacy in the longitudinal model but not in the early attrition model. This may have
been due to the use of self-efficacy data from wave 1 in the early attrition model; at this
time point, students had not had as much time to develop accurate self-efficacy beliefs as
at later time points. The difference in betas was noticeable for the relationship between
social integration and satisfaction (.21) and for the relationship between academic

integration and academic self-efficacy (.15), but it was .07 or less for all other significant

relationships.

In the next stage, all three hypothesized relationships were supported in both
models, though the differences in relationship strength between GPA and withdrawal
cognitions and behaviors and between commitment and withdrawal cognitions and
behaviors (both .21) were noticeable. In the final stage of the model, withdrawal

cognitions and behaviors were weakly related to dropout in the unexpected, negative

73

direction in the early attrition model, but were positively and more strongly related to

dropout in the longitudinal model.

The overall fit indices give a reasonable idea of the adequacy of the models as a
whole. The relative chi-square (i.e. the ratio of the chi-square value to the degrees of
freedom) is less than the upper bound of 5.0 that has been provided as a guideline for
good fit (Kline, 2005, p. 137) for the longitudinal model (relative chi-square=4.93) but
slightly above this range for the early attrition model (relative chi-square=6.30). It
should be noted, though, that the chi-square statistic frequently produces Type II errors
and is often disregarded by researchers in favor of other fit indices when poor (Garson,
2009). For both models, the CFI values were below what is generally considered
acceptable, while the RMSEA values were within the limits of good fit (Hu & Bentler,
1999). This difference indicates that while the models do not do well when compared to
a baseline model in which indicators and latent variables are uncorrelated, the data do fit
the hypothesized models acceptably when degrees of freedom are taken into account.
That is, the data do not fit the hypothesized model that well in the absolute sense, but
when a proxy for the complexity of the model (i.e. the number of parameters to be
estimated) is taken into account, the data fit reasonably well. It has also been suggested
that RMSEA be relied upon more in confirmatory contexts whereas CFI be relied on
more in exploratory contexts (Rigdon, 1996). Given this recommendation, it might be

more appropriate to rely on the RMSEA in this confirmatory context.

The measurement model appeared to contribute most strongly to lack of model fit.
The most problematic variables were those that had to be combined due to

multicollinearity. The satisfaction, commitment, and withdrawal cognitions and

74

behaviors constructs were not ideally represented, negatively impacting fit. Also,
variables such as the three aspects of goal orientation did not demonstrate great fit.
Taken together, these variables negatively affected overall fit. The final CFA of all
variables indicated that this was the case, and the addition of the causal paths only served
to maintain similar fit rather than overcoming the existing measurement issues. While
these fit indices present a general picture of these models, it is also useful to examine the

individual paths and their associated hypotheses; the next seetion is such an attempt.

Support for Hypotheses

Figure 1 presents the general model with numbered hypotheses. A summary of
support for each of the hypotheses is presented in Table 13. The correlations provide
support for the bivariate relationships, while the regression weights provide support for
these relationships when they are considered as part of the full model. Support is
classified as “full” if all associated correlations and standardized regression weights
support a hypothesis, “partial” if some but not all evidence supports a hypothesis, “none”
if none of the evidence supports a hypothesis, and “uncertain” if the support for a
hypothesis is unclear due to evidence simultaneously supporting and contradicting (i.e.,
having a sign opposite of what is hypothesized) a hypothesis. The table also displays
whether the support classified as “partial” is due to correlational evidence only (indicated
by “corr.” next to the designation of “partial”) or if the evidence is from correlations and

beta weights (indicated by “mixed” next to the designation of “partial”).

75

Hypotheses 1a, 1b, and 1c stated that the personality factors of conscientiousness,
agreeableness, and emotional stability would be positively related to academic
integration. The correlation with academic integration was significant for
conscientiousness, but neither of the beta weights were, partially supporting Hypothesis
la. The correlations and betas for the latter two personality factors were all significant,
providing full support for Hypotheses 1b and 1c. Hypotheses 2a, 2b, and 2c proposed
similar relationships between the three personality factors and social integration.
Hypothesis 2a (which concerned conscientiousness) received no support, while the
hypotheses concerning the other two variables received full support. Thus,
conscientiousness demonstrated little effectiveness as a predictor, butlagreeableness and
emotional stability demonstrated effectiveness for academic and social integration both
when considered at the bivariate level and when considered in the context of the full

model.

Hypotheses 3a, 3b, and 3c posited that mastery goal orientation (positively),
performance approach goal orientation (positively), and performance avoid goal
orientation (negatively) would be related to academic integration. The correlation
between mastery goal orientation and academic integration provided support for this
hypothesis. Contrary to this hypothesis, however, the beta for the longitudinal model was
negative and significant, indicating uncertainty in the support of this hypothesis. This
indicated suppression in the model that resulted from mastery goal orientation correlating
with other variables. Performance approach goal orientation received support only from
the correlation, indicating partial support for Hypothesis 3b. Performance avoid goal

orientation received support from the correlation and the early attrition model beta,

76

indicating partial support. In Stun, the conjectures that mastery goal orientation would be
related to academic integration received uncertain support, while the hypotheses

regarding performance approach and performance avoid goal orientation received partial

support.

Hypotheses 4, 5, and 6 posited that career orientation, knowledge, and
perseverance (respectively) would be positively related to academic integration. Career
orientation was related to academic integration all possible ways, indicating full support
for Hypothesis 4. This relationship with knowledge was supported by the correlation and
by the longitudinal model beta, indicating partial support for Hypothesis 5. The
relationship with perseverance was supported only by the correlation. Contrary to the
hypothesis, however, the standardized regression weight for the longitudinal model was
in the opposite direction than expected due to multicollinearity, resulting in uncertainty
about the support for this hypothesis. It seems that the hypothesis is supported when the
relationship is considered by itself, but not supported when considered in the context of
the full model. Lastly for this stage, Hypothesis 7 stated that social/cultural involvement

would be positively related to social integration. This hypothesis received no support.

Hypotheses 8a, 8b, 8c, and 8d stated that academic integration would be
positively related to academic self-efficacy, academic satisfaction, institutional
commitment and degree commitment respectively. Hypothesis 8a, which concerned
academic self-efficacy, received full support. Because of the aforementioned issues with
the model, Hypotheses 8b, 8c and 8d had to be modified to test combined satisfaction and
commitment variables. The early attrition model correlation and the betas for both

models were significant for the relationship between academic integration and

77

satisfaction, providing partial support for a modified Hypothesis 8b. All evidence for the
academic integration-commitment relationship was significant, providing full support for

modified Hypotheses 8c and 8d.

Hypotheses 9a, 9b, and 9c stated that social integration would be positively
related to social satisfaction, institutional commitment, and degree commitment
respectively. All of the evidence for these hypotheses was significant, providing full
support for Hypothesis 9a and modified Hypotheses 9b and 9c. Hypotheses 10a, 10b,
and 10c stated that financial integration would be related to financial satisfaction,
institutional commitment, and degree commitment. Because of the elimination of this

variable, these hypotheses were not tested.

Hypotheses 11a, 11b, and 11c predicted that family support would be related to
the three facets of satisfaction, while hypotheses 11d and 1 1e predicted that family
support would be related to the two facets of commitment. For the satisfaction
composite, the correlation for the early attrition model and the betas for both models were
significant, providing partial support for modified Hypotheses 11a, 11b, and 11c. For the
commitment composite, all evidence was significant, providing full support for modified
Hypotheses 11d and 1 1e. Hypothesis 12 stated that self-efficacy would be related to
GPA. Only the correlation and standardized regression coefficient in the longitudinal
model were significant, providing partial support for this hypothesis. Lastly for this
stage, Hypotheses 13a posited that shocks would be related to academic satisfaction,
while Hypothesis 13b posited that shocks would lead directly to withdrawal. For the
satisfaction composite, both correlations and the beta for the longitudinal model were

significant, providing partial support for modified Hypotheses 13a, 13b, and 13c. For the

78

path from shocks directly to withdrawal, only the correlation for the longitudinal model

was significant, providing weak but partial support for this hypothesis.

Hypotheses 14a and 14b predicted a negative relationship between the two facets
of commitment and thoughts of leaving. Both of these were originally conceptualized as
variables with distinct facets, but both were ultimately represented as composites of these
facets. There was full support for modified Hypotheses 14a and 14b. Hypotheses 15a,
15b, and 15c predicted negative relationships between the three facets of satisfaction and
thoughts of leaving. These modified hypotheses were fully supported. Hypothesis 16
predicted a negative association between GPA and thoughts of leaving. Like the others
in this stage, this hypothesis was fully supported. Finally, Hypotheses 17, 18, and 19
predicted a positive relationship between thoughts of leaving and intent to withdraw, a
positive relationships between intent to withdraw and search behaviors, and a positive
relationship between search behaviors and withdrawal. Again, these facets had to be
combined due to multicollinearity, resulting in a single direct relationship from
withdrawal cognitions and behaviors to withdrawal. The betas for both models and the
correlation for the longitudinal model were significant, providing partial support for this

hypothesis.

Model Modification

In an effort to identify a better fitting model, the previously presented information
was considered in conjunction with the modification indices to guide model trimming and

building.

79

There is a delicate balance in using both empirical and theoretical guidance for
modification. As Kline (2005, p. 149) explains, while empirical modification indices can
assist researchers in identifying opportunities for improvement in a model, uncritical
reliance on these indices increases the likelihood of capitalizing on chance and frequently
does not uncover the true model. Theory is needed to guide such respecification. In the
present study, modification indices were consulted, but an attempt was made to maintain
consistency with the general theory behind the model. Prior to modification a few rules
were developed to achieve this aim. Paths between variables in one stage and variables
in the next stage that were not included in the original model were permitted to be added
(e.g. the relationship between social/cultural involvement and academic integration
would be estimated when inclusion of this path was supported by the evidence). Existing
paths in the model were also deleted when prior analyses indicated they were
nonsignificant. No path was permitted to go “backward” (e.g., nothing in stage 2 was
permitted to lead to a variable in stage 1) or to lead to another variable in the same stage,
though other relationships across stages of data collection were estimated if there was
enough evidence to support their inclusion. Estimation of paths that skipped a data
collection stage was not included except in the case of the relationship between entry
characteristics and withdrawal. Because actual withdrawal is arguably the variable of
greatest interest, and because direct relationships between entry characteristics and
withdrawal have been supported in previous research, direct effects were permitted from

these variables to withdrawal.

It should also be noted that although there have been cogent arguments in support

of correlating error terms associated with indicators (e. g. Reddy, 1992; Cole, Ciesla, &

80

Steiger, 2007), this practice was not adopted in the present study. As Landis, Edwards,
and Cortina (2009) explain, there are a number of problems with this practice especially
when conducting post-hoe model modification. While correlating residuals will certainly
increase model fit, the practice largely capitalizes on chance, is rarely replicable, and
essentially rewards a researcher for “what could have been’” (Landis, Edwards, &
Cortina, 2009, p. 193) by accounting for whatever he or she neglected to include in a
model. The debate about correlating residuals will undoubtedly continue in the literature,
but the unresolved problems with the practice are significant enough to warrant exclusion

from model modification in this instance.

Early Attrition Model Modification.

In the early attrition model, the empirically supported modifications mainly
resulted in only small drops in the chi-square statistic. The largest possible
improvements would have resulted from allowing error terms to correlate. Allowing
some pairs of error terms to correlate, such as two of the commitment error terms or two
of the satisfaction error terms, would have dropped the chi-square by 200-400. No other
alterations to the model approached these levels of impact. While some researchers have
followed arbitrary guidelines such as only making modifications where the indices
indicate a drop in chi-square greater than 100, others have simply made the modifications
that result in the largest drop regardless of value (Garson, 2009). In the early attrition

model there were no modifications (permitted under the aforementioned rules) that would

81

have resulted in chi-square drops greater than 100, so by default, the latter strategy was

used.

One reason that there were few supported paths to be added is that the original
model included most of the theoretically possible relationships. Due to the need to
collapse facets of satisfaction and facets of commitment into single variables based on
measurement considerations, few potential paths from stage 2 variables to stage 3
variables remain to be estimated. All of the stage 3 variables are already hypothesized to
relate to withdrawal cognitions and behaviors. The only possibilities for adding paths,
then, exist in the omitted relationships between stage 1 and stage 2 (e.g., from knowledge
to social integration or from social/cultural involvement to academic integration) and the
few omitted relationships between stage 2 and stage 3 (e. g. from shocks to commitment
or from academic integration to GPA). Unfortunately, support for all such paths was

weak.

Despite this, there were some modifications that were empirically supported and
theoretically justifiable. First, academic self-efficacy appeared to work better as a stage 2
variable that resulted from entry characteristics. That is, modification indices indicated
that the chi-square would drop between 12.41 and 84.07 per path for paths added from
the stage 1 variables (i.e. the entry characteristics) to academic self-efficacy. The zero-
order correlations between these variables (presented in Table 5) also appeared to support
this move. Social/cultural involvement was the only entry characteristic for which the
evidence did not support adding a path to academic self-efficacy, with a weak zero-order
correlation between the variables and with a near-zero drop in chi-square adding this

path.

82

While paths closer to the lower end of this range of discrepancy change would not
have been added by themselves, taken as a whole, these changes are theoretically
justifiable. Though a post-hoe explanation, all of these variables were hypothesized to
relate to academic integration, so it is not entirely surprising that they should be related to
another variable that is also of an academic nature. Additionally, both the entry
characteristics and academic self-efficacy were measured at wave 1. Thus, the first step
of modifying this model involved altering the antecedents of academic self-efficacy. The
path from academic integration to academic self-efficacy was removed, and paths were
created from mastery goal orientation, performance approach goal orientation,
performance avoid goal orientation, knowledge acquisition, career orientation,
perseverance, conscientiousness, emotional stability, and agreeableness to academic self-
efficacy. This model demonstrated improved fit over the original model (x2(2180)=
13358.61, p<.01; CFI=.81; RMSEA=.04). Because these models were not hierarchically
nested, they cannot be compared using a chi-square different test. However, they can be
compared using the Akaike Information Criterion (AIC). Lower values of this index
suggest greater parsimony and increased likelihood of replication (Kline, 2005, p. 142).
The revised model (AIC=13832.98) had a lower value on this fit index than the original
model (AIC=14380.13), indicating better fit. In addition, the variance explained in

academic self-efficacy (R2=.28) was greater than in the original model (R2=.22).

After this change, there were no additional variables that displayed relationships
with variables from another stage in a manner that supported similar revision. There was
also little information to support adding more paths. The only permissible path that, if

added, would have lowered the chi-square by more than 25 was from social/cultural

83

involvement to withdrawal. This path would have lowered the chi-square by
approximately 40. Unfortunately, the estimated parameter change was positive for this
path, indicating individuals higher on social/cultural involvement were more likely to
drop out. It could be argued that this path should be added despite being inconsistent
with theory, but as was discussed, simply modifying a model based on empirical
suggestions increased the likelihood of capitalizing on chance and reduces replicability.
With no theoretical justification for adding this path, it was decided to leave it out of the

revised model.

The next step in modifying the model involved trimming paths that were trivial.
In the early attrition model, the paths between four pairs of variables demonstrated both
nonsignificant betas and zero-order correlations: the path from social/cultural
involvement to social integration, the path from conscientiousness to social integration,
the path from academic self-efficacy to GPA, and the path from shocks to withdrawal.
These paths were targeted for removal because, whether analyzed within the full model
or assessed in bivariate relationships with their associated criterion (i.e. the zero-order
correlation between, for example, academic self-efficacy and GPA), they had no

predictive ability.

First, the path from social/cultural involvement to social integration was removed.
Upon removing this path, the social/cultural involvement variable was no longer
hypothesized to relate to anything; thus, this variable was removed from the model
completely. The new model demonstrated a lower chi-square and CFI but a comparable
RMSEA (x2(1994)= 12804.60, p<.01; CFI=.80; RMSEA=.04) relative to the model that

included this variable and path (reported above). However, the AIC was lower for the

84

newer model (AIC=13370.60) compared to the older model (AIC=13832.98). These
results indicate that the structure of the social/cultural involvement variable was
positively influencing fit indices, even though the lone path coming from this variable
was not positively influencing fit. When the variable was removed, fit was affected by
the loss of the structure’s influence on the indices, but parsimony (as evidenced by the
AIC) increased. Despite the slightly lower fit indices it is not defensible to keep this
variable in the model. In theory, a host of variables with great structure could be
included in an SEM test but not hypothesized to be related to any other variable in the
model. This would improve fit but would tell the researcher nothing about the structural
relationships he or she is attempting to study. Because social/cultural involvement

contributed nothing to the model but a false impression of increased fit, it was removed

from firrther analyses.

Second, the path from conscientiousness to social integration was removed. As
might be expected, removing a single path in such a large model had only a small effect
on the chi-square (x2(1995)= 12806.30, p<.01,), and no effect on the CPI or RMSEA.
There was also a small but measurable effect on the AIC (AIC=13370.30). Because this
model was nested in the model that included this path, they can be compared using a chi-
square difference test. This test was nonsignificant (p<.05), indicating that the models fit
the data equally well. Because the more parsimonious model fits just as well, the path

was eliminated.

Though the path from academic self-efficacy to GPA was nonsignificant
statistically (p<.05), removing it was not as easily justified. If it were removed, academic

self-efficacy would be a “dead end” (i.e. it would lead to nothing else in the model).

85

GPA would become an exogenous variable with unknown causes and would have to be
set to covary with the other exogenous variables in the model. Even post-hoc, this was
difficult to justify theoretically. While empirically academic self-efficacy is indeed a
“dead end,” in accordance with good model modification technique, this path was

retained due to insufficient theoretical justification.

The final hypothesized path that was not affirmed by the data was the path from
shocks to withdrawal. It is quite possible that, at least in this college setting, there are not
many shocks that lead an individual to automatically quit. Instead, most shocks may
simply have a negative effect on satisfaction. It was deemed justifiable to remove this
direct path to withdrawal. As was the case with the conscientiousness-social integration
link, removing only one path did not have a large impact on model fit (x2(1996)=
12807.05, p<.01; CF I=.80; RMSEA=.04), though the AIC was slightly lower in this
model (AIC=13369.36) than in the previous model (AIC=13370.60). The chi-square
difference test, however, was nonsignificant (p<.05), indicating the path should be

deleted. This path was removed from the model.

Standardized regression coefficients for the paths in this modified early attrition
model are presented in Table 14. These modifications did change the pattern of
significance to some extent and the strength of some relationships. In the stage l-stage 2
links, performance avoid goal orientation was no longer related to academic integration,
but mastery goal orientation, knowledge, and perseverance were now significantly
(p<.05) related to academic integration. Of the newly added paths, only four of nine
ended up being significant (p<.05) in the full model test. Performance avoid goal

orientation, career orientation, perseverance, and emotional stability were significantly

86

e30

 

related to academic self-efficacy, while performance approach and mastery goal
orientation, knowledge, conscientiousness, and agreeableness were not related to
academic self-efficacy. As most of the changes occurred in this area of the model, the
rest of the regression weights were nearly unchanged. The squared multiple correlations
for this revised model are presented in Table 15. Academic integration was predicted
less well (R2=.21) than in the original model, but academic self-efficacy was predicted
better (R2=.28) than in the original model. No other multiple correlations were different

than in the original model.

Longitudinal Model Modification.

Though the early attrition model and longitudinal models were originally
conceptualized to be the same, it is possible that support for hypothesized mediating
effects is different. For this reason modifications were conducted separately for each
model. By examining the evidence for each model and building and trimming separately,
possible differences in these models could be uncovered.

The modification indices for the longitudinal model suggested even smaller gains
from modifying the model than those realized for the early attrition model. While
reassigning academic self-efficacy to the second stage was a supported change in the
early attrition model, it was clearly not appropriate in the longitudinal model. The largest
drop in chi-square (resulting from connecting performance avoid goal orientation to
academic self-efficacy) would only have been about 15. With no evidence supporting the

addition of paths, the focus shifted to removing paths.

87

In the longitudinal model, there were only two paths for which both the zero-order
correlation and beta were nonsignificant (p<.05): the path from social/cultural
involvement to social integration, and the path from conscientiousness to social
integration. The model that omitted the path from social/cultural involvement to social
integration, and also removed the social/cultural involvement variable altogether, resulted
in slightly worse fit (x2(2001)= 10308.11, p<.01; CFI=.82; RMSEA=.04) than the old
model (x2(2188)= 10786.96, CFI=.83, RMSEA=.04). The new model, however, resulted
in a lower AIC (10860.11) than the old model (AIC=11380.96). The chi-square
difference test for this modification was significant (p<.05), but this was due to the
positive impact of the measurement of social/cultural involvement on fit (i.e. leaving the
social/cultural involvement variable in the model, but with no relationships with other
variables, did improve the chi-square). For the same reasons discussed above, this
modification was retained. When the path from conscientiousness to social integration
was removed the CFI and RMSEA did not change, but the chi-square (x2(2003)=
10309.99, p<.01) and AIC (10859.99) changed slightly. The chi-square difference test

was nonsignificant (p<.05), indicating the path should be removed.

Finally, the correlation matrix indicated that the stage 3 variables of satisfaction,
commitment, and GPA were directly related to withdrawal. It was decided to move this
variable from the final stage of the model to the stage with withdrawal cognitions and
behaviors. Conceptually, this alters this portion of the model. Withdrawal cognitions
and behaviors can be considered as a criterion just as withdrawal can. It is likely that
many students who consider withdrawing do not follow through. However, these

students may not be as involved in the community, or may not represent the university

88

well to outsiders by expressing their intent to leave. Thus, it is still a criterion of interest.
Also, in the longitudinal model withdrawal was summed across years two through four.
The withdrawal cognitions and behaviors measured in students’ third semester may not
have captured consideration of withdrawing that occurred later. Low satisfaction,
commitment, and GPA may have accumulated over time, or simply may have not
changed despite a student’s short-term decision to stay in college; in this case, these
variables could eventually lead to a withdrawal decision that would not be mediated by

withdrawal cognitions and behaviors measured during the third semester.

The withdrawal variable was moved to stage 4 of the model. The path from
withdrawal cognitions and behaviors to withdrawal was removed, and paths were added
from satisfaction, commitment, and GPA to withdrawal. The CFI (.83) and RMSEA
(.04) remained the same, but the chi-square (x2(2001)= 10210.08, p<.01) and AIC
(10762.07) decreased. The betas for the paths from commitment and GPA to withdrawal
(-.32 and -.25) were significant (p<.05), but the beta for the path from satisfaction to
withdrawal (.04) was not significant Qr<.05). In addition, the variance explained in both
withdrawal cognitions and behaviors (18:34) and withdrawal (R2=.16) were lower than
in the original model. Because fit was not notably improved, and because the variance
explained in both withdrawal cognitions and behaviors and actual withdrawal decreased,

this modification was not retained.

The model identified prior to moving withdrawal was accepted as the final
modified longitudinal model. Standardized regression coefficients for the paths in this
model are presented in Table 16. Because there were minimal changes, the modified

model exhibited the same pattern of significance and very similar betas to those found in

89

the original model. Because the squared multiple correlations did not change, they are
not presented. A summary of model fit across the early attrition model, the longitudinal

model, and the revised version of both of these models is presented in Table 17.

Discussion

Main Findings

These findings provide mixed support for the model as a whole. The main
problem with the model was the measurement portion. Many of the variables did not
demonstrate discriminant validity. These problems caused estimation issues, and while
combining variables resolved these problems, fit was negatively impacted. In the full
model, the addition of structural paths did not compensate for this poor measurement.
The differences in fit indices are evidence of these measurement issues. Because many of
the indicators and latent variables in this model are indeed correlated, a low CF I resulted,
as this fit index compares the hypothesized model to a baseline model in which indicators
and latent variables are uncorrelated. Even though the model fared poorly when assessed
by the CPI, the model appeared to fit the data much better when assessed by the lack of

fit relative to degrees of freedom (as indicated by the RMSEA).

Despite these measurement issues, regression weights from the structural portion
of the model provide support for many of the hypotheses. Of the stage l-stage 2 paths,
five were supported by the early attrition model and the longitudinal model. Career
orientation, agreeableness, and emotional stability were related to academic integration,
and agreeableness and emotional stability were related to social integration. In the early
attrition model, performance approach goal orientation was related to academic

90

integration, and in the longitudinal model, knowledge was related to academic
integration. While only seven of 13 hypothesized paths were supported in the full model
by their associated regression weights, this number was reduced because of
multicollinearity. The zero-order correlations indicate that 11 of these 13 paths were
significant (the paths between social/cultural involvement and social integration and
between conscientiousness and social integration being the two exceptions). If these
variables were to be measured with more precision and discriminant validity was
increased (reducing multicollinearity), it is possible that the regression coefficients

associated with the paths would be significant.

These results increase our understanding of the relationships of these variables to
withdrawal. While the broad category of entry characteristics has been found to directly
affect withdrawal (Pascarella & Terenzini, 1980; Braxton, Sullivan, & Johnson, 1997), it
may be more useful to understand these relationships as being affected by a series of
intervening variables. More specifically, conscientiousness, agreeableness, emotional
stability, mastery goal orientation, career orientation, and perseverance have all been
directly related to withdrawal or withdrawal intent in previous research. The zero-order
correlations indicate that all of these variables were more strongly related to academic
integration than to withdrawal or withdrawal intent in both the early attrition and
longitudinal models, with one exception (i.e. emotional stability was correlated r=.15
with academic integration and r=-.15 with withdrawal cognitions and behaviors). In the
full models, the modification indices did not suggest that direct paths should be added
from these entry characteristics to withdrawal. Similarly, although conscientiousness and

social/cultural involvement did not demonstrate hypothesized relationships with social

91

integration, agreeableness and emotional stability demonstrated stronger relationships
with social integration than with withdrawal or withdrawal cognitions and behaviors,
despite previous research indicating direct relationships with these outcomes. These
results suggest support for mediation and a more process-oriented approach to the study

of relationships between individual differences and withdrawal.

The strongest findings involve the variables in the later stages of these models. In
the stage 2-stage 3 paths, zero-order correlations indicated relatively strong relationships
between the stage 2 variables of academic integration, social integration, and family
support and the stage 3 variables of commitment and satisfaction. Again, these effects
were attenuated in the full model for the relationships between academic integration and
satisfaction and between family support and satisfaction due to similarity of constructs in
measurement. Even so, the results concerning the relationships between integration and
both satisfaction and commitment are consistent with previous research (Braxton,
Sullivan & Johnson, 1997; Liu & Liu, 2000) supporting these relationships. The findings
regarding family support are consistent with previous research that has found a
relationship between family support and commitment (Cabrera, Nora, and Castaneda,
1993) but are incremental in that they also identify a relationship with satisfaction,

though this bivariate relationship was weakened when included in the full model.

Lastly for the second stage, the shocks variable was not an effective predictor of
satisfaction or dropout in the full model, but it did exhibit significant (p<.01) negative
bivariate correlations with satisfaction. It is important to note that the typical richness
found in a measure of shocks was decreased in this study. While a number of different

shocks were measured, the representation of shocks in the full model combined all shocks

92

into a single binary composite, simply indicating whether a student had experienced any
of the shocks or none of the shocks. Different shocks may have varied in the extent to
which they decreased satisfaction or resulted in withdrawal decisions. Future research
should investigate the impact of shocks in this context more thoroughly without such a

simple representation of this phenomenon.

One of the most significant differences between the early attrition model and the
longitudinal model was the relationship between academic self-efficacy and GPA, which
was nonsignificant (p<.05) in the former and moderate and significant (p<.01) in the
latter. Because self-efficacy was measured at different time points in these models, this
difference is consistent with previous literature that has found academic self-efficacy
only demonstrates predictive ability when it is measured later in college. It appears that
students do need to develop accurate perceptions of ability based upon feedback and

performance in a context before developing a sense of self-efficacy in that context.

Zero-order correlations between the stage 3 variables of GPA, satisfaction, and
commitment and the stage 4 variable of withdrawal cognitions and behaviors were
consistent and moderate to strong. While the relationship between satisfaction and
withdrawal cognitions and behaviors was somewhat smaller (but still significant, p<.05)
in the full model, the other two paths remained strong. This evidence provides support
for the idea that poor performance, reduced contentment with one’s situation in college,
and reduced dedication to the goal of graduating and to one’s institution increase the
likelihood a student will evaluate alternatives to college and consider dropping out. It is
interesting to note, then, that these withdrawal cognitions and behaviors lead to a decision

to actually withdraw much less frequently in the first year than in the second through

93

fourth years. It may take time for a student to fully evaluate his or her options, hence the
support from the longitudinal model. Also, students making such deliberations may
either decide to “stick it out” and see if college life improves, or may simply not
encounter the same pressures from family, friends, or the university to withdraw earlier in

their college experience.

Overall, this evidence provides a general picture of the factors affecting college
student withdrawal. Students appear to go through a process from the time they enter
college to the time they decide to leave a university, rather than simply withdrawing as a
result of background variables, despite previous research having demonstrated this
relationship. As was mentioned, previous research has found that entry characteristics do
affect withdrawal (Pascarella & Terenzini, 1980; Braxton, Sullivan, & Johnson, 1997),
but these associations may be better understood when considered in the context of a full
model especially when assessed over time. In this study, the entry characteristics were
largely unrelated to eventual withdrawal. The exception was the variable of
social/cultural involvement. Though the modification indices (and zero-order
correlations) for the early attrition model most strongly supported a direct path from
social/cultural involvement to withdrawal, it is unclear why this association existed.
Perhaps students who spent more of their time pursuing non-academic endeavors simply
concluded they enjoyed these activities more than college. Future research should

investigate possible reasons for this relationship.

This was not the only possible improvement to the model that was not ultimately
undertaken. In the early attrition model, though not retained, there were nonhypothesized

direct relationships from commitment and GPA to withdrawal. These variables could

94

perhaps be rearranged in a manner that takes advantage of such relationships while not
decreasing the variance explained in withdrawal and withdrawal cognitions and
behaviors. If these direct relationships are supported in future research, such findings
would imply that students who lack comrrritrnent or have a low GPA early on in college
may be more likely to withdraw without evaluating alternatives to the same extent as

students who withdraw later in college.

Theoretical Implications

These findings build upon the academic withdrawal literature in a couple ways.
First, the entry characteristics that were used in previous research were varied and
inconsistently employed. This study represents a more exhaustive set of entry
characteristics, and also explicates the nature of their relationships with academic and
social integration. These variables are related to withdrawal, but I show that this is a
mediated relationship. Second, this model expands upon more abbreviated models (e. g.
Cabrera, Nora, & Castaneda, 1992) that utilized variables from Tinto’s (1993) and Bean’s
(1981) frameworks. While more complex models that use variables from both
frameworks have been proposed on theoretical grounds (e. g. Bean and Eaton, 2000), little

empirical testing has been done on these complex models.

These findings also build upon the organizational turnover literature. First,
Munson and Rubenstein (1992) advanced the idea that school is work for the students
involved, and that these two contexts are more similar than most believe. Peterson

(2007) tested a model somewhat similar to the present model (i.e. combining some of the

95

same variables from the academic and organizational literatures) in work organizations,
finding that initial goals and career decision-making self-efficacy were related to
employee fit or integration, and that this integration predicted commitment and
satisfaction. Similarly, the present study demonstrated that career orientation and facets
of goal orientation are related to academic integration, and that this integration predicts
commitment and satisfaction. These shared relationships aid our understanding of
Peterson’s (2007) representation of organizational turnover and our general

understanding of the relationship between the academic and organizational contexts.

Second, satisfaction, commitment, and turnover intentions have been extensively
researched in organizational settings. This study provides evidence that these variables
are important in decisions to leave the institution one is associated with, and also are
important across contexts. Fit has been found to relate to commitment (Sekiguchi, 2004)
and satisfaction (Verquer, Beehr & Wagner, 2003) and commitment has been found to
relate to withdrawal cognitions (Cohen & F reund, 2004) and intent to leave (Williams &
Hazer, 1985). These findings were replicated in this study. In addition, the relationship
between satisfaction and turnover is better clarified by the results of this study.
Satisfaction has been found to have direct relationships with turnover (Griffeth, Horn &
Gaertner, 2000) and to mediate the relationship between fit and turnover (Wheeler,
Coleman, Gallagher, Brouer, and Sablynski (2007). The results of this study are more
consistent with findings that suggest the relationship between satisfaction and turnover is
not direct but is mediated by withdrawal cognitions and behaviors “(e.g. Hom, Caranika-

Walker, Prussia, & Griffeth, 1992).

96

Despite the similarities of these results to findings in the organizational literature,
there are some differences. The shocks variable was a poor predictor of satisfaction and
withdrawal. As was mentioned, this could have been a result of collapsing all of the
shocks into a single binary variable, losing much of the richness of the shocks measure.
Another possible explanation is that students do not experience as many shocks as
employees. Even if the rate of experiencing shocks is the same in the two contexts,
students may be less likely to drop out as a result of a shock than employees. Whatever
the reason, the shocks variable did not display patterns consistent with evidence from the
organizational literature. Also, the intent to leave, search behaviors, and thoughts of
leaving variables also could not be investigated in the order supported by Sager, Griffith,
and Hom (1998) due to multicollinearity among these variables. If these variables had
demonstrated discriminant validity, it is possible such an order could have been
supported, but the results of this study only support relationships of satisfaction,
commitment, GPA and withdrawal with a more general variable of withdrawal cognitions

and behaviors.

The similarity of these relationships in both contexts could be interpreted to
support a logical extension of Munson and Rubenstein’s (1992) hypothesis: this is
organizational research, and attempts to categorize it based solely on the contextual label
(i.e. school or work) under which it occurs rely on artificial assumptions. This is not to
say that the contexts are equivalent; as was discussed in the introduction to this study,
there are important differences between them. However, attrition occurs in both contexts,
and there appear to be similar cognitive processes for individuals who voluntarily leave

either context.

97

mi;

.)
(nt\

Practical Implications

These findings also have several practical implications. Measuring entry
characteristics such as career orientation, perseverance, and goal orientation may be
useful for academic institutions. These characteristics are largely developed before one
gets to college, but they affect the match one perceives between individual characteristics
and the university environment. These variables could play a part in the admissions
process. Selecting students based on these could increase the likelihood that a student
will perceive this fit, and eventually could lower withdrawal rates. Admittedly, there are
potential problems with this use such as applicant faking and reactions, but admissions

are still a potential use.

Alternatively, students could be measured on these variables when they enter
college for counseling purposes. Students likely to drop out could be counseled in order
to lessen the chances of this occurrence. The variables in later stages of the model could
be used similarly. The degree of integration a student perceives with the academic and
social aspects of college could be measured, and counselors could target these areas if
they are in need of improvement. Satisfaction, commitment, and withdrawal cognitions
and behaviors could also be monitored throughout college to provide assistance to these

students and reduce attrition.

98

Strengths and Limitations

This study has a number of strengths. First, this represents the first test of a
model in the collegiate context that combines variables from the dominant models of the
academic literature and variables from the organizational literature. While models have
been tested in this context using variables from Tinto’s and Bean’s frameworks and a
model combining academic and organizational variables has been tested in a workplace,
no previous study has tested a model in the academic setting that has drawn on the
academic withdrawal and organizational turnover literature. Second, this is a
longitudinal study assessing withdrawal over four years. Many other studies of
withdrawal (e. g Terenzini, Pascarella, Theophilides, & Lorang, 1983; Cabrera, Nora &
Castaneda, 1993; Kahn & Nauta, 2001) have not done this. Third, this study involves
participants in multiple institutions, so the results may be more generalizable than studies
that have only focused on participants in one institution (e.g. Mashburn, 2000/2001;

Braxton, Milem, & Sullivan, 2000; Edens, 2006).

Despite these advantages, there are some limitations associated with this study.
First, the correlational nature of the data prevents establishing causation. Even though
correlations support many of the associations between variables, it cannot be concluded
that one necessarily causes another. Also, as has been discussed, the measurement of
these variables was not optimal. Facets of a number of the variables were highly
correlated. For instance, the items in the goal orientation measure may not have been
optimal. For example, the third situation presented pertained to volunteering to solve a
problem in front of a class. Individuals may not want to do this for a number of reasons

including anxiety and low confidence, but these could have little to do with low mastery

99

orientation, which two of the items are intended to capture. The motivation for
performing a number of the actions in the scenarios presented may not be as simple and
clear-cut as the items appear to suggest, and this may have resulted in the strong
correlations between the facets. The collinearity of these facets may have affected the

paths from these facets to academic integration.

The measurement of satisfaction, commitment, and withdrawal was especially
problematic. Analyzing a model with such a large number of variables undoubtedly
added to multicollinearity issues. Given the multicollinearity among the entry
characteristics (e. g., perseverance correlated 2.40 with a majority of the other variables in
this stage), it might be useful to employ a smaller model with fewer of these variables.
This could prevent some of the significant correlations from displaying nonsignificant
betas in the full model, or even eliminate the suppression that occurred in the relationship
between mastery goal orientation and academic integration or the relationship between
perseverance and academic integration. In the later stages of the model the inability to
discriminate between variables clearly caused difficulties. This prevented testing the
model exactly as hypothesized, so questions remain about the relationships involving

facets of satisfaction, commitment, and withdrawal cognitions and behaviors.

In addition, although common in withdrawal studies, the attrition rate in this study
(which summed to 23.1%) was lower than the actual rate, which is consistently estimated
at approximately 40% (ACT, 2008). This statistic, however, is for public universities;
private universities typically have lower attrition rates than public universities. Two of
the universities in this study (the students of which comprised 14.8% of the sample) were

private universities and have typical attrition rates of 23% and 14% according to data

100

from those universities, so the attrition rate for the entire sample should not be expected
to be as high as the ACT estimation. Also, some students who had not yet graduated
may have dropped out after the completion of this study, though given the low attrition
rates in the 3rd and 4th years of college, this number is not likely to be high. In addition,
students who did not respond to the survey may be different than students who did
respond. For instance, poorer performing students may have been less likely to respond
to the survey, but they would also be more likely to drop out. However, without having
these data it is not possible to know. All of these factors could have contributed to the

lower attrition rate found in this study.

Future Work

The deficiencies in this study and the questions raised by the findings could be
illuminated with future research. Suggestions for future research have been offered in
this section, generally falling into the categories of measurement and operationalization
improvements, structural improvements, or research design and focus improvements. In
the measurement category, the scales in this study could be improved. Most of these
scales were developed specifically for use in this study. While this increased contextual
specificity, it may have led to too much overlap between the scales, causing the
aforementioned measurement issues. Future research should investigate if different
measures impact the relationships in this model. Additionally, even with retaining the
same shocks measure, a different representation in the full model could improve the
utility of this variable. A study with a narrower focus on these shocks could aid in

understanding them.

101

Structural improvements could also be investigated. While the results of this
study represented an attempt to include more variables than other studies have, there is
always the possibility that important variables were omitted. Other entry characteristics,
such as locus of control and coping strategies, could demonstrate relationships with fit or
integration. Variables such as study habits, hours worked at a non-academic job, or stress
levels could affect the withdrawal process at some point. More variables would make the
model more complex, but they could help better identify how students move from
enrollment to withdrawal. Aside from adding variables, those measured in this study
may not function as originally hypothesized. Model modification indicated relationships
between GPA, commitment and satisfaction with withdrawal in the longitudinal model.
Reconfiguration of the variables in the model could account for these relationships
without losing the important link between withdrawal cognitions and behaviors and
withdrawal. Family support was also significantly (p<.05) correlated with academic and
social integration. Just as an argument was made for the hypothesized relationships
between family support and the outcomes of satisfaction and commitment, an argument
can be made for why such support facilitates integration into the academic and social
environment. This variable could be measured at the initial data collection and entered
into the model as a stage 1 variable that leads to integration. Such reconfigurations of

variables in the model would aid understanding of the nature of these variables.

Lastly, the research design or focus could be altered. This method did not focus
on institutional characteristics such as whether a university was public or private,
selectivity, or regional or cultural differences. It is possible the withdrawal process

differs based on these characteristics. Also, student characteristics, such as race, gender,

102

or whether they were traditional or continuing education students were not assessed.

Again, it is possible focusing on one group or another would affect results.

Conclusion

Assessing withdrawal with a more complex model aids our understanding of this
phenomenon. It seems that entry characteristics do affect integration, and that integration
and family support affect commitment and satisfaction, which lead to withdrawal
cognitions and behaviors and eventually to withdrawal. The large number of variables in
the model and the measurement issues involving those variables complicated
interpretation of the results, but many of the hypothesized relationships were supported.
These results have helped bridge the student withdrawal and organizational turnover
literatures, and they also have practical implications for universities looking to reduce
withdrawal. With further research into this model and these relationships, we can
increase our understanding of student withdrawal, and perhaps eventually affect the rates

that have been consistent over so many years.

103

APPENDICES

104

Appendix A

Biographical data- Career Scale

1. How many times have you gathered information (e. g., from the library, on the
Internet) about a career in which you were interested?

A. None

B. One or two

C. Three or four

D. Between five and ten
B. More than ten

2. Which of the following best reflects your current career plans?

A. You know what you want many years in advance and plan to stick to you goals
B. You have a general idea of what you would like to do

C. You know what you want to do a few years in advance,
but are not concerned much beyond five years or so
D. You are really only concerned with your immediate goals

B. You make no plans at all, but take advantage of opportunities
as they present themselves

3. In the last year, how many times have you talked to a career counselor or used
materials at a career center?

A. None

B. One or two

C. Three or four

D. Between five and ten
B. More than ten

4. What steps have you taken to gather information about possible careers? Pick
the answer that best demonstrates the amount of effort that you have made.

A. I have not done anything actively yet

B. I have mostly listened to friends and observed a parent's or
relative's career

C. I have sought information from my school or career counselor

D. I have gotten library or web-based information on careers that I thought were
interesting

5. How committed are you to achieving your career goals?

105

A. Extremely committed
B. Very committed

C. Somewhat committed
D. Not very committed
E. Not at all committed

6. To what extent have you tried to prepare yourself for a particular job that you
hope to have in the future? (For example, you contacted a company and talked
to someone about the work.)

A.A great deal

B. A lot

C. Somewhat

D. A little

E. Not very much

7. How confident are you about what your career will be?

A. Very confident

B. Somewhat confident

C. About as confident as others are

D. Somewhat less confident than others
E. Not confident

8. How many awards or scholarships have you applied for that were directly
relevant to your career interests?

A.0

B. 1

C. 2 or 3
D.4or5

B. More thanS

9. How confident are you about what your college major will be?

A. Very confident

B. Somewhat confident

C. About as confident as others are

D. Somewhat less confident than others
E. Not confident

10. When did you develop a sense of what you want out of a career (whether

it is money, respect, interesting work, the opportunity to help people, or something
else)?

106

A. Before high school

B. During my first year or second year in high school
C. During my third year in high school

D. During this past year

E. Still undecided

107

Appendix B
Biographical data- Perseverance scale
1. How important is it to you to succeed in whatever task you are engaged in?

A. Extremely important
B. Very important

C. Important

D. Not very important
E. Not at all important

2. When encountering problems that take a long time to solve, how impatient do
you tend to become?

A. Extremely impatient

B. Very impatient

C. Somewhat impatient

D. Slightly impatient

E. Not at all impatient

3. How often do others tend to compliment you on your determination to continue
with a project under difficult circumstances?

A. Very often
B. Often

C. Sometimes
D. Rarely

E. Never

4. To what extent would your friends describe you as someone who goes after
what you want?

A. Not at all

B. A slight extent

C. A moderate extent

D. A large extent

E. A great extent

5. How ofien have you achieved a personal goal that seemed unattainable at first?

A. Very often
B. Often

C. Sometimes
D. Rarely

E. Never

108

6. How frequently do you fail to get what you want because you did not put in enough
effort?

A. Very often

B. Often

C. Sometimes

D. Rarely

E. Never

7. To what extent has it been important to you to do your very best whenever you
take on a project?

A. Extremely important
B. Very important

C. Important

D. Not very important
E. Not at all important

8. How often have you accomplished something you initially thought was very
difficult or almost impossible?

A. Very often
B. Often

C. Sometimes
D. Rarely

E. Never

9. How often have you finished a project when faced with difficult
circumstances?

A. Very often
B. Often

C. Sometimes
D. Rarely

E. Never

10. How often do you tend to give up on a task after being told that you were
not doing well?

A. Almost all the time
B. Most of the time
C. Sometimes

D. Rarely

E. Never

109

Appendix C
Biographical data— Knowledge scale

1. Think about the last several times you have had to learn new facts or concepts about
something. How much did you tend to learn?

a. usually not enough

b. sometimes not enough

c. just what is needed

(1. a little more than what is needed
e. much more than what is needed

2. How do you compare your standards for learning to the standards teachers in high
school or college gave you?

a. much lower

b. lower

c. about the same
(1. higher

e. much higher

3. How do you compare with other people your age in having specific knowledge on a
wide variety of topics (both inside and outside of school)? .

a. well below average
b. below average

c. average

d. above average

e. well above average

4. Think about those courses in high school you were most interested in. Generally how
determined were you to learn the facts and concepts from the class?

a. extremely determined
b. very determined

c. rather determined

d. sort of determined

e. not very determined

5. When you were in high school, how much of a priority was knowledge and learning
(both inside and outside of school)?

a. not much of a priority
b. somewhat of a priority
c. a priority

(I. a high priority

“0

e. a very high priority

6. Out of the courses that interested you in high school/college, how often were you
interested in knowing all the information (vs. getting a general idea of the infonnatron)?

a. never

b. not very often

c. sometimes

(I. often

e. always

7. Generally, whenever you learn about a topic or how to perform a task, how often do
you learn all the details as well as the general principles?

a. hardly ever
b. not very often
c. sometimes

(1. often
e. almost always

8. In your college/high school courses, how effective would you say you are at learning
knowledge and mastering general concepts?

a. very effective

b. effective

0. not effective but not ineffective
d. ineffective

e. very ineffective

9. In college, how do you tend to work on your homework assignments?

a. You always spend the most time learning what you think the teacher/professor will
include on tests

b. you learn enough to get the work done, and eventually learn the concepts

c. you make sure you learn the concepts, and get the work done

d. you learn the concepts, and after the concepts are learned, you practice some more

10. How often do you tend to learn all the rules of a complicated game, or all the details
of a complicated task, before trying it?

a. almost always
b. most of the time
c. sometimes

d. not often

e. almost never

1]]

Appendix D
Biographical data- Social/cultural involvement scale

1. Over the past year how many art exhibitions have you attended?

A. None

B. One

C. Two

D. Three or four
E. Five or more

2. How many times in the last year have you attended cultural events, even when you
weren't certain about whether you would like them?

A. Never

B. Once

C. Twice

D. Three or four times
E. Five times or more

3. The number of high school clubs and organized activities (such as band, sports,
newspapers, etc.) in which you took a leadership role was:

A. I did not take a leadership role

B. 1
C. 2
D. 3
E. 4 or more

4. How many times in the past year have you volunteered in social service or charity
organizations?

A. Never

B. Once

C. Twice

D. Three

E. Four times or more

5. In the last six months, how many times have you tried to talk to someone from a
different country or culture just to learn about their background?

A. Never

112

B. Once

C. Twice

D. Three or four times
E. Five times or more

6. In the past year, how many hours were you engaged in community service or volunteer
activities?

A. None

B. Less than 10 hours
C. 11-40 hours

D. 41-80 hours

E. More than 80 hours

7. How important has it been in the past for you to be involved in community or
volunteer work?

A. Extremely important
B. Very important

C. Important

D. Not very important
E. Not at all important

8. In the last year, how many times did you go to a play, musical, or other live theater
performance?

A. Never
B. Once
C. Twice

D. 3 to 5 times
E. More than 5 times

9. How many times each year do you visit museums, art galleries, or exhibitions?

113

10. Compared with others your age, how much do you know about art (e.g., types of
painting, sculpture, and music) both historically and across cultures?

A. Much more than others

B. Somewhat more than others
C. About the same as others
D. Somewhat less than others
E. Much less than others

114

Appendix E
Goal Orientation Measure

Instructions: Below are descriptions of diflerent situations. Each situation has a
number of options following it that describe different reactions. Please indicate how
likely you would be to react in each of the ways described, using the scale below.

 

l 2 3 4
definitely not probably not probably definitely
react this way react this way react this way react this way

 

Think of each item by itself In other words, just think about the situation and how you
would react to the specific item you are answering at the time. Don ’t worry about how
you said you would react in the previous items.

A teacher says something that you don’t understand in class.

1. I would raise my hand to gain a better understanding. (MASTERY)

2. I would raise my hand if I get credit for class participation. (APPROACH)

3. I would raise my hand to make sure that what she is saying is similar to what I am
thinking. (MASTERY)

4. I wouldn’t raise my hand because I wouldn’t want to look stupid. (AVOID)

You are having trouble finishing a statistics project that the teacher said should be pretty
easy.

5. I would go to the TA. for help because I want to get the best grade I can.
(APPROACH)

6. I would go to the TA. for help, so the TA. would recognize that I am trying and
showing my effort to understand. (APPROACH)

7. I wouldn’t go to the T.A., because I’m embarrassed that I do not understand when the
teacher said it would be easy. (AVOID)

Your teacher asks for a volunteer to solve a problem on the blackboard in front of the
class. You think you know how to solve the problem, but you are not sure.

8. I would volunteer to solve it to see if I am correct in knowing how to solve it.
(MASTERY)

9. I would volunteer to solve it to see if I am wrong and then learn how to solve it.
(MASTERY)

10. I would volunteer to solve it to show class participation. (APPROACH)

115

11. I wouldn’t volunteer to solve it, because I fear the teacher’s criticism if I don’t do it
correctly. (AVOID)

Please continue to use the following scale:

 

1 2 3 4
definitely not probably not probably definitely
react this way react this way react this way react this way

 

In a class discussion, your teacher asks what the class thought of an assigned reading that
you read thoroughly.

12. I would raise my hand and share my opinion so the teacher knows I did the
assignment. (APPROACH)

13. I would raise my hand and share my opinion, so that I can make sure that what I got
out of the reading was correct. (MASTERY)

14. I wouldn’t say anything, because I have a fear of speaking in public and sounding
unintelligent. (AVOID)

You are studying with some friends, and you don’t understand some of the materials they
are reviewing. You feel like they are going too fast.

15.1 would ask them to slow down and explain it so I have a good understanding of the
material. (MASTERY)

16. I would ask them to slow down and explain it, because I want to do well in the class.
(APPROACH)

17. I would decide to learn it on my own later, because I don’t want the others to think
that I’m not intelligent. (AVOID) '

You’re in a classroom and your teacher asks if anyone knows the answer to the problem
that she says is really difficult. You know the answer.

18. I would raise my hand and give the answer to show myself that I can tackle difficult
topics. (MASTERY)

19. I would raise my hand and give the answer if I get class participation points.
(APPROACH)

20. I wouldn’t raise my hand because I don’t want to be wrong or seem stupid in front of
other people. (AVOID)

116

Please continue to use the following scale:

 

1 2 3 4
definitely not probably not probably definitely
react this way react this way react this way react this way

 

There is an academic honor society for individuals in the major that you have. You have
been invited to join.

21. I would join the honor society because it looks good on a resume. (APPROACH)

22. I wouldn’t join the honor society because I don’t want to risk any potential
embarrassment. (AVOID)

23. I wouldn’t join the honor society because I fear rejection from others already
involved once I get there. (AVOID)

You are asked to be a T.A. for a class that you took last semester. You did well in the
class but aren’t sure you know the material perfectly.

24. I would agree to be the T.A. to advance my knowledge in that area. (MASTERY)
25. I would agree to be the T.A. to help build my resume. (APPROACH)
26. I would decide not to be the T.A. because l don’t want to look stupid. (AVOID)

117

Appendix F
IPIP Personality Items

Instructions: 0n the following pages, there are phrases describing people '3 behaviors.
Please use the rating scale below to describe how accurately each statement describes
you Describe yourself as you generally are now, not as you wish to be in the fitture.
Describe yourself as you honestly see yourself in relation to other people you know of the
same sex as you are and roughly your same age. So that you can describe yourself in an
honest manner, your responses will be kept in absolute confidence. Please read each
statement carefully and then fill in the bubble that corresponds to the number on the
scale.

 

l 2 3 4 5
very moderately neither accurate moderately very
inaccurate inaccurate nor inaccurate accurate accurate

 

1. Make people feel at ease (A+)

2. Have a rich vocabulary (0+)

3. Don't talk a lot (E-)

4. Have difficulty understanding abstract ideas (O-)
5. Am interested in people (A+)

6. Feel comfortable around people (E+)

7. Follow a schedule (C+)

8. Insult people (A-)

9. Get chores done right away (C+)

10. Make a mess of things (C-)

11. Sympathize with others' feelings (A+)

12. Don't mind being the center of attention (E+)
13. Keep in the background (E-)

14. Leave my belongings around (C-)

15. Feel little concern for others (A-)

16. Change my mood a lot (ES-)

17. Often forget to put things back in their proper place (C-)
18. Am full of ideas (0+)

19. F eel others' emotions (A+)

20. Have a soft heart (A+)

21. Pay attention to details (C+)

22. Shirk (i.e. skip out on) my duties (C-)

23. Am not interested in other people's problems (A-)
24. Am the life of the party (E+)

118

25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
. Have little to say (E-)
45.
46.
47.
48.
49.
50.

Am always prepared (C+)

Get irritated easily (ES-)

Have excellent ideas (O+)

Use difficult words (O+)

Get stressed out easily (ES-)

Start conversations (E+)

Get upset easily (ES-)

Do not have a good imagination (O-)

Am relaxed most of the time (ES+)

Often feel blue (ES-)

Talk to a lot of different people at parties (E+)
Have frequent mood swings (ES-)

Take time out for others (A+)

Spend time reflecting on things (0+)
Have a vivid imagination (O+)

Am not interested in abstract ideas (O-)
Don't like to draw attention to myself (E-)
Like order (C+)

Seldom feel blue (ES+)

Worry about things (ES-)

Am exacting in my work (C+)

Am quick to understand things (0+)
Am easily disturbed (ES-)

Am quiet around strangers (E-)

Am not really interested in others (A-)

119

Appendix G
F it/Integration Items

Please indicate the extent to which you agree with the following statements about your
school.

 

 

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

|
Academic Fit
1. The courses available at this university match my interests
2. I know other students here whose academic interests match my own
3. My current courses are not really what I would like to be doing
4. All things considered, my current major suits me
5. I feel like my current major is not the right academic program for me
6. I feel that my academic goals and needs are met by the faculty at this school
7. I'm sure there must be another major for which I am better suited
8. I am able to use my talents, skills, and competencies in my current courses

Social Fit

9. The social activities on campus suit me

10. I have a lot in common with the students around me

11. My ideas of fun are not shared by the students around me
12. There are students here who really understand me

Financial Fit

13. I am in a similar financial position as most students here
14. Sometimes my college friends and I can't do things together because of money
15. The things I can afford are different from what my college friends can afford

120

Appendix H
Family support scale

Please indicate the extent to which you agree with the following statements about your
school.

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

I

 

 

My family supports my decision to attend this school

My family supports my decision to get a college degree

My family supports me emotionally if I'm having a hard time in school

My family supports me financially if I'm having a hard time in school

My family supported me throughout the college application process

My family wanted me to attend this school more than any other school

If I left this school to go to another, I feel like I would be letting my family down
Going to this school is part of a family tradition

?°>'?‘E"PP°!":"

121

Appendix I
Commitment scales

Please indicate the extent to which you agree with the following statements.

 

 

 

 

 

 

 

 

 

 

l 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

I

 

 

Goal (degree) commitment

. It is hard to take the goal of graduating from college seriously

It is unrealistic to expect that I will graduate from college

. I might rethink my goal of graduating from college, if things go differently than I

expect

. Quite frankly, I don't care if I graduate from college or not

I am strongly committed to pursuing the goal of graduating from college

It would not take much for me to abandon the goal of graduating from college

I think that the goal of graduating from college is a good goal to shoot for

I am willing to put forth a great deal of effort, beyond what I would normally do, to
graduate from college

9. I will not gain much by trying to achieve the goal of graduating from college

KAN—i

wsewe

Institutional Commitment

. I would be very happy to graduate from THIS particular school

. I enjoy discussing my school with people outside it

I really feel as if this school's problems are also my own

I think that I could easily become as attached to another school as I am to this one

. I do not feel "emotionally attached" to this school

. This school has a great deal of personal meaning for me

. I do not feel a strong sense of belonging to my school

. I am afraid of what might happen if I dropped out of this school without being
accepted somewhere else first (OMITTED)

9. It would be very hard for me to leave this school right now, even if I wanted to

10. Too much in my life would be disrupted if I decided I wanted to leave this school

right now

11. I would not lose many credits or money if I left this school now

12. I feel that I have too few options to consider leaving this school (OMITTED)

13. I do not believe that a person must always be loyal to his or her school

14. I believe in the value of remaining loyal to one school

15. I think alumni should remain actively involved in their school's activities

16. I believe allunni should continue to contribute financially to their school

“\IONUIPEJJNF‘

122

Appendix J

Satisfaction Items

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

I

 

 

Academic Satisfaction

1. I'm satisfied with the intelligence of my teachers here

2. I'm satisfied with the extent to which my education will be useful for getting
future employment

3. I'm happy with the amount I learn in my classes

4. I am satisfied with the extent to which attending this school will have a positive
effect on my future career

5. I generally enjoy my coursework here

6. All in all, I'm satisfied with the education I get at this school

Social Satisfaction
7. I'm satisfied with the number of social activities I have had since attending this
school

8. I'm satisfied with the number of friends I have here

9. The friendships I have developed with other students at this school have been
personally satisfying

10. I'm satisfied with the extent to which I've made friends I can talk to about my
problems

1 1. It is satisfying to spend time with other students at this school

12. Overall, I'm satisfied with my social life at this school

Financial Satisfaction
13. I have to hold a job outside of school just to make ends meet
14. I often feel upset about my financial situation
15. I am dissatisfied with the amount of time I have to spend working at a job outside
of school
16. I worry about having enough money to pay my bills
17. I'm satisfied with the amount of college debt I am accumulating
18. In general, I'm satisfied with my financial situation

123

Appendix K
Academic Self-Efficacy Items

Using the following response scale, please indicate how accurately these statements
reflect your perceptions.

 

1 2 3 4 5
not true not true somewhat true true very true
at all

 

l. 1 am confident in my ability to succeed as a college student
2. I believe I can achieve good grades in college

3. I worry that I won’t be successful in college

4. I have the ability to excel in school

124

Appendix L
Thoughts of leaving

Please indicate the extent to which you agree with the following statements about your
school.

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

I

 

 

l. I am considering transferring to another school.
2. I am considering other job options instead of continuing in school.

125

Appendix M

Intentions to withdraw

Please indicate the extent to which you agree with the following statements about your
school.

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

I

 

 

 

1. I intend to be enrolled at this school 6 months from today
2. I intend to transfer to a different school at or before the end of this academic year
3. I intend to leave school and get a job at or before the end of this academic year

 

126

Appendix N
Search Behaviors

Please indicate the extent to which you agree with the following statements about your
school

 

1 2 3 4 5
Strongly Disagree Neither agree Agree Strongly Agree
Disagree nor disagree

 

 

 

 

 

 

 

 

 

 

 

 

 

l. I am gathering lots of information about other schools I could transfer to.

2. I am gathering lots of information about job options as opposed to continuing in
school.

127

Appendix 0

Shock scale items

Below you will see a list of events that could happen to a college student. For each event,
indicate whether this has happened to you and/or to a fiiend/significant other during
college (you may check both). If the event has not happened to you or a friend/significant
other, cheek "neither of the above."

_ This happened to me.

_ This happened to a friend or significant other.

_ Neither of the above.
1. Was the victim of theft
2. Was the victim of assault
3. Became pregnant
4. Was recruited by another job or institution
5. Receive an unexpectedly bad grade
6. Had roommate conflicts
7. Lost financial aid
8. Became ill
9. Death or illness of a family member
10. Became clinically depressed
11. A close friend or significant other left school
12. Became addicted to a substance
13. Conflict with a faculty member
14. Came into a large sum of money
15. Family member lost job, family in need of financial help
16. Lost job that was needed to pay tuition
17. Large increase in tuition/living costs
18. Experienced a significant injury
19. Became engaged or married, or entered a civil union
20. Received a job offer
21. Was unable to enter intended major at school

128

Appendix P

Table 1. Demographics for Entry, Semester 1, Semester 2, Semester 3, and

 

 

GPA/Graduation Samples.

Entry Sample Semester 1 mg

N % N %‘ N %
Gender
Male 945 34.8 391 34.0 340 34.1
Female 1692 62.3 758 66.0 654 65.9
Race/Ethnicity
White 1460 53.8 738 64.2 650 65.5
Black 652 24.0 123 10.9 95 9.5
Asian 201 7.4 130 11.3 110 11.0
Hispanic 152 5.7 63 5.5 49 4.9
Other 156 5.7 90 8.0 90 9.0
Age at entry
18 2336 86.0 1008 87.7 861 86.4
19 247 9.1 118 10.3 115 11.5
2 20 49 1.8 23 2.0 21 2.1

 

129

Table 1 Continued.

 

Gender

Male

Female
Race/Ethnicity
White

Black

Asian
Hispanic
Other

Age at entry
18

19

220

W
N %
301 34.9
562 65.1
570 66.0
67 7.8
1 1 1 12.9
43 5.0
72 8.3
745 86.3
88 10.2
16 1.7

 

2008 GPA Sample
N %
632 33.3

1265 66.7

1211 64.0
346 18.3
161 8.5

67 3.5
107 5.7
0 0
0 0

1894 99.8

 

130

Table 2 Timeline of the Variables Collected.

 

Wave 1 Wave 2 Wave 3 Wave 4

(beginning (end of (end of (end of

of college) first second third
semester) semester) semester)

 

Biodata (Knowledge, X

Perseverance, Careers,

Social/Cultural

Involvement)

Goal Orientation X

Personality X

Academic Self-Efficacy X X X
GPA HS College College College
Commitment X X X
Satisfaction X X X
Integration X X X
Family Support X X X
Withdrawal Cognitions X X X
and Behaviors

Shocks X X X

 

131

armm.

oo

aamfl.
**—_t

aaom.

**h—.

an

m_

nawv.

awe.

«*fim.
anOt
**@~.

aavfi.
aaam.

9%.

~—

mo.

anom.

*ho.
mo.
mot

*oor

mo.

mm.

o—

**c_{ *awo.
No. ~o.
*ammr aawm.
Nor mo.
we. we.
ace. vet
aao~r **_~.
aaoor **oo.
aawfir **¢—.
vw. aafiwt
aw.
@ w

rho.
aawfir

saoM.
co.

«not
*Nfi.

*o_.

**_N.
aahmr
*ahc.
mm.

aamfi.
name.

neon.
we.
we.

_or
**——.

**h_.

aawm.
*awmr
aahm.
**_v.
hm.

mo.
rahwr

*ahN.
mo.
oo.

mot
aaoc.

aaoM.

anoN.
aammr
*amm.
*amm.
**_¢.
QR.

eav—.
*aaN.

*amm.
vet
Nor

nor
cc.

**©_.

**0N.
**ONr
aemm.
anom.
*aav.
*amm.
no.

ham“.
_or

scum.
aa-r

camfi.

*awo.
aamN.

aamfi.

ave.
ramNr
ace.
samfi.
**a_.
same.
4*m_.
NQ.

**m~.
awe.

**hm.

oo.

*no.

**o~.

vo.

*umfi.

*aoo.
aavmr
*aoN.
sham.
*aav.
*ahm.
aaflv.
ramfi.

QQ

N

**©_.
**wo.

saw“.
_0.
mo.

_o.
aahfi.

aamfi.

aa—m.
**hmr
**wN.
*awm.
aa—m.
**o_.
*aom.
**V—.
**mm.
0%.

Nm.m_4V
Na. cod

©©.o_4‘
hv.wcgv
Vm.mhAH

Nb.wadn
vn.voau
mm._Num
mh.ohdu
mm.one_
vv.a_an
av.wo&n
av.mnau
mo.m0fi
5v.mfim
Nh.wman
_©.oWM
hm.oo4y

Aum.:5e4

SE

cowofimgm £80334 .fi
Aeaomaamv $6 .2
FE

cacao—bum bum 280334 .3
SE 3.8% .2

as: eoaaam been; .3
SE

“8:835 Boga .m _
8.3V £58on 38m .2

$6
E5832: outage/w .:

3080335
33538.6. .2
0o 22% Em .o

00 on? use .w

oo been: .5
oo§co>omcom .c
Maopzxu.m
omen—390m .v
35% 3:285 .m
mmocmsomucomomconv .N
amuse—smoocw< ._

 

.mcoufloboeoE— 23m $82 coring. beam m 033.

132

 

meg?

armmr

**~vr

*ammr

**_N.

**om.

**N—.

meg?

**Nmr

**_vr

**mmr

haem.

*amm.

**m_.

aaom.

*hor

rho:

VOA?

**N_.

tho.

_OA¥

*aogr

moxv

_ogv

vogfi

**h_f

meg?

gear

NOAV

Noam

mogv

**m_.

**m_.

meg?

ace.

Nogv

_OAV

*noo.

**—_.

**m_.

*hor

*ho.

hoot

acct

:wo

**wN.

**__.

Nogv

NOAV

Noam

noo.

**©~.

**o~.

mog¥

woo.

voafi

**__r

enact

**_N.

voav

*oo.

Noam

*amflr

**o_r

*hor

**N_.

mogv

**w_.

aver

vogfi

*aoflr

mogw

*aom.

voxv

anao.

avo.

*hor

*awor

«cot

**_N.

**N_.

mogfi

mmgv

:Yo

_wAv

vngv

vvav

ngv

owgv

N_Av

hm:—

on;

no;

hmac

hug”

mfim

2 not
“sonpxu.©m
as: good
9 685 .3

SE
wages ho
flaw—85. .vm
$6 o3
wctofimmv
.mm

AS:
308:8800
Boo .NN
56
acogmfifioo
353335
._N

66
cocomfiumm

Began .8

 

N—

__

o_

Aum caysg

 

.cosfiaou m 2an

133

.S. w Are... .3. w a... H802

 

 

wig-8.- 8.- 3. a: _. 8.- S. :2. :2. 8. .8: 3.- .8.- 2 not 88er em
bu . nine. 1.50. aamm... *anr 1.3.- ......Nm.r Siam-t “SKA.- _.....m_... amo. ..:..@~.. .31; .t ANBV 023..— 3 38% .va
N3
Amt“; .I.Nw. 3.9V... ..:..mm... nan—N6 ..:..wm.- *awmt a... —N.. 3.2.- *wo. .3..me 1.2... 9:23..— mo 3.3393- .VN
MV.H.~ aahmr 1.3!. a... 8.- .1. fl m.. 1.94.- ..:..mm.- two—.- amo. ..:..om.- aim—.- ANBV 85 mam-855W .mw
N3
mm. SEN. 1.2. _.:..©_. them. aamfi. *avm. fio. 3.2. aaofi. Even—MEEOU ECU .NN
56 doggone
Fe. a. 1.3». 1.3. 3. #0. wet .Ibm. mo. 3:35:me AN
3%
mm. ..&o. 1.2. **w_. it. 1.2.- «av-n... *aom. :omuoflmfimm Rwanda—mm .ON
65
2m. ..:..Nm. No.. 1.2. ELL..- aamfi awe. comaommmumm Rmoom .o~
326
as. ..:..N_. «*3. we... .I. .N. 3.3. comuommmflmm omEovmo< .m—
E 3.- as.- S. :8. canoe-roe $6 .2
£26 Succum—
Nm. mot 1.2. Ibo. tom 380934 .0—
...2 :2; .8.- §5 38% .3
me. :2. SE condom been .3
$5
ma. dozens..— aoeee... .2
on mm vm mm mm 3 ON a a t 2 2 E 2 .

 

62.5.80 m 2.3

134

SE 8:335

 

:8. :2. :2. :3. :2. :2. :2. :2. ”no Em 28082.:
2880202:
:2. :3. :3. :8. :3. :3. :8. :8 m3 and afiao%_oom.o_
:3.- :$.- :3. :8- :8.- ..MN; :§.. :5- N2 32 causififld
as :S. :R. :3 :2. .8. :8. :3. 3d 3m 00.23.20...”
ts :3. :2. :8. :2. :2. :3. 2.... m2 00239.2.“
2 6 : :.. :3. :2. :3. : _ m. 23 m? 85.53%; .0
a s :2. :8. :R. :2. 8... 2mm 2830 .m
as :2. :3. :8. :3 2.... omBEoé .q
23 :2. :2. N2. mam 2:93 3:285 .m
a s :2. Ed o2 mmoamsoufiséoo .N
as $5 8... 3803823 ._

w h o m e m N 2 am :82

 

.mcowflotoouofim 23m 3608 .mfiuamwcog .v 035-

135

56

 

20825800

......2. *...N~. ._:..mm. 1.5—. SKA. cod .33. *LN. 36 5%? FCC .NN

56

25.5.8800

two. 1.3. 23 3; g. 86 mod 86 Lb. de Sim 3.83355 .3

33v 5:89QO

mod. mod- cod mod- wed .28. mod 85. wad vma 305:5 .om

65 couoagm

35o. 1.3. 2.8. 1.3. mod- 1.2. mod 1.2. end :6 38m .3

SE 8.5...QO

86 mod 1L fl. .._:..o_. 1.3. 1.3. *8. 1:. Ed nod £8033. .M:

3:82:56

mod- 3.8.- Lo. 1N7 3mm. 56. *wo. mod mud find «50 .5

2%

385m row

cod IE. 3.2. 3.3. .33. 3.2. .26 1.8. $6 cad omfiocmoax .3

mod mod vod mod vod- 1. _ Hr cod 56 two 3.0 33v 302% .m g

SE 225%

3d cod mod cod med- 1.2. *8. mod de Sum bmfimm .3

33v 5389::

wed- 1b.- Ed- mod- Nod- Imo. 2.2. 56 mud wad 328:3 .2

3.5 8283.5

1:. IE. 1.2. 1.8. cod 3mm. vod 3:. Ed we...” 38m .S
m N. o m w m N _ Om :82

 

3:528 v 2.3

136

 

de

wed

$4

and

:6

am;

9}“

mm;

a.

-N $3.5 Soap—Q 6N
$2:

263 8 ESE .mm
as: was»:

.8 mawsofi .vm
as: as

@558 .mm

 

Gm

532

 

3:588 v 22¢

I37

55 :osofimumm

 

:mm. :8; :8. :2. :mm. :2. :5. 36. Ed Se 28082.2
322,68
:3. :8: 3o :2. 8.? :s. :2. :8. m3 :8 $5.:
3.2:
hoaoamaom
N: 8.? :2. :2. :3. :3. 8d- :2: So am 28233.2
Sc :2; :3- ::.- :8: 8.0 85. :3 3d $350292
Amavcaasm
was :2. :8. :2. 8.? :3 Ed m: 559:
SE Sages
as :2. :3. :8: :8. NS ”2 3235.2
Q26 8:835
as :3. mod :2; Ed $40. 308.2
ANBV aozfiwofiu
as Ed :8; ”2 Rd 28032.:
«Goao>_o>:~
Beauioom
Es :3- m3 m2 .2
00
33 mg C: EOE Em .o
2 2 E S Q : S 0 am :32

 

3:528 v 033

I38

 

 

mod- *ao. 1.0—.- ..hor ..:..N_.- *wcr *mor Sd- de _ —.o CNN mbwo>v
528.5 .3.
t..:.- cod firefi- .I.o~.- **©N.- .IBNV mod- :2. mod am; $26 023..—
3 335 .mN
**ON.- *Bo. fwawf ._....m—.- fromr twvmr Nod- *ho. ~04 and Oak/v war/mo..—
.«o 3.38.: «N
3.2.- ood _.:..m~.- *aor *...mN.- .3..me mod- *wo. mud am; $26 £5
3558 .8
35%. mod- 1:. EL —. :2. .IbN. 35¢. *._.N~.- 3.0 5%? Am}:
HGQEMEEOU
BOO .NN
*wo. mod- 1mm. wed 3.3% ivm. 1.5:. Ed- de 3mm Ami/v
6258880
ficomuaumfi AN
nihm. Infi- 1mm. .I.Nm. 1.2. 3.2. cod Mod «6.0 vmd 33v comuommmflwm
13053; .ON
.I.w—. 3.2.- 3.3. 3.2. .150. 3.7:: vod ‘25—.- cwd tum $.26 :omuoflmumm
330m d~

2 2 3 m— S _ — E a Gm 53:

 

62.5.80 v 2.3

139

.S. w at. .8. w a... 582

 

a.

V2 **mm. *...mm. .ILN. ood- ._._..©N.- 1.0—.- ..:_.m—.- ..:..h~.- file.- de :6 -Nm.ao>vu:omo.~n—.©N
$5

3.3.. 1.2.. giwo. ._.._.Nm.- ...*@N.- 1.0—.- ......©N.- ..:..mm.- ..:..w~.- mod mm; “gnu-Houaofihfim
$335.63

ELY-IL *ivw. sir-m.- ..:..vM.- Egvmr *...oM.- .Ibmr *...NN.- $4 and memuswzofiém

as».

firmwwx ..:_.om.- .3..th 1.2.- ......_N.- ..:..mm.- ..:..ON.- and am; omcmwéofimOdN

AME/v EoEflEEOU

va6 Elm. 1.5—. 31%. ..:..mm. ibm. :6 Cl“ ECONN

33v «Goggou

Mde Nod- .I.Nv. ......mm. cod find vim EcomuameAN

SE sonogram

mmd 1.2. .I.NN. ._:_.wm. wad vmd EBSwGEdN

33v cowommflumm

ems ..:..wN. mod- cwd 55m Hmoomd—

$26 zomwoflmflam

vad *2. de bod $825353

.2 MS .2 gauge $6.:

 

om mm vm mm mm S cm 2 w_ E Om :82

 

.8328 v 2.3

140

 

:8. :2. :2.- :2. :2. :8. :2. :2. :2. :3. :8. 2.. 8.... 833858882
:3. 8. :2.- 8. 8. :2. 8. .8. :8. :2. :2. 8m. 88 8% 88888 .2
:8. .8. :8..- :8. :8. :8 :8. :8. :8. :8. :2. 3m. 8.". :38 8 28082 .2
:8. :8. 8. 8. :2.- :8. :2.- :8. 8.- .8. :8. 8. 8.8 83:80.2
.8.- 8. 8.- 8. 8. 8. 8. 8.- :2.. 8. 8. 3. 8.8 838.88%.3W
25
:2. 8.- S. 8. 8 8. 8. 8.- :2. .8. 8. 8. 8.8 88% 85.2228
83
:8. 8. :2.- :2. :2. ::. :8. 8. :8. 8. :2. E. 8..“ 88825 38m .2
82: 88.8.5
2.. 8. :8.- :8. :2. :2. :8. :2. :2. :2. :2. 8. :8 2.88%.:
2880302:
8. 3.2.- :8. :8. :8. :8. :8. .8. :8. :8. 8. 8.8 83323882
8.. :2..- :8.- ..8- :8.- :8.- :8.- :8.- ..8.- 8m. 82 0629:... 80.2.8
8. :8. :8. :8. :8. .8. :8. :8. 3. 88 00.8.8288
8. :3. :8. :8. :2. :8. :8. 8. 8.8 cc 22828
8. :3. :8. :8. :8. :8. 8. 88 85.38.88
R. :8. :8. :8. :2. 8. 8.... £8.88
8. :2. :3. :8. 3. 2.... 88.8.2...
8. :2. :2. 8. 88 883 3886-”;
8.. :8. G. 88 388856888
8. 8. 8.3 382888.... ._

: 2 a w 8 8 m 3 m 8 _ am an:

 

.mcoufloboouowfi 28m 6on 88:53. ram waged .m 033.

I4]

 

88- :8. :2.- 88 .8. .8. 8.8 .8. 8.8- .8.- .8. 8.8 2.8 23538298
SE £0323
25888

:3.- .8.- o 8.8 8.8 88- 8 :8.- :2.- 8.8- .8.- 88 8.2 33583.2

2 o. o w h c w v m m _ Gm .802

 

.8288 m 038

I42

.3. W a: .3. w a... $ch

 

 

3.2. 8.- .2. 8. 8. .2. 8.- 8.- 8.- 230858058
838
22>mnom 8:05:on
8. :8.- :2..- :2.- :8.- :2. :8.- :3.- . 23388832
2. :2.. :2. .8. 8.- :8. :8. 833858882
8. :8. :2. :2.- :8. :8. 225 88288 .2
2.. 8.- 8.- :2. :2. :38 mm 2888.3 .2
22 :8.- 8. 8. 8550.2
3.2 :2.- :2.. 838288.:
838
2. :8. E88 858 .2
55
8. 88825 38m .2
8 2 2 2 2 2 3. 2 2

 

.8888 m 038

I43

 

58

 

:8. :2. 8.- ..2. :2. ..2. :2. 8. .8. .8. :2. 3.. 88 8088882
:8. 8. 8.- 8. 8. .8. 8. 8. :8. 8. .8. 8. 28 83888888.:
:8. 8.- :2.- 8. :2. :2. :3. .8. :2. :2. .8. 8. 88 33888882
328388
:2. :2. 8. 8. :8.- :8. :2.- :8. 8.- .8. .8. 8. 20m $6 .2
.8.- 8. 8.- 8. 8. 8. 8. 8.- ..:.- 8. 8. 2.. 8.8 85284.2
28
:2. 8.- 8. 8. 8. 8. 8. 8.- ..2. .8. 8. 8. 88 8288 888 .2
SE 88888—
:8. 8. :2.- ..:. ..2. ..:. :8. 8. :8. 8. :2. E. 8.". 38m .2
$8 888888
8. 8. :8.- :8. :2. :2. :8. :2. ..2. :2. :2. 8. :8 88888::
2880235
8. :2.- :2. :8. :2. :8. :8. .8. :8. :8. 8. 8: 8888088882
8. :2..- :8.- :8: :8.- :8.- :8.- :8.- :R.. 8. o: co 205-888
E. :8. :8. :8. :8. .8. :8. :8. 3.. 28 00.8388.»
8. ..:.. :8. :8. ..2. :8. :8. 8. 88 00 8882.8
E. ..:.. :8. :2. :2.. :8. 2.. 88 8588888
R. :8. :8. :8. :2. 8. 88 8880.8
8. ..2. ..:.. :8. 2.. 28 082380:
88%
8. ..2. :3. 8. 8.... 8588-88
8. :8. G. 8.8 888882888 .N
8. 8. 8.2. 8283883 ._

: 2 a m 8 8 m 2. m N _ am 832

 

acorn—88885 28m 382 88:55.84 323% .o 033-

I44

 

 

8N

. . . .- .- .- . 88 N8 :8 880888858

.8.- .8- 8 8- 88 o 88 o 88 88 :8 N 838885

88880

8N; 88- :. 8.- 8.8- .8. 8.8- :8. :2.- ..:.- :2.- 88 8.2 83888232
: S o w 8 8 w v m N 2 am so:

 

88888 8 288

I45

.8. w 8...... .8. w 8... H882

 

 

 

22.-N
.2 :8. :VN; :2.- 8.- :8.- :8. :2.- :2.- Nm. :8 msotsaoadN
as: 88283
8.8880
8. :8.- ..8.- :2.- :_N.- 8. :2.- :8.- 8. 82 838883728
25
R. :8. LN. 88. 8.- :8. :3. 3.. 88 808888 .2
8. :2... :2. :2.. :8. :2.. 8. 83 83888882
8. :2. 8.- ..:. :2. 8. 8.8. $358822
Abmfioficbv
82 :2.- .8. 8. 8. :8 $5 .2
82 :2.- ..:.- 8. 8.8 838288.:
ANBV
2. :8. N8. 8N 888m 888 .2
326 82:23::
E. E. 8.». 38m .2
8N 2 2 2 2 2 E 2 2 am 88.2
/II

3888 8 8:

I46

 

 

82 E.- 8. 8.- 8. 8. 8. 8. 8.- 2.- 8. 8.- 2.0-32288.2

ASE toga—2m

82 8.- 8.- 8. 8. 8. 8. 8. 2m. 2. 2. baa-.222

2880335

283258208

82 8.- 8. 8. G. 8. 22.. 8. 2. 8. 2:

82 8.- 8.- 8.- 8.- 8.- 8.- 8.- 8.- 00 295. 23.2.8

82 8. 2.. 8. 8.. :. 8. 8. 00.88288

82 8. 8. 8. 2. mm. 8. ooh-232.8

82 8. 8. 8. 8. 8. 88.3038 .8

82 8. 2. 8. 2. £858

82 2. 8. 8. 082.8522.

828%

82 mm. 2. 3883 .m

wmocmzomwcomomcomv

82 3.. .N

GOA mmDfiO—Q—woouw< .—
2228888mvmm2

 

2032 count-<- bfim 05 E 832.22g 2:83 msocowoxm 05 mo macaw—2.80 s 2an

I47

 

82 2.- 8. 8.- 8. 8. 8. 8. 8.- 22.. 8. 8. 8322822828
83

82 8.- 8.- 8. 8. 8. 8. 8. 8. 2. 2. 2888 888.:
2880235

82 8.- 8. 8. 8. 8. 8. 2.. 2. 8. 833082822
82 8.- 8.- 8.- 8.- 8.- 8.- a..- 8..- oc 203. 28.8

82 8. 8. 8. 8. :. 8. 8. 8.85:8.»

82 8. 8. R. 2. 8. 8. 82832.8

82 8. 8. 8. 8. 8. 85.8028 .8

82 8. 2. 8. 2. 285 .m

82 2. 8. 8. 88.3202 .22

82 8. 2. 8288 28085 .8
oo._ 3.. 80530356800 .N

82
mmocofimoaw< .—

 

N2 : E a m N. c w v m N _

 

2382 Ecmcamwcoq 2: 5 2.32.22; 2:83 msocowoxm 05 mo macaw—2.80 .w oBm-w

I48

 

 

2.28- 288. N8- m8- N8. 28- 28- 88 82 ER ~22
cc 2283.. .222 .8
28- 88- 28- m8- :8- 2.28- 28- 88 82 ER 222
ow 28283.. 288 .2
228 28 :8 88 2.28 288 88 88 88 2.28 8 00 .228. 20.2 .2
28 N8 28 88 28 28 2.28 88 2.8 ER 2 oc .228. .20-2 .2
28 288 88 m8 28 2.28 28 88 88 ER 222 00 822222 2.5-2 .2
28 88 88 88 88 28 28 88 28 2.28 8 8 E28: .2
28 S8 88 m8 :8 N8 28 88 88 :8 ~22 ow E2822 .2
2.28 88 88 88 28 N8 28 88 8.8 ER :2 ow E2322 .2
228 28 88 88 2.28 28 28 88 28 2.28 22 2825 .2
28 88 88 R8 28 28 28 88 E8 ER 8 2828 .2 2
28 288 2.88 88 28 2.28 :8 :8 28 28 222 .8826 82
88 28 88 288 88 28 288 88 8.8. 2228 22 228222
.226 228208 8
88 28 88 88 28 28 2.8 288 88 2.28 S 223222
.2226 228288 .22
2 28 88 88 N8 28 88 88 88 28 2.2 223222
.2226 228288 .8
2 88 :8 88 2.8 88 88 N: 2.28 8 .892 8
2 88 88 88 88 $8 88 28 2 28.2 8
2 88 88 88 88 88 ER :2 20.2 .2.
2 88 28 88 88 2.22 8 .2852 .m
2 28 88 8 ER ~22 23852 .N
2 88 2.2-. 22 :2 23852 .2

8 8 m m 8 d8 22832 z

 

20on scum-£82 ham 2: 5 832.29 32030 05 waofi< 222882828285 d 033-

I49

 

 

 

 

 

88.8- 8.8 8 88.8 28.8- 28.8- 8 88.8 88.2. 2 22 82
28.2.2228 82258.8 .82.
88.8 8.8 88 88.8 88.8 88.8 28.8 22.8 88.2. 82 22 82
28.2.2228 8222.288 .88
8 88.8 88.8 88.8 28.8 88.8 88.8- 8.8 88.2. 282 2 22
8.2.2.28 82258.2 .88
8.8 88.8 8.8 82.8 8 88.8- 28.8 88.8 88.8 8822 2.2 22.2 28288 .88
8 88.8 88.8 88.8 88.8 28.8. 88.8- 2 88.8 882 2 82 22.2 28288 .88
28.8 22.8 88.8 8.8 28.8- 28.8 8 8.8 88.8 8822 82 222 28288 .88
88.8 88.8 8.8 88.8 2.8 8 8.8 28.8 88.8 82 2 22 22.2 28288 2.8
88.8 2.8 88.8 2.8 2.8 22.8 2.8 2.88 28.8 822 8.2 22.2 825888... .88
28.8 82.8 8.8 2.28 82.8 88.8 22.8 88.8 88.8 8822 8.2 22.2 8228888.. .88
88.8 88.8 8.8 2.28 82.8 88.8 88.8 88.8 88.8 222 2.2 22.2 8282888... 28
88.8 82.8 82.8 8.8 2 2.8 22.8 2.8 8.8 2.8.8 288 8.2 .8888. .88
8.8 2.8 2.8 8.8 2 2.8 22.8 2.8 88.8 88.8 8288 8.2 .828... .88
8.8.8 8.8 88.8 88.8 22.8 2.8 2.8 28.8 8.8 8288 2.2 3.828 .88
288 8.8.8 2.8.8 8.8 2 2.8 88.8 8.8 88.8 28.8 2288 8.2 .2828 .8882 .88
88.8 88.8 8.8 82.8 2.8 88.8 88.8 88.8 28.8 2288 8.2 .8828 28.2282 .88
28.8 8.8 88.8 88.8 2.8 82.8 88.8 88.8 2.8 8288 2.2 .8828 288282 .8
2.8 88.8 8.8 828 88.8 88.8 88.8 88.8 8.8 8288 8.2 .888 2.8
2.8 88.8 88.8 82.8 2.88 28.8 88.8 28.8 28.8 288 8.2 .880 .88
8.8 88.8 28.8 88.8 82.8 88.8 2.88 88.8 88.8 8288 2.2 .880 .88
.88- 88.8- 88.8- 8.8- 88.8- 2.8- 82.8- 88.8 8.2 .8288 8.2
00 8283.. 8.8.2 .28
8 8 8 2. 8 8 2 d8 8.82 z

 

88882288 8 82.288

150

 

 

8.8 28.8 8.8 22.8 88.8- 88.8 88.8- 2.88 88.8 82.22 2..2 .888 .2882 .88
8.8 88.8 2.8 2.8 2.88 28.8- 8 8.8 88.8 82.22 8.2 5&8 .2882 .88
8.8 88.8 2.88 8.8 88.8 88.8 8.8 28.8 88.8 82.2 2 8.2 8&8 .2882 .88
88.8 88.8 88.8 82.8 88.8 88.8 88.8 88.8 88.8 82.2 2 2.2 .5288 .2882 .8
88.8 28.8 82.8 88.8 82.8 2.8 82.8 82.8 88.2. 2.22 8.2 .888 280 .8
22.8 2.8 82.8 82.8 2.8 2 2.8 2.2.8 8.8 28.2. 82.22 8.2 .8288 28.5 2.8
82.8 28.8 2.8 88.8 82.8 2.28 82.8 88.8 82.2. 2.22 2.2 .888 280 .88
8 88.8 2.88. 8 88.8 88.8 88.8 2.88 88.8 82.22 8.2 28 82.2 .88
28.8 2.8 88.8- 28.8 88.8 2.88 88.8 2.88 88.8 82.22 2.2 .288 82.2 .28
88.8 88.8 2.8 22.8 28.8- 88.8- 88.8- 8.8 88.8 82.22 8.2 .288 28288 .8
88.8 22.8 22.8 82.8 28.8 88.8 28.8- 28.8 88.8 82.22 2.2 .28 28288 .82.
8.8 82.8 2.8 2.2.8 828 8.8 88.8 2.88 8.2. 82.2 2 8.2 .288 .888. .82.
2.88 88.8 88.8 2.8 2.28 88.8 88.8 .888 88.2. 82.22 2.2 .288 .882. .82.
28.8- 82.8 2.8 2.28 82.8 28.8 82.8 88.8 8.82. 888 2.2 8888.282
.2288 .88... .82.
8.8- 2 2.8 88.8 88.8 88.8 8 8 82.2 88.8 888 82 8888.282
8288 .88... .82.
8 2.8 88.8 2.8 2.8 28.8 88.8 88.8 82.2. 88 82 8882.228
8288 .888. .2.2.
88.8 82.8 2.8 82.8 2.8 8.8 2 2.8 88.8 8 2 .2. 2.88 22 8882-882
8288 .88... .82.
88.8- 88.8 88.8 88.8 88.8- 8 8.8- 8.8 88.2. 2 2 2 82
2.8.2.288 822582 .82.
8 28.8- 8.8- 88.8- 2 2.8- 88.8- 88.8- 8.8 88.2. 882 2 2.2
2.8.2.288 822882 22.

8 8 8 2. 8 8 2 .828 882 z

 

882.2288 8 82888

IS]

 

 

 

8.8 88.8 88.8 88.8 88.8 2.8.8 88.8 88.8 82.8 8888 2 2 .8282 28.28.82 .88

2.8- 2 2.8- 2.8.8- 88.8- 2.8.8- 88.8- 88.8- 88.8 2.8. 2 82 2 2 82 388.2223

3 “:35 .mm

88.8- 88.8- 28.8 88.8- 88.8- 28.8- 28.8 88.8 28.2 82 2 2 22 38.8.2223

ow H8235 .50

2 2.8- 8 2 .8- 88.8- 82.8- 82.8- 2.28. 82.8- 2.8.8 8.2 882 2 82 8822.3

88 82.288228 .88

88.8- 8 28.8 88.8- 88.8- 88.8- 88.8- 8.2 88.8 282 2 22 88223-2

28 82.2882: .88

88.8- 88.8- 28.8- 88.8- 2.8- 88.8- 2.8- 28.8 82.2 22.22 82 8.282 8825.228 .2.8

88.8- 88.8 88.8 28.8 88.8- 88.8- 88.8- 2.8.2 88.2 82.22 22 .8282 882.228 .88

8.8 88.8 2.88 22.8 2.88 82.8 88.8 88.8 88.8 8288 28.22.82.882 $288 .88

88.8 8 88.8 88.8 88.8- 88.8- 2.8.8- 2.8 88.8 882 2 8.28.28 .28

2.8 88.8 88.8 88.8 288 2.8.8 288 88.8 88.8 82.2 2 8.2 .8680 822882 .88
8 8 8 2. 8 8 2 .828 882 z

/

«80385580 Q 03er

152

 

153

28.8- 88.8 88.8 88.8 88.8 88.8 88.8 2.8 88.8 88.8 88.8 2888288888288
82.8 82.8 88.8 82.8 82.8 82.8 82.8 28.8 82.8 228 2.2.8 8.2888988
82.8 82.8 8.8 8.8 88.8 82.8 2.28 2.88 82.8 22.8 88.8 8.2.8.5088
88.8 82.8 8.8 82.8 82.8 82.8 82.8 88.8 82.8 88.8 88.8 2.2.8-880.88
88.8- 88.8- 88.8- 22.8- 888- 88.8- 82.8- 82.8- 82.8- 82.8- 82.8- 8.2
cc 8283.. 8.8.2 .28
2.88. 22.8- 88.8- 22.8- 82.8- 88.8- 22.8- 82.8- 828. 2.2.8- 228. 8.2
00 8283.. 8.5.2 .88
82.8- 88.8- 8.8- 88.8- 28.8- 28.8- 82.8- 88.8- 82.8- 88.8- 82.8- 2.2
09 28>... 8.8.2 .82
2 88.8 3.8 28.8 88.8 88.8 82.8 82.8 82.8 2.8 82.8 8.2 082.888.85.282
2 28.8 8.8 2.8.8 88.8 82.8 88.8 82.8 82.8 82.8 8.2 8.2.28. 85.2.82
2 88.8 22.8 22.8 82.8 88.8 88.8 82.8 82.8 2.2 8.8.2.... 85.2.82
2 88.8 3.8 82.8 88.8 828 2.28 82.8 8.2 cc 852882 .82
2 88.8 28.8 88.8 8.8 82.8 82.8 8.2 ow 8.2882 .2.2
2 28.8 88.8 8.8 2.8 82.8 2.2 00 8.828822 .82
2 8.8 3.8 88.8 82.8 8.2 285 .82
2 88.8 88.8 82.8 8.2 8.828 .2 2
2 2.28 88.8 2.2 8.8.882 .82
2 2.2.8 8.2 9282.2
.225 228288 .8
2 8.222.582
.225 88288 .8

82 82 82 82 2.2 82 82 22 82 8 8 I

 

/

6225250 a 035.

 

88.8 2.8 88.8 88.8 2 2.8 82.8 88.8 2 2.8 88.8 88.8- 28.8- 82 38.-2.282
22.28 .888. .2.2.
8 88.8 88.8 88.8 82.8 82.8 88.8 82.8 88.8 8 88.8 22 882.2282
2288 .88... .82.
88.8 88.8 88.8 8 88.8- 28.8 28.8- 8 2.8.8- 888- 88.8- 82
2.8.2.288 822.58 .82.
88.8- 28.8- 8 88.8- 2.88- 888- 88.8- 288. 2.8.8- 888- 2.8.8- 2.2
2.8.2.288 82288.2 .22.
28.8 88.8 88.8 28.8 88.8 88.8 28.8 88.8 88.8- 88.8- 2.88. 82
2.8.2.288 82288.2 .82.
8.8 2.8.8 2.8.8 2.8.8 2.88 88.8 8 88.8 28.8 88.8 28.8 82
2.8.2.288 82288.2 .88
88.8 88.8 8 28.8 2.8.8 88.8 88.8- 28.8 2.88. 28.8 88.8- 22
28.2.2228 82225-2 .88
88.8 22.8 2.8 88.8 22.8 88.8 88.8 2.8 2.88 88.8 88.8 2.2 22.2 28288 .88
28.8- 88.8 88.8 28.8- 88.8 88.8 8 2.88 88.8 88.8 8 82 22.2 28288 .88
2.88 88.8 2.8 2.8.8 8.8 88.8 2.8 82.8 288 2.8.8 2.8.8 82 .2.2 28288 .88
88.8 88.8 88.8 88.8 2.88 88.8 88.8 82.8 88.8 88.8 88.8 22 22.2 28288 .2.8
2.88 88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 88.8 8.2 .2.2 82282.8... .88
88.8 88.8 88.8 88.8 88.8 88.8 8.8 8.8 82.8 88.8 88.8 8.2 22.2 82252.8... .88
2.8.8 2.88 88.8 88.8 88.8 88.8 82.8 28.8 88.8 88.8 2.88 2.2 22.2 82.58828 .28
82.8 82.8 82.8 82.8 8.8 82.8 88.8 88.8 88.8 2.8.8 88.8 8.2 8.82.. .88
82.8 82.8 82.8 82.8 82.8 82.8 88.8 88.8 88.8 88.8 88.8 8.2 8.8.... .88
82.8 88.8 28.8 88.8 88.8 88.8 88.8 82.8 82.8 88.8 2.88 2.2 .8288. .88
28.8- 88.8 88.8 88.8 22.8 22.8 88.8 88.8 2.88 88.8 88.8 8.2 .838 28882 .88
88.8 82.8 88.8 82.8 82.8 82.8 88.8 228 2.8.8 288 88.8 8.2 .8828 28882 .88
82 82 82 82 2.2 82 82 22 82 8 8

 

/

32222-22250 a 035-

l54

 

 

88.8 88.8 88.8 88.8 88.8 88.8 88.8- 8.8. 2.8.8 8.8. 2.8.8- 22 8828882
88 828888.28 .88
88.8- 28.8 88.8- 88.8 88.8 88.8 88.8 88.8 28.8 88.8- 88.8- 82 .8282 882.2228 2.8
88.8 228 2.8.8 88.8 88.8 88.8 88.8 2.8.8 88.8 88.8- 28.8- 22 .8882 882.2280 .88
88.8 88.8- 88.8 88.8- 88.8- 2 2.8- 82.8- 8 2 .8- 2 2.8- 8 2 .8 2.2 .8 288.22.82.82 $228 .88
88.8 88.8 2.88 8 88.8 88.8 2.88 88.8 2.88 88.8 88.8 9288.28 .28
88.8 82.8 2.8 22.8 82.8 82.8 2.8 2.8 82.8 88.8 82.8 8.2 88880 822882 .88
88.8 88.8 88.8 88.8 88.8 2.8 88.8 2.8 88.8 8 88.8 2..2 .2888 222882 .88
88.8 88.8 88.8 88.8 88.8 2.8.8 88.8 88.8 88.8 2.8.8 88.8 8.2 .5880 822882 .88
88.8 88.8 88.8 28.8- 88.8 88.8 2.8.8 2.8.8 28.8 88.8 2.88 8.2 8.8288 222.82 .88
88.8 88.8 22.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 2.2 8:88 8228.82 .88
82.8 88.8 82.8 2.8 88.8 82.8 828 2.2.8 2.8 88.8 88.8 8.2 .8888 28828 .88
2.8 2.8 2.8 88.8 88.8 88.8 2.8 22.8 88.8 88.8 2 2.8 8.2 .8888 28888 .2.8
88.8 2.8 88.8 2.88 88.8 88.8 88.8 2 2.8 88.8 22.8 2.28 2.2 .558 28882 .88
88.8- 88.8- 2.8.8- 88.8- 2.8.8- 2.88. 88.8- 88.8- 88.8- 88.8- 88.8- 8.2 .288 82.2 .88
28.8- 28.8- 88.8- 2.88. 88.8- 2.8.8- 888- 88.8- 2.8.8- 2.8.8- 8 2.2 .288 .82-2 .28
88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 88.8 88.8 8.2 .288 282888 .88
88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 2.8.8 88.8 2.2 .288 282888 .82.
88.8 88.8 28.8 88.8 88.8 88.8 88.8 88.8 8 88.8 88.8 8.2 .288 .88... .82.
2.8.8 88.8 88.8 28.8- 88.8 2.88 88.8 88.8 28.8 88.8 88.8 2.2 .288 .8882.- .82.
88.8 2 2.8 88.8 828 2.2.8 82.8 8 2 .8 8 2 .8 88.8 2.8.8- 88.8 2.2 888822-282
2288 8882.. .82.
88.8- 88.8 88.8 88.8 2.8 82.8 88.8 8 2 .8 2.88 88.8- 2.8.8- 82 88882-2282
2288 .888. .82.

82 82 82 82 2.2 82 82 22 82 8 8

/

6025250 a 073%

155

 

 

28.8 88.8 88.8 88.8 88.8 88.8 88.8 8 88.8 22.8 82.8 22 .8820 288.285 .88
88.8- 2.88- 888- 88.8- 8 288. 2.8.8- 88.8- 88.8- 88.8- 2.8- 82 >8828822.22
82 28282 .88

28.8- 88.8 88.8 28.8 2.88 88.8 28.8- 28.8 88.8 88.8- 2.8.8- 22 38.8.2223
82 288282 .88

88.8- 2.88. 88.8- 88.8- 88.8- 2.88. 2.88. 88.8- 88.8.. 88.8- 88.8- 82 88288-2
28 82.288228 .88

82 82 82 82 2.2 82 82 22 82 8 8

 

888822880 8 82.288

156

 

 

88.8 2.8 82.8 82.8 2.8 88.8 88.8 88.8 8 2.8.8- 888. 2.5.2.288 8222888 .82.
22.8 2.2.8 88.8 88.8 88.8 88.8 828 2.8.8 2.8- 88.8- 88.8- 82
2.8.2.288 82288.2 .88
2.8 82.8 2.8 82.8 88.8 22.8 22.8 88.8 2.8.8- 88.8- 88.8. 22
. 2.8.2.288 82225.2 .88
82.8 82.8 82.8 82.8 82.8 2.8 2.88 88.8 88.8- 88.8- 88.8- 2.2 22.2 282888 .88
88.8 88.8 82.8 82.8 82.8 88.8 28.8- 88.8- 88.8- 88.8- 88.8- 82 22.2 282888 .88
2.8 82.8 288 2.88 82.8 88.8 88.8 88.8 88.8- , 88.8- 88.8- 82 228 2.2288 .88
88.8 82.8 22.8 82.8 82.8 88.8 28.8 88.8- 88.8- 88.8- 2.8.8- 2222.2 2.2288 .2.8
88.8 22.8 88.8 2.8 88.8 8.8 88.8 22.8 88.8- 2.8- 88.8- 8.2 22.2 822882.88... .88
88.8 82.8 88.8 22.8 22.8 82.8 88.8 2.8 88.8- 88.8- 88.8- 8.2 22.2 82.8888... .88
88.8 228 2.2.8 82.8 828 2.2.8 88.8 88.8 2.8.8- 88.8- 88.8- 2.2 22.2 8282888.... .28
88.8 8.8 2.8.8 828 88.8 82.8 8.8 88.8 88.8- 82.8- 82.8- 8.2 .8828... .88
2 88.8 8 2.8 88.8 82.8 88.8 88.8 82.8- 82.8- 22.8- 8.2 .8828... .88
2 22.8 88.8 82.8 28.8 88.8 88.8 2.8.8- 88.8- 8.8- 2.2 .8828... .88
2 28.8 28.8 88.8- 82.8 88.8 2.8- 82.8- 28.8- 8.2 .2828 28.8-.2 .88
2 88.8 88.8 2.8.8 82.8 82.8- 28.8- 88.8- 8.2 .8828 28.882 .88
2 88.8 88.8 82.8 82.8- 82.8- 82.8- 2.2 .8828 28.882 .88
2 82.8 28.8 82.8- 2.2.8- 828. 8.2 .8880 .2.8
2 88.8 88.8- 88.8- 88.8- 8.2 .888 .88
2 82.8- 82.8- 82.8- 2.2 .8880 .88
2 88.8 8.8 8.2
cm 2.283.. 228.2 .28
2 88.8 8.2
00 22288.... 28.2 .88
2 2.2
00 8288... 828.2 .82

88 88 88 88 88 88 88 88 28 88 82

 

888822880 8 82.288

157

 

 

2.8.8 2 2.8 88.8 88.8 88.8 88.8 8 88.8 88.8- 88.8- 82.8- 8.2 88880 22282.2 .88
88.8 82.8 88.8- 2.88 28.8 22.8 88.8 2.8 28.8- 288- 2.8.8- 2..2 8.82828 222882 .88
88.8 2 2.8 88.8 2.8 88.8 88.8 28.8- 28.8- 28.8 28.8- 8 8.2 28.88 222.82 .88
88.8 88.8 88.8 2.8 88.8 88.8 8 88.8- 8 88.8- 88.8- 8.2 .8828 222882 .88
22.8 2.28 88.8 88.8 88.8 2.2.8 88.8 2.8.8 28.8- 88.8- 2.8.8- 2.2 82.8882 222882 .88
82.8 82.8 88.8 82.8 88.8 82.8 82.8 88.8 82.8- 22.8- 82.8- 8.2 .8888 28.-2 .88
2.2.8 828 2.8.8 828 88.8 8.8 82.8 82.8 82.8- 88.8- 88.8- 8.2 .8828 28822 .2.8
2.8 82.8 88.8 82.8 22.8 82.8 82.8 88.8 82.8- 2.2.8- 88.8- 2.2 8:880 28822 .88
88.8- 288. 2.2.8 828 82.8 88.8 88.8 88.8 88.8- 88.8- 28.8- 8.2 .288 82.2 .88
88.8- 88.8 82.8 828 2.2.8 88.8 22.8 2.8 28.8 28.8- 8 2.2 .288 82.2 .28
88.8 82.8 288 2.8.8 8.8 88.8 28.8 88.8 88.8- 88.8- 88.8- 8.2 .288 282888 .88
88.8 82.8 8.8 2.88 28.8 88.8 88.8 88.8 88.8- 88.8- 88.8- 2.2 .288 282888 .82.
88.8 2.28 88.8 82.8 88.8 82.8 22.8 88.8 88.8- 88.8- 22.8- 8.2 .288 .888. .82.
82.8 82.8 88.8 82.8 88.8 2.2.8 88.8 88.8 88.8- 88.8- 88.8- 2.2 .288 .888. .82.
2.88 88.8 82.8 82.8 88.8 2 2.8 2.2.8 88.8 82.8- 82.8- 2.28- 2.2 888822282
2288 .88... .88
2.8.8 88.8 82.8 82.8 2.2.8 88.8 2 2.8 88.8 88.8- 2 2.8- 82.8- 82 888.2282
2288 .8888. .82.
88.8 88.8 2 2.8 8 2 .8 2.8 88.8 2 2.8 2.8.8 2 2.8- 2.2 .8- 2.2.8- 82 88882-228
2288 8882.. .2.2.
88.8 2.8 82.8 8 2 .8 2.28 88.8 82.8 88.8 88.8- 2 2.8- 82.8- 22 888.2282
2288 .88.... .82.
88.8 88.8 82.8 82.8 88.8 88.8 88.8 88.8 88.8 8 28.8- 82
228.2888 82258.2 .82.
88.8 2.8.8 88.8 88.8 88.8 2.88 88.8 88.8 28.8 2.88 28.8 2.2
2288.288 82222.88 .22.

88 88 88 88 88 88 88 88 28 88 82

 

88882288 8 82.288

158

 

 

 

 

 

159

2.8.8 2.8.8 2.8.8- 288. 88.8- 88.8- 2.8.8- 88.8- 2 2.8- 88.8- 88.8- 22 28826 288.285 .88

88.8- 82.8- 88.8- 2.2.8- 888. 2.8- 88.8- 88.8- 2.8 88.8 28.8 82 38.28.2223

82 288282 .88

8 88.8- 88.8- 8 2 .8- 88.8- 2.88. 8 88.8 28.8- 88.8 8 22 38.822223

82 288282 .88

88.8- 82.8- 88.8- 2.2.8- 88.8- 82.8- 82.8- 2.8- 2.8 88.8 88.8 82 8828882

28 82.288228 .88

28.8 88.8- 2.88- 888- 88.8- 2.8- 88.8- 28.8- 88.8- 8 88.8- 22 88288-2

28 8222888.:- .88

88.8- 82.8- 8 22.8- 88.8- 82.8- 88.8- 2.88- 88.8 88.8 88.8- 82 .8282 8822882822 .28

88.8 88.8- 88.8- 88.8- 88.8- 88.8- 8 2.8.8 88.8- 88.8- 88.8- 22 .8282 88282828 .88

88.8 2.8.8 28.8 88.8- 8 82.8 88.8 88.8- 88.8- 28.8 88.8 288.282.2232 $228 .88

88.8 8 88.8- 2.8- 2.8- 28.8- 28.8- 28.8 28.8- 88.8- 88.8- 88288.28 .28
88 88 88 88 88 2.8 88 88 28 88 82

 

88882288 8 82.282.

 

 

82.8 82.8 82.8 88.8 88.8 82.8 22.8 8.8 2.2.8 82.8 88.8 8288822282
228888282.
828 88.8 82.8 2.8 88.8 88.8 22.8 88.8 8.8 82.8 88.8 82 888822282
2288 .8888. .88
2.8 2.2.8 2.2.8 88.8 22.8 2.28 82.8 28.8 82.8 82.8 88.8 8288822282
2288 .888... .2.2.
2.28 82.8 82.8 88.8 22.8 2.28 82.8 88.8 82.8 82.8 88.8 22888822282
2288 .888... .82.
88.8 88.8 82.8 82.8 82.8 88.8 82.8 82.8 22.8 82.8 88.8 82
2.8.2.288 82288.2 .82.
82.8 28.8 88.8 88.8 88.8 82.8 88.8 82.8 88.8 82.8 28.8 2.2
2.8.2.288 82288.2 .22.
2 88.8 82.8 82.8 82.8 82.8 82.8 82.8 82.8 8.8 88.8 82
2288.288 82288.2 .82.
2 82.8 88.8 88.8 88.8 82.8 22.8 82.8 82.8 22.8 82
228.2888 82888.2 .88
2 82.8 82.8 8.8 82.8 82.8 82.8 28.8 22.8 22
228.2888 82288.2 .88
2 88.8 88.8 82.8 2.88 82.8 88.8 8.8 8222.2 28288.88
2 88.8 88.8 8.8 2.28 88.8 82.8 8222.2 28288.88
2 88.8 88.8 88.8 88.8 82.8 82 22.2 282888 .88
2 88.8 88.8 88.8 88.8 2222.2 282888 .2.8
2 88.8 88.8 88.8 8.2 22.2 82.228888... .88
2 28.8 88.8 8.2 22.2 82.88888... .88
2 88.8 2.2 22.2 8288888... .28
2 8.2.8888<.88

88 88 88 88 88 88 88 88 88 28 88

 

888822880 8 82.288

160

 

 

2 2.8- 2.8.8- 88.8- 88.8- 2.88- 2.8- 2.2.8- 888. 88.8- 88.8- 8 8.288228 .28
88.8 88.8 88.8 82.8 82.8 28.8 82.8 88.8 88.8 82.8 88.8 8.2
228880 .222882 .88

2.8 88.8 82.8 28.8 88.8 88.8 2.88 82.8 2.2 .8 8.8 2 2.8 2..2
88.88 .222882 .88

2.8 82.8 88.8 88.8 88.8 88.8 88.8 82.8 82.8 8.8 88.8 8.2
288.80 .222882 .88

82.8 82.8 88.8 88.8 88.8 82.8 88.8 88.8 8.8 88.8 88.8 8.2
82.8822 .222882 .88

82.8 2.2.8 2.8.8 88.8 88.8 2.8 88.8 88.8 88.8 88.8 2 2.8 2.2
.5800 £me 6m

82.8 88.8 82.8 82.8 82.8 82.8 82.8 28.8 2.8.8 2.8.8 82.8 8.2
8.82822 28828 .88

2.2.8 8.8 82.8 22.8 82.8 828 2.2.8 828 8.8 28.8 2.2.8 8.2
828280 28822 .2.8

82.8 88.8 8.8 2 2.8 82.8 82.8 8 2 .8 88.8 88.8 88.8 2 2.8 2.2
88880 28882 .88

82.8 82.8 82.8 88.8 2.8.8 22.8 88.8 82.8 82.8 82.8 88.8- 8.2 .288 82.2 .88
828 2.28 82.8 88.8 2.8 22.8 88.8 82.8 82.8 82.8 2.8.8- 2.2 .288 .82-2 .28
82.8 88.8 82.8 88.8 88.8 88.8 88.8 88.8 828 2.8.8 828 8.2 .288 282888 .88

16]

E dam Ecom .98
mm .228m 6280.88 .wv
E gum 6283-... .28

:6 mod n 2 .o $5 2.6 and 2.0-o mmd 26 36 m 2 .o
36 02.9 vmd 36 E .o mmd NN-o de 2.22 3.6 mod
56 0N6 mmd mud end 36 mmd wmd de 32.-o 2.2-o

 

om mm 8m 0m mm Vm mm mm 3 Om

 

888822880 8 82.288

 

 

 

 

mod- 222.22- 022.22- No.22- 222.22- mod- 222.22- 822.22- mod 2.22.22- 2.22.22 2 2 ENC
2222082929 do
222. mmd- mmd- 2.2 .22- 22.22- 222.22- 82.22. m 2 .22- 82.22- NN.22- 2.22- N2 3822222223
02 2228822 .228
22.22- 82.22- omd- mmd- mmd- 28.8-.22- N022- NNd- de- 2022- M2222- 22 32822222223
8 2228222 .3
m 2.22- 022. mmd- o 2 .22- 2N.22- NN.22- o 2 .22- SS. 3.22- R22- 2 2.22- N2 magma-2
28 8222888.28 .88
m 2.22- 82.22- 8.22- omd- 8.8-.22- 2.22.22- 28.22- wmd- mad- 22..22- 222.22- 22 @2332
2o 82382222. .3
o 2 .22- 2 m .22- 8.22- 2.2 .22- a 2 .22- N22- 2.8-.22- 2 N22- «N22- vmd- 822.22- 2 62:2
watofiamv .280
2.2.22- 82.22- 3.22- 9.6- mmd- 22.22- mmd- 36. 8.22. 8022- 22 22
.8282 88822828 .88
mod 822.22 222.22 3.22- 222.22- 2.22.22- 822.22 2222.22 22.22 822.22 22.22 22.882822
.22882 $222 .88

222. am mm 2.2” cm mm .2.2-- mm mm 2.8. cm

 

.8.-888228822 8 82.288

162

 

 

 

 

 

 

82.8 82.8 82.8 82.8 88.8 88.8 22.8 88.8 88.8 88.8 8.2 8.880.222.8288
88.8 88.8 82.8 88.8 88.8- 28.8 88.8- 88.8- 88.8 88.8 8.2828288222288288
28.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 8.2 82828082288288
88.8 88.8 88.8 8.8 88.8 28.8 88.8 88.8 82.8 88.8 8.2 580.222.282.88
8.8 88.8 88.8 88.8 88.8 88.8 88.8 22.8 8.8 88.8 2.2688022288288
82.8 82.8 88.8 8.8 82.8 2.8 82.8 28.8 82.8 2.8 8.2.8.2880 2880.88
228 82.8 88.8 88.8 82.8 88.8 82.8 82.8 88.8 88.8 8.282.882.2888
22.8 22.8 88.8 88.8 88.8 82.8 28.8 88.8 82.8 82.8 2.2.8.8280 288.88
88.8 88.8 82.8 82.8 28.8 8.8 88.8 88.8 28.8 88.8 88.288.82.288
88.8 88.8 82.8 82.8 8.8 82.8 82.8 88.8 88.8 8.8 22.288.82.228
2 88.8 28.8 88.8 88.8 82.8 2.8 22.8 82.8 88.8 8.2.288 28288888
2 88.8 88.8 88.8 82.8 2.8 82.8 8.8 88.8 2.2 .288 282888 .88
2 88.8 82.8 82.8 22.8 82.8 82.8 82.8 8.2 .288 .8888. .88
2 82.8 82.8 2.2.8 828 82.8 82.8 2.2 .288 .8888. .88
2 88.8 88.8 88.8 88.8 2 2.8 2.2 888882282
2288 88828.82.
2 28.8 88.8 82.8 2 2.8 82 888882282
2288.888<.82.
2 88.8 88.8 2 2.8 82 88.28.2282
2288 .888... .88
2 2.8 82.8 22 888822282
22888882888
2 88.8 82
2.8.2.288 82288.2 .88
2 82
2.5828286 22288.2 .22.

88 88 88 88 88 88 88 88 88 28

 

||

88888888 8 82.288

I63

 

[

 

 

 

8 28.8 82.8 88.8 88.8- 28.8- 88.8- 88.8- 2.8- 88.8 22 288282 58.285 .88

82.8- 82.8- 82.8- 88.8- 82.8- 88.8- 82.8- 82.8- 828. 2.2.8- 82 >88.8.82.2,

82 288282 .88

88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 82.8.. 88.8- 22 38.822223

82 288282 .88

82.8- 8 2 .8.. 88.8- 28.8- 82.8- 82.8- 8.8- 28.8- 82.8- 82.8- 82 8828882

28 822288228 .88

88.8- 28.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 82.8.. 88.8- 22 8828882

28 82.2888228 .88

82.8- 82.8- 88.8- 88.8- 82.8- 88.8- 82.8- 8 2 .8- 82.8- 8 2 .8- 82 .8282 882285828 .88

88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 82.8- 2.8- 2 2.8- 88.8- 22 .8282 8.22.8882 .88

88.8- 88.8- 2 2.8 88.8 88.8 88.8 82.8 82.8 28.8 88.8 222888222882 .8822 .88

82.8- 82.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8. 88.8- 82.8- 8.288228 .28
88 88 88 88 88 88 88 88 88 28

 

88882288 8 82.288

164

 

 

82.8 88.8 88.8 88.8 82.8- 28.8 28.8 88.8 88.8- 8 22288282222888.8288
88.8- 88.8- 82.8- 82.8- 88.8- 88.8- 88.8- 88.8- 22.8- 82.8- 82 >8828822.22
8228828288
88.8- 88.8- 8.8- 88.8- 88.8- 82.8- 82.8- 82.8- 88.8- 88.8- 22382822223
82 288.82 .88
88.8- 82.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 82.8- 88.8- 82 8828882
28 82.288888 .88
88.8- 88.8- 88.8- 88.8- 88.8- 82.8- 82.8- 88.8- 22.8- 82.8- 22882882
28 82.288.28.88
88.8- 88.8- 88.8- 82.8- 28.8- 88.8- 88.8- 28.8- 82.8- 88.8- 82.8282 882288282288
88.8- 88.8- 8.8- 8.8- 88.8- 82.8- 82.8- 88.8- 82.8- 82.8- 22828288228088
88.8- 88.8- 88.8 88.8 88.8 88.8 2.8 88.8 82.8 82.8 288.282.228.28..282.88
28.8 28.8 88.8- 88.8- 88.8- 28.8 88.8 88.8- 2.8- 28. 8288.28.28
2 88.8 8.8 28.8 88.8 82.8 2.8 88.8 28.8 28.8- 8.2 88888222288288
2 88.8 88.8 88.8 82.8 82.8 88.8 28.8- 28.8 8.2 .5880 .222882 .88
2 88.8 88.8 82.8 82.8 82.8 8 28.8 8.2 .8580 222882 .88
2 28.8 82.8 82.8 28.8 88.8 88.8 8.2 82.22882 .222882 .88
2 88.8 82.8 82.8 88.8 88.8 2.2 8.82882 222882 .88
2 28.8 88.8 88.8 82.8 8.2 88880 28828 .88
2 88.8 2 2.8 82.8 8.2 2882882 28828 .88
2 82.8 28.8 2.2 .8288 28822 .88
2 88.8 8.2 .288 82.2 .88
2 2.2 .288 82.2 .28

88 88 88 88 88 88 88 88 88 28

 

88882288 8 82.288

165

 

 

2 28.8- 88.8- 88.8- 28.8- 88.8- 28.8- 82.8 88.8- 22.882822888888288
2 88.8 88.8 88.8 88.8 88.8 82.8- 88.8 82 28828.2282,
82 288282 .88
2 88.8 88.8 88.8 88.8 82.8- 88.8 22382822223
02 222882280
2 88.8 88.8 88.8 28.8- 88.8 82 882888-2
28 8222888.:- .88
2 88.8 8.8 82.8- 2.8 22 8828882
28 828888.22- .88
2 88.8 88.8- 88.8 82 .8282 88228222822 .88
2 82.8- 2.8 22 .8282 882222822 .88
2 88.8- 288.282.228.82 <20 .88
2 82288.28 .28
88 88 88 88 88 88 88 88 28

 

888822880 8 82.288

166

Table 10. Standardized Regression Coefficients in the Early Attrition and Longitudinal
Models.

 

 

Early attrition Longitudinal
STAGE 1 -- STAGE 2 r b r b
Perf. Avoid GO -- Acad. Int. -.09 -.15** -.09 -.09
Perf. App GO -- Acad. Int. .09 .07 .09 .11
Mast. GO -- Acad. Int. .10 -.15 .10 -.26*
Knowledge -- Acad. Int. .16 .07 .16 .23"
Careers -— Acad. Int. .30 .35" .30 .41“
Perseverance -- Acad. Int. .17 .06 .17 -.19*
Conscientiousness -- Acad. Int. .15 .05 .15 .01
Agreeableness -- Acad. Int. .12 .09" .12 .15"
Emot. Stab. -- Acad. Int. .15 .14" .15 .121”
Soc./Cult. Involvement —- Soc.
Int. .05 .01 .05 .01
Conscientiousness -- Soc. Int. .04 -.04 .04 -.05
Agreeableness -- Soc. Int. .17 .22" .17 .20"
Emot. Stab. -- Soc. Int. .23 .27" .23 .25"
STAGE 2 -- STAGE 3
Acad. Int. -- Satisfaction .44 .03 .38 .08’ll
Acad. Int. -- Ac. SE .21 .47" .26 .32M
Acad. Int. -- Commitment .43 .23" .29 .16"
Soc. Int. -- Satisfaction .62 .88" .48 .67"
Soc. Int. -- Commitment .53 .57" .44 .56"
Family Support -- Commitment .28 .13" .25 .12"
Family Support--Satisfaction .31 .02 .27 .07*
Acad. SE -- GPA -.04 -.04 .17 .21"
Shocks -- Satisfaction -.17 -.06** -.14 -.02
Shocks -- Dropout -.05 -.02 .09 .04
STAGE 3 -- STAGE 4
GPA -- Withdrawal Cog./Beh. -.21 -.24** -.21 -.45**
Satisfaction -- Withdrawal
Cog./Beh. -.48 -.09** -.35 -.10**
Comm. -- Withdrawal Cog./Beh. -.59 -.62** -.38 --41 **

STAGE 4 -- STAGE 5

Withdrawal Cog-lBeh. -- Dropout -.03 -.07* .23 .45"
*p S .05. **p S .01.

167

 

Table 11. Squared Multiple Correlations in the Models.

 

 

Early
Dependent Attrition Longitudinal
Independent Variables Variable R2 R2
Performance Avoid GO,
Performance Approach GO,

Mastery GO, Knowledge, Career

Orientation, Perseverance,

Conscientiousness, Agreeableness,

Emotional Stability Academic Int. .31 .21

Conscientiousness, Agreeableness,
Emotional Stability, Social/cultural

Involvement Social Int. .13 .11
Academic Integration Academic SE .22 .10
Academic SE GPA .00 .05
Academic Int., Social Int., Family
Support Satisfaction .78 .49
Academic Int., Social Int., Family
Support Commitment .44 .39
Withdrawal
Cognitions/
GPA, Satisfaction, Commitment Behaviors .51 .42
Withdrawal Cognitions/Behaviors Withdrawal .01 .20

 

168

 

 

 

88.8 88.8 8.8 88.8 88.8 82.8 82.8 88.8 88.8 82.8 8288 2.2 ..2828 28.882 .88
82.8 82.8 88.8 8.8 88.8 88.8 88.8 88.8 88.8 88.8 8288 8.2 .8888 .88
88.8 2.8 88.8 88.8 88.8 88.8 28.8 88.8 28.8 88.8 8288 8.2 .8880 .88
88.8 88.8 88.8 28.8 88.8 82.8 88.8 88.8 88.8 88.8 8288 2.2 .8888 .88
82.8- 8.8- 88.8- 88.8- 8.8- 88.8- 82.8- 82.8- 88.8 88.2 8288 8.2 00 82888. 228.2 .28
22.8- 82.8- 88.8- 8.8- 8.8- 8.8- 2.8- 2.8- 88.8 88.2 8288 8.2 00 82882.. 228.2 .88
82.8- 82.8- 88.8- 82.8- 8.8- 82.8- 82.8- 82.8- 88.8 8.2 8288 2.2 cc 8288... 228.2 .82
82.8 22.8 82.8 82.8 88.8 82.8 88.8 88.8 88.8 88.8 8288 8.2 co .2.2... 228.2 .82
82.8 82.8 8.8 28.8 88.8 82.8 82.8 82.8 88.8 82.8 8288 8.2 00 .2.2... 228.2 .82
82.8 82.8 28.8 88.8 8.8 82.8 82.8 82.8 88.8 8.8 2.288 2.2 ow .2.2... 228.2 .82
82.8 82.8 88.8 88.8 88.8 88.8 82.8 82.8 88.8 82.8 8288 8.2 cc 8.2828882 .82
82.8 82.8 88.8 88.8 8.8 82.8 8.8 82.8 88.8 88.8 8288 8.2 cc 8282882 .82
82.8 82.8 88.8 88.8 88.8 82.8 8.8 82.8 88.8 88.8 8288 2.2 ow 828.8882 .82
82.8 82.8 82.8 88.8 88.8 82.8 82.8 82.8 88.8 82.8 8288 8.2 828880 .82
82.8 82.8 88.8 88.8 88.8 82.8 82.8 82.8 88.8 88.8 8288 8.2 8288280 .2 2
88.8 82.8 28.8 88.8 88.8 82.8 82.8 82.8 28.8 82.8 8288 2.2 888280 .82
88.8 88.8 82.8 88.8 28.8 8.8 82.8 28.8 88.8 88.8 8288 8.2 8828882 .2280 2282888 .8
2 88.8 82.8 88.8 88.8 82.8 82.8 88.8 28.8 88.8 8288 8.2 8828882 .225 2282888 .8
2 82.8 88.8 88.8 8.8 82.8 88.8 88.8 88.8 8288 2.2 8828882 .286 2282888 .8
2 88.8 88.8 88.8 88.8 88.8 88.8 88.8 8288 8.2 828.2 .8
2 88.8 88.8 88.8 88.8 88.8 88.8 2.288 8.2 828.2 .8
2 88.8 88.8 88.8 88.8 88.8 8288 2.2 828.2 .8
2 88.8 8.8 88.8 88.8 8288 8.2 23852 .8
2 88.8 8.8 8 8288 8.2 23852 .8
2 88.8 82.8 8288 2.2 23852 .2

8 8 8 8 8 8 8 2 .828 88822 z

 

20228222 28:22282w220-2 222 :2 802828228> 28208820 222 mecca 82822282228285 .2 2228-8

169

 

 

 

 

88.8 88.8 88.8 88.8 22.8 88.8 88.8 88.8 88.8 2.8 888 8.2.288 282888 .88
2.8 88.8 88.8 88.8 2.8 82.8 88.8 88.8 28.8 88.8 888 2.2 .288 282888 .88
88.8 88.8 88.8 88.8 88.8 28.8- 88.8- 88.8- 88.8 88.8 888 8.2 .288 .8888. .88
88.8 88.8 88.8 88.8 88.8 28.8- 28.8- 88.8- 28.8 8.8 888 2.2 .288 .8888. .88
88.8 28.8- 82.8 2.8 82.8 82.8 28.8 82.8 88.8 88.8 888 82 888882282 2288 .8888. .88
88.8- 88.8- 22.8 88.8 88.8 88.8 8 8 82.2 88.8 888 82 88822282 2288 .8888. .88
28.8- 8 82.8 88.8 82.8 82.8 28.8 88.8 88.8 82.8 888 82 888822282 2288 .8888. .88
88.8 88.8 82.8 2.8 82.8 2.8 88.8 22.8 88.8 82.8 888 22 88882-2282 2288 .8888. .88
88.8- 88.8- 88.8 88.8 88.8 88.8- 8 88.8- 88.8 88.2. 82 22 82 228.2888 82258.2 .88
88.8- 8 28.8- 88.8- 88.8- 2 2.8- 88.8- 88.8- 88.8 88.8 8822 82 228.2888 82258.2 .28
88.8- 88.8- 88.8 8 88.8 28.8- 28.8- 8 88.8 88.8 82 22 82 228.288 82288.2 .88
28.8 88.8 88.8 88.8 88.8 88.8 88.8 28.8 28.8 88.8 82 22 82 228.288 82288.2 .88
88.8- 8 88.8 88.8 88.8 28.8 88.8 88.8- 8.8 88.8 2822 22 2.8.2.288 822.2.8.2 .88
88.8 88.8 88.8 88.8 82.8 8 88.8- 28.8 88.8 88.8 8822 82 22.2 282888 .88
8 8 88.8 88.8 88.8 88.8 88.8- 888- 2 88.8 882 2 82 22.2 282888 .88
88.8 88.8 2 2.8 88.8 88.8 28.8- 28.8 8 88.8 88.8 882 2 82 22.2 282888 .88
88.8 88.8 88.8 88.8 88.8 2.8 8 88.8 28.8 88.8 8822 2222.2 282888 .88
88.8 88.8 82.8 88.8 82.8 82.8 22.8 82.8 88.8 28.8 8822 8.2 22.2 8288888. .88
88.8 88.8 82.8 88.8 82.8 82.8 88.8 22.8 88.8 88.8 882 2 8.2 22.2 828288888. .88
88.8 88.8 88.8 88.8 82.8 82.8 88.8 88.8 88.8 88.8 282 2 2.2 22.2 828288888. .28
88.8 88.8 82.8 82.8 8.8 2 2.8 2 2.8 82.8 8.8 88.8 8288 8.2 .8288. .88
88.8 8.8 82.8 82.8 8.8 2 2.8 2 2.8 82.8 88.8 88.8 8288 8.2 8828.8 .88
88.8 88.8 88.8 88.8 88.8 22.8 82.8 82.8 28.8 88.8 8288 2.2 88288. .88
88.8 88.8 88.8 88.8 88.8 22.8 88.8 88.8 88.8 88.8 2288 8.2 88.8 28.882 .88
88.8 88.8 88.8 88.8 82.8 2.8 88.8 88.8 88.8 88.8 2288 8.2 .8828 88.8-.2 .88
8 8 8 8 8 8 8 2 22.8 88882 z

 

888822880 82 82.288

170

 

 

 

 

88.8- 88.8- 88.8- 88.8- 28.8 88.8- 88.8- 28.8- 88.8 2 2.8 8888 28.8 8288282 28888282 .88
88.8- 28.8- 88.8- 8 28.8- 82.8- 28.8- 88.8- 88.8 88.2 888 82 288282222222 82288282 .88
8 28.8- 88.8- 8 88.8- 2.8- 88.8- 88.8- 28.8 8.2 888 22 >88282222222 82 2882.22 .88
88.8- 88.8- 82.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8 88.2 888 82 8828882 28 8222888228 .88
28.8 88.8- 88.8- 28.8 88.8- 2.8- 88.8- 88.8- 22.2 88.2 888 22 8828882 28 82.288888 .88
88.8- 88.8- 22.8- 88.8- 88.8- 82.8- 88.8- 22.8- 88.8 8.2 888 82 .8282 8822882828 .88
28.8- 88.8- 88.8- 88.8 88.8- 88.8- 88.8- 88.8. 888 8.2 888 22 .8282 882.2280 .88
82.8 82.8 88.8 88.8 88.8 82.8 82.8 88.8 88.8 28.8 8882 2822828828822 8820 .88
88.8 88.8 8 88.8 88.8 88.8- 88.8- 88.8- 88.8 88.8 8822 8.28888 .28
2.8 82.8 88.8 88.8 88.8 88.8 28.8 28.8 88.8 8.8 888 8.2 2828828 22282.2 .88
88.8 88.8 8 28.8 88.8 88.8- 8 88.8- 28.8 88.8 888 8.2 22.28822 222882 .88
88.8 2.8 88.8- 88.8- 88.8- 88.8 28.8 88.8 88.8 88.8 888 8.2 28.228 22.882 .88
88.8 88.8 88.8 88.8 88.8 88.8 88.8 28.8- 28.8 88.8 888 8.2 .8880 222882 .88
2.8 22.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 888 2.2 22.28882 222882 .88
2.8 88.8 82.8 82.8 8.8 88.8 88.8 82.8 88.8 8.2. 888 8.2 828280 28828 .88
88.8 88.8 82.8 82.8 82.8 88.8 88.8 2 2.8 88.8 88.8 888 8.2 .8888 28828 .88
88.8 2.8 82.8 2 2.8 82.8 82.8 88.8 82.8 88.8 88.8 888 2.2 .8288 28828 .88
28.8 28.8 88.8 28.8- 88.8- 88.8 88.8 88.8 82.8 2.8 888 8.2 .288 82.2 .88
8 28.8 88.8 88.8- 88.8- 88.8 28.8 88.8 28.8 88.8 888 2.2 .288 .822 .28

8 8 8 8 8 8 8 2 .828 8882 z

 

888822880 82 82.288

171

 

 

 

88.8- 88.8 88.8 88.8 88.8 88.8 88.8 8.8 8.8 82.8 88.8 8.2 22.2 82.888888. .88
88.8- 88.8 88.8 88.8 88.8 88.8 88.8 82.8 28.8 88.8 88.8 2.2 22.2 822888888. .28
82.8- 82.8 82.8 82.8 82.8 8.8 82.8 88.8 88.8 88.8 88.8 8.2 88288. .88
22.8- 82.8 82.8 82.8 82.8 82.8 82.8 88.8 88.8 88.8 88.8 8.2 88.88. .88
8.8- 82.8 88.8 28.8 88.8 88.8 88.8 88.8 82.8 82.8 .888 2.2 88288. .88
28.8- 28.8- 88.8 88.8 88.8 22.8 22.8 88.8 88.8 88.8 88.8 8.2 .8828 882882 .88
88.8- 88.8 82.8 88.8 82.8 82.8 82.8 88.8 22.8 88.8 28.8 8.2 .8828 28.882 .88
82.8- 28.8- 88.8 88.8 88.8 88.8 88.8 88.8 2.8 88.8 88.8 2.2 .8828 882882 .88
82.8- 82.8 82.8 88.8 82.8 82.8 82.8 82.8 28.8 82.8 22.8 8.2 .8880 .88
88.8- 82.8 82.8 8.8 8.8 88.8 82.8 82.8 88.8 82.8 22.8 8.2 .8880 .88
82.8- 88.8 82.8 8.8 82.8 82.8 82.8 82.8 88.8 82.8 88.8 2.2 .8880 .88
8.8 88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 82.8- 82.8- 82.8- 82.8- 8.2 022 82888. 228.2 .28
88.8 88.8- 28.8- 88.8- 28.8- 88.8- 88.8- 2 2.8- 82.8- 82.8- 82.8- 8.2 00 82888. 228.2 .88
2 82.8- 88.8- 8.8- 88.8- 28.8- 28.8- 82.8- 88.8- 82.8- 88.8- 2.2 00 82888. 228.2 .82
2 88.8 88.8 28.8 88.8 88.8 82.8 82.8 82.8 2.8 8.2 00 .888. 228.2 .82
2 28.8 8.8 88.8 88.8 82.8 88.8 82.8 82.8 8.2 oo 8.28. 228.2 .82
2 88.8 28.8 28.8 82.8 88.8 88.8 82.8 2.2 082 8.28. 228.2 .82
2 88.8 88.8 82.8 88.8 82.8 82.8 8.2 00 8.2828882 .82
2 88.8 28.8 88.8 8.8 82.8 8.2 00 888288282 .82
2 28.8 88.8 8.8 2.8 2.2 00 828288822 .82
2 8.8 88.8 88.8 8.2 8288280 .82
2 88.8 88.8 8.2 82882822 .2 2
2 82.8 2.2 82888882 .82
2 8.2 8828882 .2280 2282888 .8

82 82 82 82 82 82 82 82 22 82 8

 

888888882 82 82.288

172

 

 

28.8- 8 88.8 28.8 28.8 88.8 88.8 88.8 88.8 88.8 88.8 8.2 228880 222882 .88
88.8- 88.8 88.8 88.8 88.8 88.8 88.8 88.8 22.8 88.8 88.8 2.2 82.88 .222882 .88
88.8- 88.8 88.8 22.8 88.8 82.8 2.8 82.8 82.8 82.8 28.8- 8.2 .8880 28828 .88
88.8- 82.8 82.8 22.8 88.8 2.8 22.8 82.8 82.8 88.8 88.8 8.2 828880 28822 .88
88.8- 88.8 88.8 28.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 2.2 .8828 28822 .88
88.8 88.8- 8 88.8- 88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 88.8- 8.2 .288 .82-2 .88
88.8 88.8- 8 88.8- 88.8- 28.8- 88.8- 8 88.8- 88.8- 88.8- 2.2 .288 .82-2 .28
88.8- 8 88.8 88.8 28.8- 88.8 88.8 82.8 88.8 88.8 88.8 8.2 .288 282888 .88
88.8- 28.8- 88.8 8 28. 88.8 88.8 88.8 88.8 8 88.8 2.2 .288 282888 .88
88.8- 88.8 88.8 2.8 88.8 88.8 88.8 88.8 2.8 88.8 88.8 8.2 .288 .8888. .88
88.8- 88.8 88.8 88.8 88.8 88.8 2.8 88.8 22.8 88.8 88.8 2.2 .288 888.8 .88
2.2.8- 888 2 2.8 88.8 82.8 82.8 82.8 82.8 82.8 88.8 88.8- 2.2 888822282 2288 .8888. .88
82.8- 88.8- . 88.8 88.8 88.8 2.8 82.8 88.8 82.8 88.8 88.8- 82 888822282 2288 .8888. .88
82 .8- 88.8 2.8 88.8 88.8 2 2.8 82.8 88.8 2 2.8 88.8 88.8- 82 888822282 2288 .8888. .88
82.8- 8 88.8 88.8 88.8 82.8 82.8 88.8 82.8 88.8 8 22 888822282 2288 .8888. .88
28.8- 88.8 88.8 88.8 8 88.8- 28.8 88.8- 8 88.8- 88.8- 82 228.288 82288.2 .88
28.8 88.8- 28.8- 8 88.8- 88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 82 2.8.2.288 8228282 .22.
88.8- 28.8 88.8 88.8 28.8 88.8 88.8 28.8 88.8 88.8- 88.8- 82 228.2888 82288.2 .88
88.8- 88.8 88.8 88.8 88.8 88.8 88.8 8 88.8 28.8 88.8 82 28.2.2.8 82288.2 .88
88.8- 88.8 88.8 8 28.8 88.8 88.8 88.8- 28.8 88.8- 28.8 22 2.8.2.288 82288.2 .88
88.8- 88.8 2 2.8 2.8 88.8 2 2.8 88.8 88.8 2.8 88.8 88.8 82 22.2 282888 .88
88.8- 28.8- 88.8 88.8 28.8- 88.8 88.8 8 88.8 88.8 88.8 82 22.2 282888 .88
88.8- 88.8 88.8 2.8 88.8 88.8 88.8 2.8 82.8 28.8 88.8 82 22.2 282888 .88
88.8- 88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 88.8 22 22.2 282888 .88
88.8- 88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 88.8 8.2 22.2 82.888888. .88
82 82 82 82 82 82 82 82 22 82 8

 

88882288 82 82.282-

 

173

 

 

 

 

88.8- 28.8 8 88.8 8 28.8- 8 28.8- 28.8- 28.8 88.8- 2218 8.88282 288.285 .88
88.8 2.8- 88.8- 88.8 88.8- 88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 82 >8882222.22 82 288.82 .88
88.8 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 2.88 88.8- 22 88828.2282, 8. 288282 .88
88.8 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8 88.8- 82 8.228882 28 822288228 .88
88.8 88.8- 88.8- 88.8- 88.8- 28.8- 88.8- 88.8- 88.8- 88.8 88.8- 22 8.228882 28 82.288888 .88
88.8 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 28.8- 88.8 88.8- 82 .8282 88228222822 .88
28.8 88.8- 88.8- 28.8- 88.8- 28.8- 28.8- 88.8- 88.8- 88.8 88.8- 22 .8282 882.8822 .88
88.8 28.8 88.8- 28.8- 88.8- 88.8- 88.8- 2.8- 82.8- 88.8- 82.8 282282882822 88222 .88
88.8- 88.8 88.8 88.8 8 88.8 88.8 88.8 88.8 88.8 88.8 88288.28 .28
82.8- 88.8 2.8 2.8 88.8 82.8 82.8 82.8 82.8 2 2.8 88.8 8.2 8282822 222882 .88
88.8- 8 88.8 88.8 88.8 88.8 88.8 88.8 2 2.8 88.8 28.8 8.2 .8680 222882 .88
88.8 28.8 8 88.8 88.8 88.8- 88.8- 88.8 8 88.8- 88.8 8.2 2828822 222882 .88
82 82 82 82 82 82 82 82 22 82 8

 

88882288 82 82.288

174

 

 

88.8 2.8 82.8 82.8 82.8 88.8 82.8 88.8 88.8- 2 2.8- 22 888882-282 2288 .8888. .88
88.8 88.8 82.8 82.8 88.8 88.8 88.8 88.8 88.8 8 82 228.2888 822.2282 .88
88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 28.8 88.8 82 28.2888 82228882 .28
88.8 2.8 82.8 82.8 2.8 88.8 88.8 88.8 8 2.8.8- 82 88.2888 8222.82 .88
22.8 82.8 88.8 88.8 88.8 88.8 82.8 88.8 2.8- 88.8- 82 228.2888 82288.2 .88
2.8 82.8 2.8 82.8 88.8 22.8 22.8 88.8 88.8- 88.8- 22 2288.288 82258.2 .88
82.8 82.8 82.8 82.8 82.8 2.8 88.8 88.8 88.8- 88.8- 82 22.2 282888 .88
88.8 88.8 82.8 82.8 82.8 88.8 28.8- 88.8- 88.8- 88.8- 82 22.2 282888 .88
2.8 82.8 28.8 88.8 82.8 88.8 88.8 88.8 88.8- 88.8- 82 22.2 282888 .88
88.8 82.8 22.8 82.8 82.8 88.8 28.8 88.8- 88.8- 88.8- 2222.2 282888 .88
88.8 22.8 88.8 2.8 88.8 8.8 88.8 22.8 88.8- 2.8- 8.2 22.2 828288888. .88
88.8 82.8 88.8 22.8 22.8 82.8 88.8 2.8 88.8- 88.8- 8.2 22.2 828288888. .88
88.8 22.8 82.8 82.8 82.8 82.8 88.8 88.8 88.8. 88.8- 2.2 .2.2 822888888. .28
88.8 8.8 88.8 82.8 88.8 82.8 8.8 88.8 88.8- 82.8- 8.2 88.88. .88
2 88.8 8 2.8 88.8 82.8 88.8 88.8 82.8- 82.8- 8.2 88288. .88
2 22.8 88.8 82.8 28.8 88.8 88.8 88.8- 88.8- 2.2 88288. .88
2 28.8 28.8 88.8- 82.8 88.8 2.8- 82.8- 8.2 .8828 28.882 .88
2 88.8 88.8 88.8 82.8 82.8- 28.8- 8.2 ..2828 288282 .88
2 88.8 88.8 82.8 82.8- 82.8- 2.2 .8828 288282 .88
2 88.8 28.8 82.8- 82.8- 8.2 88822 .88
2 88.8 88.8- 88.8- 8.2 .8880 .88
2 82.8- 82.8- 2.2 .8880 .88
2 88.8 8.2 8228 82888. 228.2 .28
2 8.2 00 82888. 228.2 .88

88 88 88 88 88 88 88 88 28 88

 

888822880 82 82.288

175

 

 

28.8 88.8- 82.8- 8.8- 88.8- 88.8- 88.8- 88.8- 28.8- 88.8 22 88888-2 28 82.288228 .88
88.8- 82.8- 88.8- 82.8- 88.8- 82.8.. 22.8- 2.8- 88.8 88.8 82 .8282 8888222882 .88
88.8 88.8- 22.8.. 82.8- 88.8- 88.8- 28.8- 88.8- 28.8- 88.8 22 .8282 882.2280 .88
88.8 88.8 88.8 28.8- 28.8- 82.8 88.8 28.8- 28.8- 88.8 2822828828822 8.282 .88
88.8 8 88.8- 2.8- 2.8- 28.8- 28.8- 28.8 28.8- 88.8- 8.288228 .28
28.8- 88.8 88.8 88.8 88.8 8 8 88.8 8 88.8- 8.2 82.8822 222882 .88

8 88.8 88.8 28.8 88.8 88.8 28.8 88.8 88.8 88.8 8.2 82.8882 222882 .88
28.8 88.8 88.8- 88.8- 88.8- 88.8 88.8- 28.8- 88.8 88.8 8.2 25880 222882 .88
88.8 88.8 88.8 88.8 88.8 88.8 88.8 8 88.8 88.8- 8.2 82.88 222882 .88
88.8 2.8 88.8 22.8 2.8 88.8 88.8 88.8 8 88.8- 2.2 8:880 222882 .88
82.8 82.8 28.8 2.8 88.8 82.8 82.8 82.8 2 2.8- 88.8- 8.2 .2588 28822 .88
82.8 82.8 8 88.8 88.8 82.8 2.8 88.8 2 2.8- 2.8- 8.2 82.8822 2880 .88
82.8 82.8 88.8 2.8 88.8 82.8 2 2.8 88.8 2.8- 88.8- 2.2 288.80 28882 .88
88.8- 88.8 82.8 8.8 82.8 88.8 88.8 88.8 28.8 88.8 8.2 .288 82.2 .88
88.8- 28.8 82.8 82.8 82.8 88.8 88.8 28.8- 28.8 88.8 2.2 .288 82.2 .28
88.8 82.8 88.8 82.8 82.8 88.8 88.8 88.8 28.8 28.8- 8.2 .288 282888 .88
88.8 88.8 88.8 88.8 2.8 88.8 88.8 88.8 88.8- 28.8- 2.2 .288 282888 .88
88.8 82.8 82.8 8.8 82.8 88.8 88.8 88.8 88.8- 88.8- 8.2 .288 .8888. .88
88.8 82.8 82.8 82.8 82.8 88.8 8 8 88.8- 2.8- 2.2 .288 .8888. .88
88.8 88.8 82.8 82.8 88.8 2 2.8 82.8 88.8 82.8- 82.8- 82 888882282 2288 .8888. .88
88.8 88.8 82.8 828 2.28 88.8 22.8 88.8 88.8- 22.8- 82 888822282 2288 .8888. .88
88.8 88.8 2 2.8 82.8 2.8 88.8 2 2.8 88.8 2 2.8- 82.8- 82 888882222 2288 .8888. .88

88 88 88 88 88 88 88 88 28 88

 

88882288 82 82.288

176

 

 

 

8 28-8 8288282 28888282 .88

 

2.22.22 22 222.22- 222.22- 222.22- 222.22- 86 222.22 222.22-

36. N222- 2.22- 2. 2 .22- 2.22- 282 .22- «22.22- 3.0- 22.22 @2222 N2 382282223 9 28202222 .wo

222.22 wo-o- N222- 52.22- 522.22- 2.22- 222.22- 36. M2222 2.22 22 3822222223 02 2828222 .3

22.22- N222- 022.22- m 2 .22- mod- m 2.22- N222- 2.22- 282222 822.22 N2 magma-2 .20 8223:922- .8
am mm mm on mm 8N mm mm 2m 2

 

8888888 82 82.288

177

 

28.8 82.8 88.8 88.8 88.8 88.8 82.8 82.8 82.8 82.8 2.2 .8828 2880 .88
82.8 82.8 88.8 2.8 82.8 2.8 82.8 22.8 82.8 88.8- 8.2 .288 82.2 .88
82.8 82.8 88.8 88.8 82.8 88.8 82.8 22.8 82.8 88.8- 2.2 .288 82.2 .28
28.8 88.8 82.8 8.8 88.8 88.8 88.8 88.8 88.8 88.8 8.2 .288 282888 .88
82.8 82.8 88.8 88.8 82.8 8.8 82.8 88.8 88.8 88.8 2.2 .288 282888 .88
82.8 82.8 88.8 8.8 88.8 88.8 8.8 82.8 88.8 82.8 8.2 .288 .8888. .88
82.8 82.8 88.8 88.8 88.8 8.8 88.8 8.8 88.8 2.8 2.2 .288 .8888. .88
82.8 82.8 88.8 88.8 82.8 22.8 8.8 82.8 82.8 88.8 82 88822282 2288 .8888. .88
88.8 82.8 2.8 88.8 88.8 22.8 88.8 8.8 82.8 88.8 82 8882-2282 2288 .8888. .88
82.8 82.8 88.8 22.8 82.8 82.8 28.8 82.8 82.8 88.8 82 888822282 2288 .8888. .88
82.8 82.8 88.8 22.8 82.8 82.8 88.8 82.8 82.8 88.8 22 888822282 2288 .8888. .88
88.8 88.8 82.8 82.8 88.8 82.8 82.8 22.8 82.8 88.8 82 28.2888 822288.2 .88
28.8 88.8 88.8 88.8 82.8 88.8 82.8 88.8 82.8 28.8 82 228.288 822.282 .28
88.8 88.8 82.8 82.8 82.8 82.8 82.8 82.8 8.8 88.8 82 2.8.2.288 822288.2 .88
2 88.8 88.8 88.8 88.8 82.8 22.8 82.8 82.8 2 2.8 82 2.8.2.288 822288.2 .88
2 82.8 82.8 8.8 82.8 82.8 82.8 28.8 22.8 22 28.2888 822288.2 .88
2 88.8 88.8 88.8 88.8 82.8 88.8 8.8 82 22.2 282888 .88
2 88.8 88.8 8.8 82.8 88.8 82.8 82 22.2 282888 .88
2 88.8 88.8 88.8 88.8 82.8 82 22.2 282888 .88
2 88.8 88.8 88.8 88.8 2222.2 282888 .88
2 88.8 88.8 88.8 8.2 22.2 822888888. .88
2 28.8 88.8 8.2 22.2 822888888. .88
2 88.8 2.2 22.2 828888888. .28
2 8.2 .8828... .88

88 88 88 88 88 88 88 88 28 88

 

88882288 82 82.288

178

 

 

 

88.8- 82.8- 88.8- 2.8- 2 2.8- 2.8- 88.8- 88.8- 88.8- 28.8 2218 82888.2 28888282 .88
88.8- 82.8- 22.8- 2.8- 2.2.8- 82.8- 82.8- 82.8- 82.8- 82.8- 82 288282222822 82 288282 .88
82.8- 82.8- 82.8- 28.8- 88.8- 88.8- 8.8- 82.8- 88.8- 88.8- 22 >8828222288 82 288282 .88
88.8- 8.8- 2.2.8- 888. 82.8- 82.8- 82.8- 8.8- 28.8- 82.8- 82 8828882 28 822288888 .88
82.8- 88.8- 8.8- 88.8- 88.8- 28.8- 88.8- 88.8- 88.8- 88.8- 22 82.28882 28 82.288888 .88
88.8- 82.8- 88.8- 88.8- 82.8- 82.8- 82.8- 82.8- 82.8- 82.8- 82 .8282 8822882882 .88
82.8- 82.8- 82.8- 88.8- 88.8- 88.8- 28.8- 82.8- 88.8- 88.8- 22 .8282 882.2280 .88
88.8 88.8 88.8- 88.8- 88.8- 88.8 82 .8 88.8 2 2.8 88.8 2822828828222 8.282 .88
88.8- 88.8- 88.8- 88.8- 2.8- 82.8. 88.8- 88.8- 88.8- 8 88288228 .28
88.8 88.8 82.8 82.8 8.8 82.8 88.8 2.8 82.8 88.8 8.2 88.8882 222882 .88
88.8 82.8 88.8 88.8 88.8 88.8 88.8 88.8 82.8 28.8 2..2 .8828 22282.2 .88
2 2.8 88.8 2 2.8 2.8 82.8 82.8 88.8 88.8 88.8 88.8 8.2 88880 222882 .88
88.8 28.8 88.8 88.8 88.8 88.8 82.8 82.8 28.8 88.8 8.2 282880 222882 .88
88.8 88.8 88.8 8.8 88.8 8.8 82.8 82.8 88.8 88.8 2.2 282880 222882 .88
8.8 82.8 88.8 82.8 82.8 82.8 82.8 82.8 8.8 82.8 8.2 .5288 2880 .88
88.8 82.8 2.8 82.8 82.8 82.8 88.8 82.8 8.8 82.8 8.2 828280 2880 .88
88 88 88 88 88 88 88 88 28 88

 

888822880 82 82.288

179

 

 

 

88.8- 88.8- 88.8- 82.8- 88.8- 82.8- 82.8- 88.8- 2.8- 22.8- 22 .8282 882.2280 .88
82.8 88.8- 88.8- 28.8 82.8 28.8 88.8 88.8 2.8 88.8 2822888828822 <20 .88
88.8- 2.8- 2.8- 88.8- 88.8- 88.8- 88.8- 88.8- 82.8- 2 2.8- 8.288228 .28
82 .8 82.8 8.8 88.8 88.8 88.8 88.8 28.8- 88.8 2 2.8 8.2 25880 222882 .88
88.8 88.8 88.8 88.8- 88.8- 88.8- 28.8- 88.8 28.8 82.8 8.2 .8880 222882 .88
82.8 82.8 82.8 88.8 2.8- 88.8 88.8 8 28.8 28.8 8.2 8.28822 .222882 .88
8.8 88.8 88.8 88.8 88.8 2.8 2.8 82.8 88.8 82.8 8.2 8.8828 222882 .88
88.8 28.8 28.8 82.8 88.8 82.8 82.8 82.8 82.8 28.8 2.2 .8880 22282.2 .88
88.8 82.8 82.8 88.8 82.8 8.8 88.8 82.8 82.8 82.8 8.2 8.2888 2880 .88
28.8 82.8 82.8 88.8 82.8 88.8 88.8 82.8 82.8 82.8 8.2 8:880 28822 .88
88.8 2 2.8 2 2.8 88.8 88.8 88.8 88.8 82.8 88.8 88.8 2.2 .8580 28882 .88
82.8 88.8 22.8 82.8 28.8 88.8 88.8 88.8 8.8 82.8 8.2 .288 82.2 .88
82.8 82.8 82.8 82.8 88.8 88.8 88.8 88.8 28.8 8.8 2.2 .288 .82-.2 .28
88.8 88.8 28.8 28.8 88.8 88.8 88.8 28.8 8.8 88.8 8.2 .288 282888 .88
2 82.8 8.8 88.8 82.8 88.8 28.8 82.8 82.8 82.8 2.2 .288 282888 .88
2 88.8 82.8 88.8 82.8 82.8 82.8 88.8 82.8 8.2 .288 .8888. .88
2 82.8 82.8 82.8 8.8 82.8 88.8 28.8 2.2 .288 .8882. .88
2 88.8 88.8 88.8 88.8 22.8 82.8 82 888822282 2288 .8888. .88
2 28.8 88.8 82.8 22.8 82.8 82 888882282 2288 .8888. .88
2 88.8 88.8 22.8 2.8 82 888822282 2288 .8888. .88
2 2.8 82.8 82.8 22 88888228 2288 .8888. .88
2 88.8 88.8 82 2288.288 822.2.8.2 .88
2 2.8.8 82 228.2888 822.2282 .22.
2 82 888.288 82288.2 .88

88 88 88 88 88 88 88 88 28 88

 

88882288 82 82.288

180

 

 

 

 

 

2 2.8- 82 .8- 82.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 88.8- 2218 8288282 288.285 .88

88.8- 82.8- 82.8- 82.8- 88.8- 82.8- 82.8.. 82.8- 2.8- 88.8- 82 38.8.2223 82 288282 .88

8.8- 88.8- 88.8- 2.8- 88.8- 22.8- 82.8- 82.8- 88.8- 2 2.8- 22 >8828822.22 82 288282 .88

8.8- 82.8- 82.8- 82.8- 2 2.8- 82.8- 82.8- 2 2.8- 82.8- 82.8- 82 8828882 28 882888.28 .88

88.8- 88.8- 88.8- 82.8- 88.8- 82.8- 8.8- 82.8- 88.8- 82.8- 22 8828882 28 8222888228 .88
8.8- 82.8- 88.8- 82.8- 88.8- 82.8- 82.8- 88.8- 22.8- 2.8- 82 .8282 82.2.8580 .88
88 88 88 88 88 88 88 88 28 88

 

88882288 82 82.282-

181

 

 

 

 

2.22. 02.22- mmd- 3.22- mod- 02222- No.22- 2.22- mod- w2.22- 22.-N mac-r2 2220220222 .00
22 2.22.22- m 2.22- w 2 .22- wmd- and- omd- No. 02-22- 2.22.22- 2 38222-2223 02 22282.22 .3
02.22- N.22- de- 582.22- 26. No. 02-22- 2.22- 2222.22- wmd- 22 38.-22222223 02 22202222 .20
2822.22- wod- m2.22- 02.22- hN-o- m-22- end- 2m.22- 2.022- wmd- N2 $222802 .20 82232202:- 00
2N.22- 2N.22- 2..22- 22.-22- NN.22- 02.22- N22- 5222- 02.22- mmd- 22 022.602 .20 28222020222- .8
2.22.22- 2.22.22- 20. N.22- 0N0- mmd- 0N0- 02.22- w2.22- wmd- N2 0.32 m22220222220 .280
c 2 .22- w 2 .22- 2m.22- 2.m.22- 2N6- 2.2.22- 02.22- m 2 .22- 2.2 .22- 8.2.22- 22 0.2222 w028222220 .mo
2.22.22- 22 222.22 mod mad 3.22 0022 mmd 2.N.22 222.22 2.222828228222222 22.820 .mw
N2222- 222.22- mod- m22.22- 222.22- mod- 2.22.22- 20. 22.22- 022-22- 82002.5 .20
22.-22 2.22 and 022-22 022.22 2222.22 22 22 222.22.. 2.2.22 mm 88222200 2228282 .220
2 2822.22 2828.22 2..22 022.22 022.22 522.22- 80. 022.22- 02.22 2.82 .8800 2228222 .3
2 2N.22 0N0 52.22 N222 mod 2.22.22- 22.22- w2-22 mm 8800 22282.22 .wm
2 00.22 w2.22 N22 2.22 022.22 3.22 mmd N82 8800 222802 Km
2 2.8-.22 3.22 w2.22 8.22 mod $.22 2.2 .8800 2228222 0m
2 3.22 2m.22 N222 22.22 2.2-:22 8.282 .8800 2000 .mm
2 28nd 2.22 282.22 $.22 N82 .8800 22800 .28m
2 .222 m 2.22 vmd 2.2 .8800 228020 .mm
2 222-.22 2N.22 N82 Jam .2222 .Nm
2 2.22 2.2.28m .02-.2 .2m
2 N82 .228m 228200m .22m

mm mm 8m 0m mm 8m mm mm 2m 22m

 

888822880 82 82.288

182

 

 

 

 

2 8.22 223 22.22 8.22 2.22 2.22 223- 8.22 8.22- 22 mambsaofido
2 ~22 8.22 2.22 one 2..2. N22- 2.22.22 222.22- N23522:? 922.25%
2 ~22 who 92.22 :2 .222. 8.22 2.22- 22382222223 92825.3
2 22.22 2.2.22 8.22 8.22- 3.22 222.22 N2 22:38.2 20 afiéfi .8
2 3.22 222.2. 2.22- 2222.22 2.22. 22 22223.2 22. mamas: .8
2 3.22 N22- 222.22 8.22 2 5.2222 222222222220 3
2 2.22- 2222.22 3.22- 22 .8222 32222.0 .8
2 8.22- 8.22- 2023222252 22220 .8
2 No.22 928% .G
2 E 8:50 22282 .8
$ 8 S 8 8 we 8 8 2o 8

 

228222280 2 2222.2

183

 

222222 22 x x 22 25222222222228 -- .2222 228222
222222 22 x x x mm .22. -- .2222 22822
2222222222 2222222222 22 x 22 2222228222223 -- .2222 228.22.
m 220.35 -- N 20.22%
222222 22 x x x .2222 .23 -- 22225 2222-22
22222 x x x x .2222 8m -- 2.2222222222282922
02.572 .2222 .oom -- mmocmflomwcomomcomv
22072 .2222 .00m - 8082520282 .2200\dom
22222 x x x x .2222 228.22 -- 22222.2 222222
222222 22 x x x .2222 222222.22. -- 22222222222823.
2222220 2.2002 22228222 X 22 .82 222202222 - 80080280282200
2223282222 x x .2222 228.222 - 82222205222022
222222 22 x x x .2222 228222 -- 228220
2222222222 2222222222 22 x x .2222 22822. -- 2222222222222222
222222282222 22 x .2222 228.22. -- 00 .2222
2222222 2.282 2222222222 22 x .2222 228222 -- 00 2222.22 222222
2222222222 222222222 22 x x .2222 228222 -- 00 22292.22 222222
N mofim -- 2 M20223
2.2022026 Sum 8022 800 .2.200

 

222222222222 2222222222 2222222223 222222222222.

22222-22

225

184

 

2222222225 222222222 222 29222222222m 22 22222222

 

 

 

 

222222 22 x x x 222222.28 22232222222223 -- 222222220
22222 x x x 22 22222222220
2232222222223 -- 22222222222222m
222222 22 x x 22 22222222220 2232222222223 -- <20
2. 20.2.5 -- m mofim
2222222 .2282 2222222222 22 225229222 -- 2,2282%
2222222222 2222222222 22 x 22 2222222222228 -- 222822m
2222222222 2222222222 22 x x 22222222222222-2222223 2222222222
22222 x x x 22 25222222222228 -- 2222222226 222222222
222222 22 x x 22 28222222222228 - .2222 .822

222222 22 x x 22

220222225 8022 80m .800 .2200

2222222223 22222222222. 222222222222 222222222222

222mm

222mm

 

2222222222280 22 02222222

185

Table 14. Standardized Regression Coefficients in the Modified Early Attrition Model.

 

 

STAGE 1 -- STAGE 2 r b
Perf. Avoid GO -- Acad. Int. -.09 -.11
Perf. App GO -— Acad. Int. .09 .12
Mast. GO -- Acad. Int. .10 -.30**
Knowledge -- Acad. Int. .16 .23M
Careers -- Acad. Int. .30 .41 **
Perseverance -- Acad. Int. .17 -.19**
Conscientiousness -- Acad. Int. .15 -.Ol
Agreeableness -- Acad. Int. .12 .16"
Emot. Stab. -- Acad. Int. .15 .12"
Agreeableness -- Soc. Int. .17 .23"
Emot. Stab. -- Soc. Int. .23 .27"
Perf. Avoid GO -- Acad. SE -.35 -.l8**
Perf. App GO -- Acad. SE .25 .05
Mast. GO -- Acad. SE .30 -.04
Knowledge -- Acad. SE .23 .02
Careers -- Acad. SE .27 .15"
Perseverance -- Acad. SE .36 .21"
Conscientiousness -- Acad. SE .27 .04
Agreeableness -- Acad. SE .16 -.Ol
Emot. Stab. -- Acad. SE .27 .16**
STAGE 2 -- STAGE 3

Acad. Int. -- Satisfaction .44 .02
Acad. Int. -- Commitment .43 .22"
Soc. Int. -- Satisfaction .62 .88"
Soc. Int. -- Commitment .53 .58"
Acad. SE -- GPA -.O4 -.O4*
Family Support -- Commitment .28 .13"
Family Support--Satisfaction .31 .02
Shocks -- Satisfaction -.17 -.O6**
STAGE 3 -- STAGE 4

GPA -- Withdrawal Cog./Beh. -.21 -.24**
Satisfaction -- Withdrawal

Cog./Beh. -.48 -.O9**
Comm. -- Withdrawal Cog./Beh. -.59 -.62**
STAGE 4 -- STAGE 5

Withdrawal Cog./Beh. --

Dropout -.O3 -.06*

 

*pS.05.**pS.01.

I86

Table 15. Squared Multiple Correlations for the Revised Early Attrition Model.

 

Early
Dependent Attrition
Independent Variables Variable R2

 

Performance Avoid GO,

Performance Approach GO, Mastery

GO, Knowledge, Career Orientation,

Perseverance, Conscientiousness,

Agreeableness, Emotional Stability Academic Int. .21

Conscientiousness, Agreeableness,
Emotional Stability, Social/cultural
Involvement Social Int. .13

Performance Avoid GO,
Performance Approach GO, Mastery
GO, Knowledge, Career Orientation,
Perseverance, Conscientiousness,

Agreeableness, Emotional Stability Academic SE .28
Academic SE GPA .00
Academic Int., Social Int., Family
Support Satisfaction .78
Academic Int., Social Int., Family
Support Commitment .44
Withdrawal
Cognitions/
GPA, Satisfaction, Commitment Behaviors .51
Withdrawal Cognitions/Behaviors Withdrawal .01

 

 

187

 

Table 16. Standardized Regression Coefficients in the Modified Longitudinal Model.

 

 

STAGE 1 -- STAGE 2 r b
Perf. Avoid GO ~~ Acad. Int. ~.O9 -.10
Perf. App GO -- Acad. Int. .09 .13
Mast. GO -- Acad. Int. .10 -.29*
Knowledge -- Acad. Int. .16 .25"
Careers -- Acad. Int. .30 .40"
Perseverance -- Acad. Int. .17 -.20*
Conscientiousness -- Acad. Int. .15 .00
Agreeableness -- Acad. Int. .12 .15”
Emot. Stab. -- Acad. Int. .15 .12"
Agreeableness -- Soc. Int. .17 .18"
Emot. Stab. -- Soc. Int. .23 .25"
STAGE 2 -- STAGE 3

Acad. Int. -- Satisfaction .38 .08”
Acad. Int. -- Ac. SE .26 .32"
Acad. Int. -- Commitment .29 .16"
Soc. Int. -- Satisfaction .48 .67"
Soc. In . -- Commitment .44 .56"
Family Support -- Commitment .25 .13"
Family Support--Satisfaction .27 .O7*
Acad. SE -- GPA .17 .21”
Shocks -- Satisfaction -.14 -.03
Shocks -- Dropout .09 .03
STAGE 3 -- STAGE 4

GPA -- Withdrawal CogJBeh. -.21 -.45**
Satisfaction -- Withdrawal

Cog./Beh. -.35 -.10**
Comm. -- Withdrawal Cog./Beh. -.38 -.41**
STAGE 4 -- STAGE 5

Withdrawal Cog./Beh. --

Dropout .23 .45“

 

*p S .05. **p S .01.

I88

 

 

 

 

3.223222 2222. «2. 22.2 222222 22.82222 2222222222222 222222.32 2.2222222222222222

2.2.322 2222. 2222. N222 3222 3.222222 22222222222222

22222222 222222222222 22222

8.2222222 22. 222. 3.22 222:2 8.2222222 2222222ch2 22222222 2.2222222222222222

22222222222 2222. 2222. 223 2222: 2.2222222 2222222222222 22222222 222222222222 2222222
0222 $5222 2220 22222222 220 222

 

2022032 2288-2- 0222 80.20.22. 2200222222 2222 .2.2 022222-2-

I89

 

 

(1)1

 
    
   
 

XXII

P

Kx9
@-

Self Concept

Xxl3

G £2

i

XX14
S

E

Figure 10. Full Model: Modified Model V

127

 

 

Psyc. Well Being

112

 

 

Gay Identity

£1

XX9
®
a XXII

 
   
   
 

Self Concept

5.2
Xx13

Figure I 1. Full Model: Modified Model VI

131

 
   
   

   
   

Psyc. Well Being

T12

 

 

  

.uauroAlo-uq
mlan/IBEOOS

  
 
 
  
   
 

£111!qu
[euonotua

  

ssau
~91qeaorfiv

  
   
 

ssausnoi
-iuaiosuo;)

  

GOURJGAQSJGd

 
   
   
 

uopmuatro
19912:)

aSpolmouX
oo £19182»:

oo
'ddv 319-1
09
now 319d

11110:) z amfir :1

 

 

 

° C?

 

 

91 SI

 

 

 

 

171

 

 

 

 

lemmpmim

 

 

 

sroiAeqag
/suonru303
[BMW‘DIM

 

 

 

 

é \
9%

 

 

 

 

luaunimmog

 

 

 

 

 

 

 

VdD

 

 

 

Z

61

     
 

luawaAIOMII
lemma/IBM

 

 

 

snooqs

 
 

 

 

 

     
 

£1“!qu
[euonourg

  

ssau
-a[qeaarfiv .

 

 

 

 

 

uouoegsues

 

  
 
 
 

as?

 
  
 

  

ssousnot
quotasuog)

 
 

 

 

 

 

 

 

 

 

 

 

 

aoumamsrod
uoumfiaiul
IBIOOS
uopeiuapo
v1 £1 21 II ‘99193

 

 

 

 

 

ofipomoux

 
 

uouarfiaiul
orurapaov

    
 

 

'ddv “319d

 

(Del
69...

oo
mow 313a

'ponupuog z emf}; :1

 

 

 

 

a GIQ-S2 Gay Identity
Xx5
° GIQ-S3 £1

 

    
   

Self Concept

E2
XXII

0
Figure 12. Full Model: Modified Model VII

135

   
   

Psyc. Well Being

n2

 

 

”2.2.
2m

> > > > > \/ > > > \/ .2.

   
    

QAIOAUI
)/19!9°S

flotsam

3 h? E '58 1'33 '53
.— 00 o
5 5.:‘2 é? a= 55
a? 0 a D <

«o

g0
O

0..

(2

go
0

D.-

ms '9 mm

 

.33
(028

H

S
El

8

a8
<1

3

0!
0:)

C‘-

‘3 6
.g g
0 n2

8
U

5

0“")1

W

I:
a)
0..

"C'-

9d

91s '9 amfii-I

 

REFERENCES

I97

REFERENCES

ACT. (2008). 2008 retention/completion summary tables. Retrieved August 20, 2008
from http://www.act.org/research/policymakers/pdf/retain_trends.pdf.

Allen, N.J., & Meyer, J .P. (1990). The measurement and antecedents of affective,
continuance, and normative commitment to the organization. Journal of
Occupational Psychology, 63, 1-18.

Ames, C. (1992). Classrooms: Goals, structures, and student motivation. Journal of
Educational Psychology, 84, 261-271.

Astin, A.W., & Oseguera, L. (2005). Pre-college and institutional influences on degree
attainment. In A. Siedman (Ed.), College Student Retention: Formula for Student
Success. Westport, CT: Praeger.

Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and job
performance: A meta-analysis. Personnel Psychology, 44, 1-26.

Barron, K.E., & Harackiewicz, J.M. (2001). Achievement goals and optimal motivation:
Testing multiple goal models. Journal of Personality and Social Psychology, 80,
706-722.

Beach, LR. (1990). Image theory: Decision making in personal and organizational
contexts. Chichester, England: Wiley.

Bean, J .P. (1980). Dropouts and turnover: The synthesis and test of a causal model of
student attrition. Research in Higher Education, 12, 155-187.

Bean, J .P. (1981). Student attrition, intentions, and confidence: Interaction effects in a
path model: Part II, the ten variable model. Presented at the Annual Meeting of
the American Educational Research Association, Los Angeles, CA.

Bean, JP. (1985). Interaction effects based on class level in an explanatory model of

college student dropout syndrome. American Educational Research Journal, 22,
35-64.

Bean, J.P., & Bradley, R.K. (1986). Untangling the satisfaction-performance relationship
for college students. The Journal of Higher Education, 5 7, 393-412.

Bean, J. P., & Eaton, S. B. (2000). A psychological model of college student retention. In
J. M. Braxton (Ed.), Reworking the student departure puzzle (pp. 48-61).
Nashville, TN: Vanderbilt University Press.

198

Bean, J .P., & Vesper, N. (1990). Quantitative approaches to grounding theory in data:
Using LISREL to develop a local model and theory of student attrition. Presented
at the Annual Meeting of the American Educational Research Association,
Boston, MA.

Berger, J .P., & Lyon, SC. (2005). Past to present: A historical look at retention. In A.
Seidman (Ed.), College Student Retention (pp. 1-30). Westport, CT: Praeger.

Betz, E. L., Klingensmith, J. E., & Menne, J. W. (1970). The measurement and analysis
of college student satisfaction. Measurement and Evaluation in Guidance, 3(2),
110-118.

Braxton, J.M., Milem, J .F., & Sullivan, AS. (2000). The influence of active learning on

the college student departure process. The Journal of Higher Education, 71, 569-
590.

Braxton, J .M., Sullivan, A.S., & Johnson, RM. (1997). Appraising Tinto’s theory of
college student departure. In J. C. Smart (ed.) Higher Education: Handbook of
Theory and Research, Vol. 12 (pp. 107-164). New York: Agathon.

Button, S.B., & Mathieu, J .E. (1996). Goal orientation in organizational research: A
conceptual and empirical foundation. Organizational Behavior and Human
Decision Processes, 67, 26-48.

Cabrera, A.F., Nora, A., Castaneda, M.B., & Hengstler, D. (1992). The convergence
between two theories of college persistence. The Journal of Higher Education, 63 ,
143-164.

Cabrera, A.F., Nora, A., & Castaneda, MB. (1993). College persistence: Structural
equations modeling test of an integrated model of student retention. The Journal
of Higher Education, 64, 123-139.

Carter, R.L. (2006). Solutions for missing data in structural equation modeling. Research
in Practice and Assessment, 1, 1-6.

Cassady, J.C. (2001). Self-reported GPA and SAT: A methodological note. Practical
Assessment, Research & Evaluation, 7(12). Retrieved April 26, 2009 from
http://PAREonIine.net/getvn.asp?v=7&n=l 2.

Cohen, A. (1993). Organizational commitment and turnover: A meta-analysis. The
Academy of Management Journal, 36, 1 140-1 157.

Cohen, A., & F reund, A. (2005). A longitudinal analysis of the relationship between
multiple commitments and withdrawal cognitions. Scandinavian Journal of
Management, 21, 329-351 .

199

Cole, D.A., Ciesla, J .A., & Steiger, J .H. (2007). The insidious effects of failing to include
design-driven correlated residuals in latent-variable covariance structure analysis.
Psychological Methods, 12, 381-398.

Colquitt, J .A., & Simmering, M.J. (1998). Conscientiousness, goal-orientation, and
motivation to learn during the learning process: A longitudinal study. Journal of
Applied Psychology, 83, 654-665.

Cope, R., & Hannah, W. (1975). Revolving college doors: The causes and consequences
of dropping out, stopping out, and transferring. New York: Wiley.

Dowaliby, F.J., Garrison, W.M., & Dagel, D. (1993). The student integration survey:
Development of an early alert assessment and reporting system. Research in
Higher Education, 34, 513-531.

Durkheim, E. (1951). Suicide. Translated by J .A. Spaulding and G. Simpson. Glencoe:
The Free Press.

Dweck, C.S., & Leggett, EL. (1988). A social-cognitive approach to motivation and
personality. Psychological Review, 95, 256-273.

Edens, K.M. (2006). The relationship of university students’ sleep habits and academic
motivation. NASPA Journal, 43, 432-445.

Elliott, A.J., & Harackiewicz, J .M. (1996). Approach and avoidance achievement goals
and intrinsic motivation: A mediational analysis. Journal of Personality and
Social Psychology, 70, 461-475.

Elliott, E.S., & Dweck, CS. (1988). Goals: An approach to motivation and achievement.
Journal of Personality and Social Psychology, 54, 5-12.

Erez, M. (2005). Goal-setting, goal-orientation. In N. Nicholson, P. Audia, and M Pillutla
(Eds.), Blackwell Encyclopedic Dictionary of Organizational Behavior, 2’“ Ed.,
13 8-141. Oxford, UK: Blackwell.

Garson, GD. (2009). Structural equation modeling. Retrieved from
http://faculty.chass.ncsu.edu/garson/PA765/structur.htm.

Geiser, S., & Santelices, M.V. (2007). Validity of high school grades in predicting
student success beyond the freshman year: High school record vs. standardized

tests as indicators of four-year college outcomes. Center for Studies in Higher
Education, Berkeley, CA.

Gore, PA. (2006). Academic self-efficacy as a predictor of college outcomes: Two
incremental validity studies. Journal of Career Assessment, 14, 92-115.

200

Grace, H. (1957). Personality factors and college attrition. Peabody Journal of Education,
35, 36-40.

Gray, E.K., & Watson, D. (2002). General and specific traits of personality and their
relation to sleep and academic performance. Journal of Personality, 70, 177-206.

Griffeth, R.W., & Horn, P.W (1988), A comparison of different conceptualizations of
alternatives in turnover research. Journal of Organizational Behavior, 13, 103-
111.

Griffeth, R., Hom, P.W., & Gaertner, S. (2000). A meta-analysis of antecedents and
correlates of employee turnover: Update, moderator tests, and research
implications for the next millennium. Journal of Management, 26, 463-488.

Haines, P., McGrath, P., & Pirot, M. (1980). Expectations and persistence: An
experimental comparison of Bandura and Rotter. Social Behavior and
Personality, 8, 193-201.

Harackiewicz, J .M., Barron, K.E., Pintrich, P.R., Elliot, A.J., & Thrash, T.M. (2002).
Revision of achievement goal theory: Necessary and illuminating. Journal of
Educational Psychology, 94, 638-645.

Harman, W.S., Lee, T.L., Mitchell, T.R., Felps, W., & Owens, BR (2007). The

psychology of voluntary employee turnover. Current Directions in Psychological
science, 16, 51-54.

Hearn, J .C. (1984). The relative roles of academic ascribed and socioeconomic
characteristics in college destinations. Sociology of Education, 5 7, 22-30.

Heilbrun, A. R. (1965). Personality factors in college dropout. Journal of Applied
Psychology, 49, 1-7.

Hom, P.W., Caranikas-Walker, F., Prussia, G.E., & Griffith, R. (1992). A meta-analytical
structural equations analysis of a model of employee turnover. Journal of Applied
Psychology, 77, 890-909.

Horn, P.W., & Griffeth, R. (1995). Employee turnover. Cincinnati, OH: Southwestern.

Hom, P.W., & Kinicki, A.J. (2001). Toward a greater understanding of how

dissatisfaction drives employee turnover. Academy of Management Journal, 44,
975-987.

Hossler, D., & Stage, F.K. (1992). Family and high school experience influences on the
postsecondary education plans of ninth-grade students. American Educational
Research Journal, 29, 425-451.

201

Hu, L., & P. M. Bentler (1999). Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling 6(1): l-55.

Hull-Blanks, E., Robsinson-Kurpius, S.E., Befort, C., Sollenberger, S., Nicpon, M.F., &
Huser, L. (2005). Career goal and retention-related factors among college
freshmen. Journal of Career Development, 32, 16-30.

J aros, SJ. (1997). An assessment of Meyer and Allen’s three-component model of
organizational commitment and turnover intentions. Journal of Vocational
Behavior, 51, 319-337.

Kahn, J .H., & Nauta, M.M. (2001). Social-cognitive predictors of first-year college

persistence: The importance of proximal assessment. Research in Higher
Education, 42, 633-652.

Kahn, J .H., Nauta, M.M., Gailbreath, R.D., Tipps, J ., & Chartrand, J .M. (2002). The
utility of career and personality assessment in predicting academic progress.
Journal of Career Assessment, 10, 3-23.

Kalsner, L. (1991 ). Issues in college student retention. Higher education extension
service review, 3, 3-10.

Kline, RB. (2005). Principles and Practice of Structural Equation Modeling. New York,
NY: The Guilford Press.

Kristof, AL. (1996). Person-organization fit: An integrative review of its

conceptualizations, measurement, and implications. Personnel Psychology, 49, 1-
49.

Kristof-Brown, A.L., Zimmerman, R.D., & Johnson, EC. (2005). Consequences of
individuals’ fit at work: A meta-analysis of person-job, person-organization,
person-group, and person-supervisor fit. Personnel Psychology, 58, 281-342.

Kuncel, N. R., Crede, M., & Thomas, L. L. (2005). The validity of self-reported grade
point averages, class ranks, and test scores: A meta-analysis and review of the
literature. Review of Educational Research, 75, 63-82.

Landis, R.S., Beal, D.J., & Tesluk, PE. (2000). A comparison of approaches to forming

composite measures in structural equation models. Organizational Research
Methods, 3, 186-207.

Landis, R.S., Edwards, B.D., & Cortina, J .M. (2009). On the practice of allowing
correlated residuals among indicators in structural equation models. In C.E. Lance
and R.J. Vandenberg (Eds) Statistical and Methodological Myths and Urban

202

Legends: Doctrine, Verity, and Fable in the Organizational and Social Sciences (pp. 193-
218). New York, NY: Routlege.

Lee, T.W., Mitchell, T.R., Holtom, B.C., McDaniel, L.S., & Hill, J .W. (1999). The
unfolding model of voluntary turnover: A replication and extension. Academy of
Management Journal, 42, 450-462.

Lee, T. W., Mitchell, T.R., Wise, L., & Fireman, S. (1996). An unfolding model of
voluntary employee tumover. Academy of Management Journal, 39(1): 5-36.

Lei, P-W., Bassiri, D., & Schulz, EM. (2001). Alternatives to the grade point average as
a measure of academic achievement in college. ACT Research Report Series,
Iowa City.

Liang, J ., Lawrence, R.H., Bennett, J .M., & Whitelaw, NA. (1990). Appropriateness of
composites in structural equation models. Journal of Gerontology, 45, 52-59.

Liu, R. & Liu, E. (2000). Institutional Integration: An analysis of Tinto’s theory. Paper
presented at the Annual Forum of the Association for Institutional Research,
Cincinnati, OH.

Lounsbury, J .W., Saudargas, R.A., & Gibson, L.W. (2004). An investigation of
personality traits in relation to intention to withdraw from college. Journal of
College Student Development, 45, 517-534.

March, J .G. & Simon, HA. (1958). Organizations. New York: Wiley.

Mashburn, A.J. (2000/2001). A psychological process of college student dropout. Journal
of College Student Retention, 2, 173-190.

Maudal, G.R., Butcher, J .N., & Mauger, RA. (1974). A multivariate study of personality

and academic factors in college attrition Journal of Counseling Psychology, 21,
560-567.

McDonald, R.P., & Ho, M.- H. (2002). Principles and practice in reporting structural
equation analyses. Psychological Methods, 7, 64-82.

McNatt, D.B.. & Judge, T.A. (2008). Self-efficacy interventions, job attitudes, and

turnover: A field experiment with employees in role transition. Human Relations,
61, 783-810.

Metz, G.W. (2002). Challenges and changes to Tinto’s persistence theory. Paper
presented at the Annual Meeting of the Mid-Westem Educational Research
Association, Columbus, OH.

203

Meyer, J. P, Stanley, D. J, Herscovitch, L., & Topolntsky, L. (2002). Affective,
continuance, and normative commitment to the organization: A meta-analySis of
the antecedents, correlates, and consequences. Journal of Vocational Behavior,
61, 20-52.

Michigan State University. (2009). Academic standing. Retrieved March 15"“, 2010 from
https://www.msu.edu/~uud/images/undgredacst.pdf

Mobley, W. H. (1977). Intermediate linkages in the relationship between job satisfaction
and employee turnover. Journal of Applied Psychology, 62(2), 237-240.

Multon, K.D., Brown, S.D., & Lent, R.W. (1991). Relation of self-efficacy beliefs to
academic outcomes: A meta-analytic investigation. Journal of Counseling
Psychology, 38, 30-38.

Munson, H. L., & Rubenstein, B. J. (1992). School IS work: Work task learning in the
classroom. Journal of Career Development, 18, 289-297.

O*NET Online. (2003). Retrieved April 24, 2009, from 0*NET Consortium website:
http://online.onetcenter.org/link/summary/27-20l 1.00

Okun, M.A., & Finch, J .F . (1998). The big five personality dimensions and the process of
institutional departure. Contemporary Educational Psychology, 23, 233-256.

Oswald, F. L., Schmitt, N., Kim, B. H., Ramsay, L. J ., & Gillespie, M. A. (2004).
Developing a biodata measure and situational judgment inventory as predictors of
college success. Journal of Applied Psychology, 89, 187-208.

Owens, W.A., & Schoenfeldt, LP. (1979). Toward a classification of persons. Journal of
Applied Psychology Monograph, 65, 569-607.

Pandey, R. E. (1973). A comparative study of dropout at an integrated university: The 16
Personality Factor Test. Journal of Negro Education, 42, 447-451.

Pascarella, E. (1980). Student-faculty informal contact and college outcomes. Review of
Educational Research, 50, 545-595.

Pascarella, E. T., & Chapman, D. W. (1983). A multiinstitutional, path analytic validation

of Tinto’s model of college withdrawal. American Educational Research Journal,
20, 87-102.

Pascarella, E.T., & Terenzini, RT. (1980). Predicting freshman persistence and voluntary
dropout decisions from a theoretical model. The Journal of Higher Education, 51 ,
60-75.

204

Pascarella, E.T. & Terenzini, RT. (2005). How college affects students: A third decade of
research. San Francisco, CA: Jossey-Bass.

Payne, S.C., Youncourt, S.S., & Beaubien, J .M. (2007). A meta-analytic examination of
the goal orientation nomological net. Journal of Applied Psychology, 92, 128-150.

Pemberton, W.A. (1970). The grade point average: Snark or boojum? Report: ED047009.
Newark, Delaware.

Peterson, S.L. (1992). Career decision-making self-efficacy and institutional integration
of underprepared college students. Research in Higher Education, 34, 659-685.

Peterson, S.L. (2004). Toward a theoretical model of employee turnover: A human

resource development perspective. Human Resource Development Review, 3,
209-227.

Peterson, S.L. (2007). Managerial turnover in US retail organizations. Journal of
Managerial Development, 8, 770-789.

Pintrich, PR. (2000). Multiple goals, multiple pathways: The role of goal orientation in
learning and achievement. Journal of Educational Psychology, 92, 544-555.

Porath, C.L., & Bateman, TS. (2006). Self-regulation: From goal orientation to job
performance. Journal of Applied Psychology, 91, 185-192.

Ramist, L. (1981). College student attrition and retention—College board report no. 81-1.
New York: College Entrance Examination Board.

Random House Dictionary. (n.d.) Perseverance. Dictionary. com Unabridged (v 1.1).
Retrieved April 12, 2009, from

http://dictionary.reference.com/browse/perseverance.

Reddy, SK. (1992). Effects of ignoring correlated measurement error in structural
equation models. Educational and Psychological Measurement, 52, 549-570.

Reed, J .G., Lahey, M.A., & Downey, R.G. (1984). Development of the college
descriptive index — A measure of student satisfaction. Measurement and
Evaluation in Counseling and Development, 1 7, 67-82.

Richards, J .M., Jr., Holland, J .L., & Lutz, SW. (1967). Prediction of student
accomplishment in college. Journal of Educational Psychology, 58, 343-355.

Rigdon, BE. (1996). CFI versus RMSEA: A comparison of two fit indexes for structural
equation modeling. Structural Equation Modeling, 3, 369-3 79.

205

Robbins, S., Allen, J ., Casillas, A., Peterson, C.H., & Le, H. (2006). Unraveling the
differential effects of motivational and skills, social, and self-management

measures from traditional predictors of college outcomes. Journal of Educational
Psychology, 98, 598-616.

Sager, J .K., Griffith, R.W., & Horn, P.W. (1998). A comparison of structural models
representing turnover cognitions. Journal of Vocational Behavior, 53, 254-273.

Saks, AM. (1995). Longitudinal field investigation of the moderating and mediating
effects of self-efficacy on the relationship between training and newcomer
adjustment. Journal of Applied Psychology, 80, 21 1-225.

Sal gado, J. F. (2002). The big five personality dimensions and counterproductive
behaviors. International Journal of Selection and Assessment, 10, 117-125.

Sandler, ME. (2000). Career decision-making self-efiicacy, perceived stress, and an
integrated model of student persistence: A structural model of finances, attitudes,
behavior, and career development. Research in Higher Education, 41, 537-580.

Schunk, D.H., & Zimmerman, B.J. (2008). Preface. In D.H. Schunk & B.J. Zimmerman
(Eds.), Motivation and Self-Regulated Learning (pp. vii-ix). New York: Lawrence
Earlbaum Associates.

Sekiguchi, T. (2004). Person-organization fit and person-job fit in employee selection: A
review of the literature. Osaka Keidai Ronshu, 54, 179-196.

Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in

work and retirement: A strategy for the study of attitudes. Oxford, England: Rand
McNally

Spady, W. (1971). Dropouts from higher education: toward and empirical model.
Interchange 2: 38-62.

Spady, W.G. (1975). Lament for the letterman: Effects of peer status and extracurricular
activities on goals and achievement. American Journal of Sociology, 75 , 680-702.

Sprandel, H. (1986). Career planning and counseling. In L. Noel & R. Levitz (Eds.),
Increasing Student Retention (pp. 302-318). San Francisco: Jossey-Bass.

Stajkovic, A., & Luthans, F. (1998). Self-efficacy and work-related performance: A
meta-analysis. Psychological Bulletin, 124, 240-261.

Summerskill, J. (1962). Dropouts from college. In N. Sanford (Ed.), The American
College (pp. 627-657). New York: Wiley.

206

Terenzini, P.T., Pascarella, E.T., Theophilades, C., & Lorang, W.G. (1983) A path
analytic validation of Tinto’s theory. Paper presented at the Annual Conference of
the American Educational Research Association, Montreal.

Tett, R.P., & Meyer, J .P. (2006). Job satisfaction, organizational commitment, turnover

intention, and turnover: Path analyses based on meta-analytic findings. Personnel
Psychology, 46, 259-293.

Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent
research. Review of Education al Research, 45 , 89-125.

Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition
(2nd ed.). Chicago: University of Chicago Press.

Tinto, V. (2005) Foreward. In A. Seidman (Ed.), College Student Retention: Formula for
Student Success (p. ix-xi). Connecticut: Praeger.

Tinto, V. (2006). Research and practice of student retention: What next? Journal of
College Student Retention, 8, 1-19.

Tracey, T.J. & Sedlacek, W.E. (1986). Prediction of college graduation using
noncognitive variables by race. Paper presented at the Annual Meeting of the
American Educational Research Association, San Francisco, CA.

Tross, S. A., Harper, J. P., Osher, L. W., & Kneidinger, L. M. (2000). Not just the usual
case of characteristics: Using personality to predict college performance and
retention. Journal of College Student Development, 41 , 323-333.

Tuominen-Soini, H., Katariina, S-A., & Niemivirta, M. (2008). Achievement goal
orientations an subjective well-being: A person-centred analysis. Learning and
Instruction, 18, 251-266.

Verquer, M.L., Beehr, T.A., & Wagner, SH. (2003). A meta-analysis of the relations
between person-organization fit and work attitudes. Journal of Vocational
Behavior, 63, 473-489.

Vollmeyer, R., & Rheinberg, F. (2000). Does motivation affect performance via
persistence? Learning and Instruction, 10, 293 -3 09.

Wheeler, A.R., Coleman Gallagher, V., Brouer, R.L., & Sablynski, Q]. (2007). When
person-organization (mis)fit and (dis)satisfaction lead to turnover. Journal of
Managerial Psychology, 22, 203-219.

207

 

 

\\

 

 

 

Williams, L.J., & Hazer, J .T. (1986). Antecedents and consequences of satisfaction and
commitment in turnover models: A reanalysis using latent variable structural
equation methods. Journal of Applied Psychology, 71, 219-231.

Wothke, W. (2000). Longitudinal and multi-group modeling with missing data. In T.D.
Little, K.U. Schnabel, and J. Baumert (Eds.) Modeling longitudinal and multilevel
data: Practical issues, applied approaches, and specific examples. Mahwah, NJ .:
Lawrence Erlbaum Associates.

Youngblood, S.A., Mobley, W.H., & Meglino, BM. (1983). A longitudinal analysis of
the turnover process. Journal of Applied Psychology, 68, 507-516.

Zimmerman, R. D. (2008). Understanding the impact of personality traits on individuals’
turnover decisions: A meta-analytic path model. Personnel Psychology, 61, 309-
348.

Zimmerman, B.J., & Schunk, DH. (2008). Motivation: An essential dimension of self-
regulated learning. In D.H. Schunk & B.J. Zimmerman (Eds.), Motivation and
Self-Regulated Learning (pp. 1-30). New York: Lawrence Earlbaum Associates.

208

     

        

     

          

 

ATE U R ITY LIB

"1111111111, 1111111113111 ,1 ,1 1111'“