An Investigation of the Causal

Report
An Investigation of the Causal
Relationship Between Academic
Motivation and Community College
Student Success
Lijuan Zhai, Director, Institutional Research, Fresno City College
Mary Ann Valentino, Psychology Faculty, Fresno City College
Chuck Kralowec, Institutional Research Coordinator, District Grant
Office, State Center Community College District
2014 RP Group Conference (April 10-11)
Introduction
• Psychology faculty investigating possible variables that
influence success rates in general psychology class, which
historically has success rates that are lower than the
college and the division (social sciences).
• Anecdotally, instructors often cite motivation or lack
thereof as a reason for student success or failure.
Introduction
• We decided to find and use an instrument to investigate
the correlation between success and motivation with our
general psychology students.
• If motivation, and specifically certain types of motivation,
is associated with higher rates of success, would it be
possible to design interventions, at the departmentand/or at the college-level to increase motivation?
Introduction
• Motivation is defined as “a process that
arouses, maintains, and guides behavior
toward a goal” (Cacioppo & Freberg, 2013)
Cacioppo, J.T. & Freberg, L.A. (2013). Discovering Psychology:
The Science of Mind. Canada: Wadsworth, Cengage Learning.
Introduction
• Formal education is an essential prerequisite of
professional progress and it is vital to identify the
psychological factors which determine academic
achievement.
• Recent research conducted in this field
emphasizes the importance of difference
motivational constructs.
Introduction
• Motivation is related to student learning and
performance (Vallerand, Pelletier, Blais, Brière, Senécal,
& Vallières, 1992).
• In their recent article, Wagner & Szamoskozi (2012)
conducted a meta-analysis on the effects of academic
motivation and achievement training programs. They
concluded that academic motivation has a great impact
on academic performance (Wagner & Szamoskozi, 2012).
It contributes to the prediction of school achievement
above and beyond intelligence.
Academic Motivation
• Academic motivation has been defined in various ways:
• “a student’s willingness, need, desire and compulsion to
participate in, and be successful in the learning process”
(Moenikia & Babelan, 2010, p. 1538).
Academic Motivation
• Several conceptual perspectives have been proposed in order
to better understand academic motivation, including intrinsic
motivation, extrinsic motivation, and amotivation (Vallerand,
et al.,1992). These three types of motivations are defined as
follows:
• Intrinsic Motivation (IM) refers to the fact of doing an activity
for itself, and the pleasure and satisfaction derived from
participation (Deci, 1975; Deci & Ryan, 1985, as cited in
Vallerand, et al.,1992)
• Extrinsic Motivation (EM) pertains to a wide variety of
behaviors which are engaged in as a means to an end and not
for their own sake (Deci, 1975, cited in Vallerand, et al.,1992).
Academic Motivation
• Amotivation is neither intrinsic nor extrinsic. When
amotivated individuals experience feelings of incompetence
and expectancies of uncontrollability, they perceive their
behaviors as caused by forces out of their own control.
Eventually, they may stop participating in academic activities
(Vallerand, et al.,1992).
•
Survey Instrument
• To measure student motivation, Academic Motivation
Scale (AMS) was used. AMS was developed and
validated by a group of scholars in Canada (Vallerand, et
al.,1992). The survey instrument was widely used and
validated by many researchers.
• Academic Motivation Scale (AMS) was developed based
on the self-determination theory (SDT) proposed by Deci
and Ryan (1985).
• Deci and Ryan basically identify “several distinct types of
motivation” (Ryan & Deci, 2000, p. 69). These types of
motivation root in the perceived locus of causality, which
can be internal, external or impersonal (see figure 1):
Ryan & Deci (2000) Self Determination
Theory
Survey Instrument
AMS contains 28 survey questions and addresses the following
seven aspects of academic motivation:
1. Intrinsic motivation - to know (questions # 2, 9, 16, 23)
Intrinsic motivation to know is defined as the fact of
performing an activity for pleasure and the satisfaction that
one experiences while learning, exploring, or trying to
understand something new (Vallerand, et al.,1992)..
Survey Instrument
2. Intrinsic motivation - toward accomplishment (questions #
6, 13, 20, 27)
Intrinsic motivation toward accomplishments is defined as
the fact of engaging in an activity for the pleasure and
satisfaction experienced when one attempts to accomplish
or create something (Vallerand, et al.,1992).
3. Intrinsic motivation - to experience stimulation (questions
# 4, 11, 18, 25)
Intrinsic motivation to experience stimulation is operative
when someone engages in an activity in order to experience
stimulating sensation (e.g., sensory pleasure, aesthetic
experiences, as well as fund and excitement) derived from
one’s engagement in the activity (Vallerand, et al.,1992).
Survey Instrument
4.
5.
Extrinsic motivation – identified (questions # 3, 10, 17, 24)
To the extent that the behavior becomes valued and judged
important for the individual, and especially that it is perceived as
chosen by oneself, then the internalization of extrinsic motives
become regulated through identification. The individual might
say: ‘I study the night before exams because it is something
important for me.”
Extrinsic motivation – introjected (questions # 7, 14, 21, 28)
Introjected means the individual begins to internalize the reasons
for his or her actions. This form of internalization, while internal to
the person, is not truly self-determined since it is limited to the
internalization of past external contingencies. Thus the student
might say: “I study the night before exams because that’s what
good students are supposed to do.” (Vallerand, et al.,1992).
Survey Instrument
6.
Extrinsic motivation - external regulation (questions # 1, 8, 15,
22)
External regulation refers to the behavior regulated through
external means such as rewards and constraints (Vallerand, et
al.,1992). For instance, a student might say: “I study the night
before exams because my parents force me to.”
7.
Amotivation (questions # 5, 12, 19, 26)
Amotivation is either intrinsic nor extrinsic motivation. Individuals
are amotivated when they do not perceive contingencies between
outcomes and their own actions. When amotivated individuals
experience feelings of incompetence and expectancies of
uncontrollability, they perceive their behaviors as caused by forces
out of their own control.
Survey Instrument
• AMS has been widely used to measure motivation in student
populations and, based upon results, various methods have
been posited to improve motivation in students.
• This instrument has been applied to elementary, high school,
and undergraduate college students, in a number of different
languages.
• A review of the literature has not uncovered previous research
in which the AMS has been applied to community colleges.
• A motivation study of community college students would
provide useful information to community college educators,
allowing them to understand the type and level of motivation
of our students and to develop motivation interventions.
Research Objective
• The objectives of the current research are to
understand student academic motivation in a
community college and to investigate casual
relationship between academic motivation and
academic performance.
Method Definitions
• GPA = Grade Point Average
• Includes letter grades only, no pass/fail
• Includes unit credit calculations
• SR% = Success Rate
• No unit credit calculations
• A, B, C, or Pass = 1
• D, F, or Fail = 0
• “Latent” refers to unobserved variables that remove error
variance from their observed variables and can serve as errorfree IV’s in path models, as in the current study
Method
• Model 1 & 2
• Missing item replacement using regression
• Multivariate outliers removed
• Seven factor oblique used to predict GPA
• Model 1
• Cumulative, career GPA as of Fall ’13
• Model 2
• Independent observations of term 1, mid-career, and last term
success rate
Before Analysis, Model 1
• Prior to analysis, subjects with missing questions from subscales 1 (n=27), 2 (n=54),
3 (n=51), or 7 (n=39) were dropped because independent t-tests showed that their
scores differed significantly on other average subscale scores than subjects without
missing items from those subscales. 103 subjects with missing items from one of
the four subscales and were dropped, and 8 subjects without cumulative GPAs as of
fall 2013 were dropped.
• Regressions were performed to determine new scale scores for any missing items
from predicting the average score on the missing subscale from the average scores
on all the other subscales. Each new score was rounded to the nearest whole
number. Afterwards, 680 subjects had no missing items, 12 had items missing on
the first subscale, 24 had items missing from the 2nd, 12 had items missing from
the 4th, 5 from the 5th, 5 from the 6th, and 2 were missing items from both the
2nd and 5th scales and were removed. All missing items were replaced with
regression-obtained scores. 8 students with missing valid GPAs were also removed.
Only 22 students had missing items replaced with regression-obtained answers.
• 53 cases where identified through Mahalanobis distance as multivariate outliers
with p < .001 and were subsequently removed.
• The final sample size for the study is N = 640 after removing 164 participants.
Measurement
Model (CFA)
Item 2
Item 9
Item 16
Item 23
Item 6
Item 13
Item 20
Item 27
Item 4
Item 11
Item 18
Item 25
Item 3
Item 10
Item 17
Item 24
Item 7
Item 14
Item 21
Item 28
Item 1
Item 8
Item 15
Item 22
Item 5
Item 12
Item 19
Item 26
Intrinsic
To Know
Intrinsic
Accomplishment
Intrinsic
Stimulation
Extrinsic
Identified
Extrinsic
Introjected
Extrinsic
Regulation
Amotivation
Item 2
Item 9
Item 16
Item 23
Item 6
Item 13
Item 20
Item 27
Item 4
Item 11
Item 18
Item 25
Item 3
Item 10
Item 17
Item 24
Item 7
Item 14
Item 21
Item 28
Item 1
Item 8
Item 15
Item 22
Item 5
Item 12
Item 19
Item 26
Intrinsic
To Know
Structural
Model (SEM)
Intrinsic
Accomplishment
Intrinsic
Stimulation
Extrinsic
Identified
Extrinsic
Introjected
Extrinsic
Regulation
Amotivation
Cum GPA as of
fa2013
Model Fit Statistics
7 Oblique-Factor Measurement Model
Structural Model
Measure
(N=640)
Value
Measure
(N=640)
Value
Χ2
Χ2(329)=1325.81,
p <.001
Χ2
Χ2(350)=1362.72,
p <.001
Χ2baseline
Χ2(378)=29315.93,
p <.001
Χ2baseline
Χ2(406)=2992.490,
p <.001
Χ2/df
4.03
Χ2/df
3.89
RMSEA
.069 [CI .065 - .073]
RMSEA
.067 [CI .063 - .071]
CFI
.966
CFI
.966
TLI
.960
TLI
.960
model
model
Item 2
1.00
Item 9
.86
Item 16
.79
Item 23
.86
Intrinsic
To Know (#1)
-.938
(.338)
.784
(.099)
.87
Item 6
1.00
Item 13
.84
Item 20
.86
Item 27
.89
Intrinsic
Accomplishment
(#2)
.69
2.102
(.847)
.90
Item 7
1.00
Item 14
.82
Item 21
.85
Item 28
.91
Extrinsic
Introjected (#5)
-1.526
(.591)
Cum GPA as of
fa2013
Significant Predictors of GPA
Factor
B
(Error)
Intrinsic – Toward Accomplishment*
2.10
(.85)
Extrinsic – Introjected*
-1.53
(.59)
Intrinsic – To Know**
-0.94
(.34)
Extrinsic – External Regulation
0.78
(.49)
Extrinsic – Identified
-.30
(.39)
Intrinsic – To Experience Stimulation
.13
(.19)
Amotivation
-.12
(.12)
GPA* full model
R2 = .22
(.10)
GPA* only the three significant predictors
R2 = .16
(.08)
Before Analysis, Model 2
• The same students that were removed for missing data in model
1 were also absent for model 2.
• 57 cases where identified through Mahalanobis distance with p
< .001 as multivariate outliers on the motivation scale and were
removed.
• 8 cases did not have complete academic records and were
removed. 71 more were removed because they only had two or
fewer terms in their academic careers, and three were needed
to detect a linear slope.
• The final sample size for Model 2 is N = 566 after removing 229
participants.
Model 2 Definitions
• “Latent Intercept”
• Y = A + b1X1
• As in regression, serves as the starting point for an equation that
would explain academic success…
• …But latent intercept removes error in academic performance
caused by sick days, family problems, etc.
• Latent intercept could be conceived of as actual (as opposed to
simply observed) academic performance
• “Latent Trajectory”
•
•
•
•
Y = A + b1X1
Serves as beta term in regression…
…but removes error between time periods
And could be conceived of as actual academic trajectory
Unconditional Means Model
Note. All semester success rate
observations are independent.
Latent
Intercept
1
1
1
1st Semester SR%
Mid Career SR%
Last Sem SR%
.712
.700
.769
Unconditional Slope Model
-.012
Latent
Intercept
(.788)
Latent (Linear)
Trajectory
(-.041)
0
1
1
1
2
1
1st Semester SR%
Mid Career SR%
Last Sem SR%
.591
.658
.471
Intercepts of time periods were set to 0
Confirmatory Latent Trajectory
Model Fit Statistics
Unconditional Means Model
Unconditional Slope Model
Measure
(N=566)
Value
Measure
(N=566)
Value
Χ2model
Χ2(2)=20.407,
p <.001
Χ2model
Χ2(1)=0.013,
p =.91
Χ2
Χ2(3)=132.094,
p <.001
Χ2
Χ2(3)=132.094,
p <.001
Χ2/df
10.20
Χ2/df
4.92
RMSEA
.128 [CI .081 - .180]
RMSEA
.000 [CI .000 - 0.046]
CFI
.857
CFI
1.00
TLI
.786
TLI
1.02
baseline
baseline
Χ2Δ
Measure
Value
Χ2difference
Χ2Δ (1)=20.394,
p <.001
Conditional Slope Structural Model
To Know
Accomplish
ment
Stimulation
Identified
Introjected
Latent
Linear
Trajectory
Latent
Intercept
1
1st Semester SR%
Regulation
1
0
1
1
Mid Career SR%
2
Last Sem SR%
Amotivation
Conditional Slope Structural Model
*p<.05 †p<.07
Note. Intercepts of SR% in parentheses.
Accomplish
ment†
To Know*
-.15
Stimulation
Introjected
†
Identified
.29
Regulation
-.21
-.012
Linear
Trajectory
(-.041)
Intercept
(.788)
1
1
0
1
1
2
1st Semester SR%
Mid Career SR%
Last Sem SR%
(0)
(0)
(0)
.712
.677
.501
Amotivation
Structural Latent Trajectory Model
Fit Statistics
Conditional Slope Structural Model
Measure
(N=566)
Value
Χ2model
Χ2(400)=1210.9,
p <.001
Χ2
Χ2(465)=26742.761,
p <.001
baseline
Χ2/df
3.04
RMSEA
.060 [CI .056 - 0.064]
CFI
.969
TLI
.964
Results
• CFA showed that a slope along with an intercept was more of a valid
model than just the intercept
• This means that there is a latent academic performance and a
(slightly negative) latent academic trajectory over the course of
one’s academic career
• Latent trajectory of acad career is assumed to be linear and in a
downward direction from first semester, mid career semester, and
last recorded semester
• On top of this, a model of academic motivation predicting intercept
and slope provided additional explanation of variance for beginning,
mid-career, and end-career performance
• The same three factors showed at least moderate significant
prediction over and above intercept and slope of academic
achievement
Conclusions
• “To know”, “Introjected”, and “Toward
accomplishment” seem to be the most
important three factors of motivation affecting
academic performance
• In both models, “To know” and “Introjected”
were negative predictors
• “Towards accomplishment” was a positive
predictor
• The other four factors did not seem to
contribute significantly
Implications
Negative Predictor
Intrinsic motivation - to know (questions # 2, 9, 16, 23)
Intrinsic motivation to know is defined as the fact of performing
an activity for pleasure and the satisfaction that one experiences
while learning, exploring, or trying to understand something
new (Vallerand, et al.,1992).
Interested in the topic but not for testing well? Pedagogy change
toward more active learning? Educational planning?
Implication
Negative Predictor
Extrinsic motivation – introjected (questions # 7, 14, 21, 28)
Introjected means the individual begins to internalize the reasons for
his or her actions. This form of internalization, while internal to the
person, is not truly self-determined since it is limited to the
internalization of past external contingencies. Thus the student might
say: “I study the night before exams because that’s what good
students are supposed to do.” (Vallerand, et al.,1992).
Engaging in activity/learning but not effective? Educational planning?
Introjected ineffective study strategies? Outside pressure?
Implication
Positive Predictor
Intrinsic motivation - toward accomplishment (questions # 6, 13, 20,
27)
Intrinsic motivation toward accomplishments is defined as the fact of
engaging in an activity for the pleasure and satisfaction experienced
when one attempts to accomplish or create something (Vallerand, et
al.,1992).
Engagement? More focus on grades/accomplishment? Educational
planning?
Implication
Motivation enhancement interventions:
• Achievement motivation training program (AMT)- is to change the
achievement motive, an unconscious and recurring preference for
emotionally rewarding experiences related to improving one’s
performance (Pang, 2010). During these interventions, the participants
are taught to think, feel and behave like a person high in need for
achievement.
• Attributional retraining (AR) - a motivational treatment that helps students
reframe what they think about success and failure, encouraging them to
take responsibility for their academic outcomes (Haynes et al., 2009)
• Multidimensional motivation and engagement intervention - is the
motivation and engagement wheel, in which motivation is characterized in
terms of four higher order groups: adaptive cognitions (self-efficacy,
mastery orientation, valuing), adaptive behaviors (persistence, planning
and task management), maladaptive cognitions (uncertain control, failure
avoidance, anxiety) and maladaptive behaviors (self-handicapping and
disengagement). (professional development for faculty)
Limitations & Future
Directions
• Some sample bias: Same sample used in all models, but
sample was smaller in LTM.
• Latent intercepts and slopes are difficult to interpret in
general.
• Model will be re-ran again on completion.
• Survey could be tested for invariance between
completers and non-completers.
• Replicate the study (post –SEPs)

similar documents