Early Identification of Introductory Major's Biology Students for Inclusion in an Academic Support Program

1
Early Identification of Introductory Major's
Biology Students for Inclusion in an Academic
Support Program
BETHANY V. BOWLING and E. DAVID
THOMPSON Department of Biological Sciences,
Northern Kentucky University
At Northern Kentucky University our first
semester Introductory Biology Course for science
majors has an attrition rate of 35-40. In an
effort to improve student success in the course
we sought to identify students early in the
semester in need of academic support. Based on
previous research we investigated the use of ACT
scores in math (ACT-M) and scores on a test of
scientific reasoning (SR Lawson, A.E. 2000) as
predictors of student success. Students scoring
in the lowest third in either ACT-M or SR were
randomly selected to be included in a
university-wide peer mentoring program in which
an upper-level student served as a small group
leader for reviewing course material. Our
results show that both SR (r0.476, plt.001) and
ACT-M (r0.325, plt0.001) are significantly
correlated to student grades in the course,
although SR appears to be a slightly better
predictor of student success. A hierarchical
multiple regression model including both
parameters was found not to be statistically
significant (r.480, pgt.05), indicating the best
model may be SR scores alone. While further
investigation into the use of SR and ACT-M scores
as predictors of success in the course is being
conducted, our results suggest peer mentoring in
its current form may not be an effective academic
support strategy as there was no significant
difference between students in the program and
those with equally low predicting factors
t(47)0.26, pgt.05.
Results
Regression Analysis of Predictors of Student
Success
Student Success vs. Predicting Factors
Parameter B SE of B ß
Step 1
Constant 56.355 3.800
SR 1.207 .259 .476a
Step 2
Constant 53.097 6.800
SR 1.099 .320 .434
ACT-M .213 .368 .073
Background
A number of studies over the last thirty years
have considered various factors related to
student success in introductory biology courses,
including prior content knowledge, self-efficacy,
cognitive style, developmental level, college
entrance exam scores, grade point average (GPA),
success in prior courses, reasoning ability, and
course pedagogy. Anton Lawson, a well-known
biology education researcher, has authored
several papers on this topic. His research has
continuously shown that scientific reasoning (SR)
ability is the best predictor of student success
in introductory biology courses (1980 1997
2007), not prior content knowledge as many
instructors mistakenly believe. Additional
research has been conducted on standardized
college entrance exams, the ACT and SAT, as
predictors of success in individual courses,
since they are considered a predictor of success
for the first year of college in general . These
studies have indicated mixed results (Freeman et
al., 2007 Johnson Lawson, 2007). Previous
research at NKU has suggested that ACT math
scores (ACT-M) are a valid predictor of student
success in our introductory biology courses. In
regards to the present study, we identified the
need to provide academic support for students in
our introductory biology course due to the high
attrition rate of around 40. Traditionally, a
university-wide sponsored peer mentoring program
was offered to students seeking assistance.
Since often students do not know of their
performance level until after the first or even
second exam, we sought to identify students early
in the semester in need of academic support.
Based on previous research in the literature and
findings at our institution, we investigated the
use of ACT-M and SR as predictors of student
success.
Note R2 .226 for step 1, ?R .003 for step
2. a. p lt .001
Note SR (N84) R2 .226, ACT-M (N123) R2
.106.

Objectives

1. Determine the best factor for identifying
   students early in the semester in need of
   academic support
2. Analyze effectiveness of peer mentoring program

Methods
With IRB approval, the SR test was administered
to students in the first semester of the NKU
Introductory Majors Biology course series (BIO
150) on a voluntary basis. Prior ACT-M scores for
the same students were also collected. Students
were notified that results of the SR test and
ACT-M scores would be used as the basis for our
selection of students to take part in a peer
mentoring program, conducted by the NKU
Department of Biological Sciences and the Office
of the Associate Provost for Student Success.
Student volunteers scoring in the lowest third of
the SR exam were selected. The same selection
process was utilized for ACT-M scores. In all,
60 students were identified, 30 each based on SR
and ACT-M scores. These students were divided at
random into two groups, those who would receive
peer mentoring and those who would not take part
in the program. As such, we had a natural
control for the study. Each peer mentoring
group consisted of 5 students and one peer
mentor. Peer mentors were upperclassmen selected
because of proficiency in the subject matter, as
evidenced by overall GPA and previous success in
the same course. Each group was instructed to
meet for one collaborative learning session each
week at a time convenient for all members. The
primary responsibility of the peer mentors was to
provide structure to the collaborative learning
session, and to ensure that problem-solving
progressed at a reasonable pace. Additionally,
peer mentors were required to shadow each
student member of their group. This entailed
meeting with each student every two weeks to
assist with management of the academic schedule,
to provide proactive mentor support, and to
monitor academic progress. Overall BIO 150
lecture percentage was selected as the measure of
student success in the course, with 70 as the
selected value required for successful completion
of the course. At the end of the semester,
overall percentages were graphed as a function of
SR or ACT-M and analyzed via correlation
coefficients. Based on these data, a
hierarchical multiple regression model was also
conducted. After determining SR to be the best
known predictor for our course, we developed a
linear regression model to predict the grade that
students were likely to receive based on SR
scores (course percentage) 56.36 (1.2)(SR),
R2 0.23. Using this model, a score of 11 or
better on the scientific reasoning test
correlates to 65.9 or higher in BIO 150. The 33
students who had SR scores of 11 or less were
compared to the remaining 113 students who
finished the course and received SR scores of 12
or higher. To discern the efficacy of the peer
mentoring program, the same 33 students who had
SR scores of 11 or less were compared based on
whether they received or did not receive peer
mentoring.
Comparison of BIO 150 final lecture scores based
on predicted success (SRgt11). Note The bars show
arithmetic means with error bars indicating
standard error of the mean, denotes values that
are significantly different (p 0.005 t-test).
Comparison of BIO 150 final lecture scores of
students predicted to be at-risk at taking part
in peer mentoring. Note The bars show arithmetic
means with error bars indicating standard error
of the mean (p gt 0.05 t-test).
Discussion
In our study, both SR (r0.476, plt.001) and ACT-M
(r0.325, plt0.001) were significantly correlated
with students grades in the course. The
correlation for SR and course grade are similar
to others in the literature. For instance, two
of Lawsons studies reported correlations between
SR and course grade as r0.43 (1998) and r0.45
(2007) . The correlation between ACT-M and
course grade does differ from one study by Lawson
which reported a correlation of the more general
ACT/SAT scores with course grade as r0.52. Our
results are also a bit contradictory to previous
studies indicating ACT-M as a predictor of
success in our introductory biology courses.
However, the previous studies conducted did not
consider SR. Since both the predictors were
significantly correlated with course grade, we
conducted a hierarchical multiple regression
model including both parameters. Interestingly,
the model was found not to be statistically
significant (r.480, pgt.05), suggesting it was
not useful as a predictive tool. Based on our
current data, it appears that SR alone may be the
best predictor for student success in NKUs
introductory biology courses. We are continuing
our research efforts in this area with slight
modifications to the protocol for administering
the SR test. Preliminary findings show weve had
a greater participation rate by the students,
thus increasing our sample size and providing a
more representative sample of the population.
This may contribute to a higher correlation
coefficient between our predictor and course
grade, providing a more accurate model. A
comparison of final lecture scores based on SR
results indicated a significant difference in the
scores of those students above the predicted
threshold (SR gt 11) compared to students who
scored below the threshold (p0.005). This
further suggests that scientific reasoning
ability is a worthy predictor for student success
in NKUs introductory biology courses. However,
students who scored below the threshold did not
seem to benefit from the peer mentoring program.
Students with at-risk SR scores who received peer
mentoring performed no better than students with
equivalent SR scores who did not receive peer
mentoring (pgt0.05). Previous versions of the
peer mentoring program have offered a financial
award for success in the course, which may have
led to poor performance on the SR test in this
study. To address this, we are exploring the
possibility of developing a peer mentoring
program in the future without financial awards.