Rethinking Grade Transfer Shock

Download Report

Transcript Rethinking Grade Transfer Shock

Rethinking Grade Transfer Shock:
Examining Its Importance In The Community College
Transfer Process
(Article published In The Journal Of Applied Research In The
Community College Vol. 14, No. 1, Fall 2006, p. 19-33).
Presented By:
Ron Pennington, Director Of Institutional Research
([email protected])
6th Annual Conference Of The Institute For The Study Of Transfer Students
January 23-25, 2008
Dallas, Texas
Introduction
2

What is Grade Transfer Shock (GTS)?
A decrease in a student’s grade point
average during their first semester at a
four-year institution when compared to
their cumulative GPA at a community
college (CC).
3
Why Is GTS A Problem?

Native student studies: CC transfers have lower
graduation rates even with SES and academic
ability controls

GTS could be an intervening variable adversely
affecting four-year student success
CC Transfer
Experience
Level of GTS
Experienced
Eventual 4-year
Success

Academic integration first/social integration later

Native four-year students will not experience GTS
4
Research Shows GTS Is A Persistent Problem

Review of the literature suggests students lose
about1/3 of a grade point
e.g. 3.0 down to 2.70
2.5 down to 2.20

Studies consistent over time
Hills (1965)
Richardson & Doucette (1980)
Diaz (1992)
Carlan & Byxbe (2000)
5
Why Might CC Transfers Experience GTS?

Poor academic prep at the CC level (Dougherty,
2000)



Within an institution – academic in-process
measures
Between institutions
Poor transfer prep (Nolan & Hall, 1978; Holahan & Kelley,
1978; Land, 1996; Laanan, 1996; Lee & Hoey, 1996; Rhine, 2000,
Debard, 1996)

Poor cognitive maps (Lovitts, 2001)

Attribution Theory (Heidner, 1958; Weiner, 1974)
6
Potential Interventions To Reduce GTS

Change the emphasis from traditional transfer
counseling strategies like:
 Where to transfer
 Meeting the prerequisites of four-year schools
To:

More proactive strategies designed to reduce GTS
 Workshops on the new four-year academic culture
 Student mentoring programs at the four-year school
(Laanen, 1996; Rhine et al., 2000)
7
Research Questions

Is GTS related to four-year student success?

Does GTS occur when student demographic
and institutional process variables are
controlled?

Do traditional two-year and four-year transfer
counseling practices reduce GTS levels?
8
Methodology
9
Measuring GTS Is Problematical

Gain score:


(4-year term GPA) – (2-year cumulative GPA)
Problem: The two GPA measures are different


Based on two schools’ grading system
4-year term GPA is less reliable than the CC cumulative
GPA



Less course taking
Shorter time period
Regression to a lower 4-year GPA scale
10
Two Basic Solutions

Using a lower level of measurement:

A dichotomous variable
GTS
No GTS
-.25

+.25
An ordinal variable
Negative Grade
Change (GTS)
Positive Grade
Change
No Grade Change
-.25

0
0
+.25
Regress the CC cumulative GPA on the 4-year
term GPA
11
Data Collection Methods

Telephone survey of MO community college
transfer students – Summer 1999

Student data came from community colleges



Demographic
Academic in-process measures
Student outcome data (MO EMSAS file)
12
Study’s Sample

Initial list of 7,055 CC transfer students completed 24
credit hours from 1995 to 1998

2,656 were surveyed using several call back attempts
(response rate = 38%)


Additional criteria used to eliminate cases




Many outdated telephone numbers
Senior transfers (>96 credits)
Pooling of 5 urban community colleges
First-time transfers prior to fall 1998 semester
Usable cases = 686
13
Findings
Is GTS Related To Four-Year Student Success?
15


Modest relationship between GTS and CC
transfer success at four-year schools.
Grade measure of GTS better predictor of
transfer success than survey measure
16

Nearly three times as many students actually
experienced GTS than reported it in the survey
17
Regression Findings

Does GTS occur when student demographic
and institutional process variables are
controlled?

Do transfer two-year and four-year traditional
counseling practices reduce GTS levels?
18
Independent Variables
+ CCGPA
- 4-Year ACT
CC Academic Challenge
CC1
CC2
±
CC3
CC4
CC5
+ Age
- Gender (0=F, 1=M)
- African American (0=AA,1=Oth)
+ Previous College (0=N, 1=Y)
- CC Lib Arts Maj (0=N,1=Y)
- CC Bus Maj (0=N, 1=Y)
+ CC Degree (0=N, 1=Y)
+ CC Financial Aid (0=N, 1=Y)
+ Cumulative CC Credits
- Dev course work
+ 4-Year First-Term Credits
Transfer Experience
+ CC Prep (0=not SAT, 1=SAT)
+ CC Acad Adv (0=no, 1=yes)
+ CC Fac Adv (0=no, 1=yes)
+ CC Couns (0=no, 1=yes)
+ Cred Transfer Success
(0=no, 1=yes)
+ 4-year Couns (0=no, 1=yes)
19
Summary Of Key Regression Findings

CC GPA was the strongest predictor variable of 4-year GPA by 4 to 1

4-year and 2-year academic challenge variables were the second
strongest set of predictors

Other significant variables were:






Taking developmental CC coursework (indicator of academic
readiness?)
Age (indicator of maturation?)
CC financial aid (indicator of financial dependency at the CC?)
4-year credits (indicator of clearer transfer goals)
Controls on demographic and institutional process variables
actually enhanced GTS
Traditional counseling variables were not significant individually or
as a set
20
Regression Findings: Results
Significant
Variable
CC GPA
4-Yr. ACT
CC4+
Age
Dev Crs Wk
Fin Aid
4-Yr. Cr
+
Not Significant
Exp.
Pos
Neg
—
Pos
Neg
Pos
Pos
Actual
Pos
Neg
Neg
Pos
Neg
Neg
Pos
Beta
0.46
-0.12
0.11
0.10
-0.10
-0.09
0.08
Gender
African-American
Previous college
Bus major at CC
Lib Arts major at CC
CC degree
CC credits
All other transfer experience variables
All CC academic vars sig as a set (F=4.99)
21
Summary of the Regression Results By Sets and CC GPA Coefficients
Set By Order of Data Entry
1. Transfer Shock (CC GPA)
2. 4-Year Academic Challenge
3. CC Academic Challenge
4. Student Background
5. CC Experience
6. 4-Year First Sem Term Experience
7. Transfer Experience
* p<,05, ** p<.01, *** p<.001
R2
Variables Change in
In Set CC GPA b R2 Change
.29
.29
.91
1
.01
.30
.93
1
.02
.32
.93
5
.01
.33
.85
4
.02
.35
.80
6
.01
.36
.79
1
.00
.36
.78
6
Change
in F
255.27 ***
11.95 **
3.15 *
3.57 *
2.55 *
5.35 *
0.86
22
Implications And Discussion
23

GTS can be measured as a

CC GPA  4-year regression study
Or

As a dummy variable in a regression study
24

Regression study question

Will the relationship between GTS and 4-year outcome
success hold up under various controls?

This study shows that the GTS variables should be split
at -.25 to -.30 if coded as a dummy variable
25

Other predictor variables should be examined

More academic process variables at the CC level




Like this study’s CC developmental coursework, CC
financial aid, and 4-year credit variables
Other examples: school attendance, course scheduling
(Karl Boughhan)
Student engagement
Inter-institutional variables like the 4-ACT and set of
CC variables


Will be needed for institutional accountability assessments
Hierarchical linear modeling could be used to “level out the
playing field”
26

Need to test if new transfer counseling
programs should be adopted

Specific program interventions



Better financial aid assistance and information
Counseling program (two or four-year) targeted to
increase students’ Cognitive maps (campus visits,
student mentoring, etc.)
More systemic strategies and explanations

Attributional Theory

vs.
Academic and Social Integration models
27

What is Attributional Theory?

A psychological theory – instead of a sociological
theory

An achievement-motivational theory that predicts
a person's future motivation to act based on
causative explanations for why certain outcomes
have occurred in the past

Concepts include:



Locus of control
Controllability
Event stability
28

Many have argued that intervention programs
based on attribution theory could improve the
academic success of CC transfers (Finley &
Cooper, 1983; Pascarella, Edison, Hagedorn,
Nora, & Terenzini, 1996; Perry, Hector, Menec,
& Weinberg, 1993; St. Clair, 1993; Valla, 1989)

But all future program interventions to improve
GTS need to be evaluated


Need a program logic for how the intervention is
suppose to work
Need to implement an experimental design to see
if it does work
29
30