Predicted
Level of Academic Challenge =
47.788 + 1.171*selectivity + 2.697*sector - .389*rural
- 2.347*small town - 2.632*large town - 1.417*fringe1
- 1.276*mid-size - .833*fringe2 - 1.012*Extensive - .157*Intensive
+ 1.956*Liberal arts + .642*General + .398*Other Carnegie
+ 2.719*on campus + 2.689*full-time – .703*humanities
–
2.624*math – 5.711*preprofessional - 5.709*multiple
- 1.589*Alien – .764*Asian + 2.561*Black + 2.958*Latina/o
- 1.716*Native American - .064*Unknown
Adjusted R2=.58
Predicted Active and Collaborative
Learning =
26.384 + 3.279*sector - 2.891*Extensive - 1.309*Intensive
+ .625*Liberal arts + .617*General + 2.462*Other Carnegie
+ 3.877*on campus + 4.366*full-time + .197*age + 4.575*humanities
+ 4.851*math – .051*preprofessional + .370*multiple
+ 3.691*Alien – .505*Asian + 6.381*Black + 7.708*Latina/o
+ 9.159*Native American - .943*Unknown
Adjusted R2=.44
Predicted Student-faculty Interaction =
35.899 + 3.235*sector - .131*enrollment + .086*Extensive
+ .079*Intensive + 2.426*Liberal arts - .218*General +
3.492*Other Carnegie + 4.311*on campus - .3.451*humanities
- 2.222*math – 7.459*preprofessional - 8.136*multiple
+ + 5.053*Alien – 7.186*Asian + 6.295*Black + 6.175*Latina/o
+ 11.939*Native American - 4.449*Unknown
Adjusted R2=.46
Predicted Enriching Educational
Experiences =
67.184 + 1.484*selectivity + 3.607*sector + .501*Extensive
+ .358*Intensive + 3.207*Liberal arts - .132*General -
2.020*Other Carnegie + 4.987*on campus - .637*age - .700*humanities
- 13.676*math – 17.588*preprofessional - 12.505*multiple
+ 14.185*Alien + 7.049*Asian + 9.629*Black + 10.868*Latina/o
+ 39.427*Native American + 3.389*Unknown
Adjusted R2=.66 |
Predicted
Supportive Campus
Environment =
50.698 + 4.182*sector - 2.756*Extensive - 2.036*Intensive
+ .983*Liberal arts + .835*General + 1.921*Other Carnegie
+ 5.702*on campus + 4.387*full-time - 2.581*Alien –
9.273*Asian - .694*Black + 11.089*Latina/o + 6.997*Native
American - 5.109*Unknown
Adjusted R2=.49
Notes:
- Residual models initially
contained all independent variables listed below.
Model reduction proceeded one variable at a time;
each iteration removed the variable associated with
the largest p value. For all but dummied variables
with more than two categories, we removed variables
that did not yield significant coefficients (p<.05
with two-tailed tests). For urbanicity, 2000 Carnegie
Classification, major field, and race/ethnicity,
we tested whether all dummy coefficients for each
variable were simultaneously equal to zero. If Wald
tests suggested that we could not reject the null
hypothesis (chi-square, p<.05), we dropped the
variable from the model (with its attendant dummy
variables), and
- As the independent variables
were not mean-centered in any way, model intercepts
have no special interpretation.
|
|
