|
|
 |
 |
|
|
|
|
Outliers |
| We conducted supplementary
analyses to ensure that outliers did not unduly impact
residual results. We identified outliers using studentized
deleted residuals. To begin with, we dropped institutions
with studentized deleted residuals that were greater than
|3.0|, and reran the regression models. Institutional
residuals did not shift appreciably after outlier schools
were dropped. To check the stability of the residuals
with a more inclusive cut-off for outliers, we dropped
institutions with studentized deleted residuals greater
than |2.25|. As with the more stringent cutoff, institutional
residuals did not shift appreciably when this larger group
of outliers was removed. Therefore, each Institutional
Engagement Index report is based on an analysis of all
institutions in the 2000-2002 National Norms. |
|
|
| |
|
|
|
|
|
