Table 3.
Model-based Marginal Proportions for Statistically Significant Effects of Primary Predictors on Ideal Behavior Outcomes*
| Effects and interactions | Disclosure of conflict of interest, % | Compliance with human-subjects requirements, % | Setting of clear rules with trainees, % | Monitoring of trainees’ development, % | Maintenance of data integrity, % |
|---|---|---|---|---|---|
| Main effects | |||||
| External funding expected by administration | |||||
| No | -- | -- | 51.7 | -- | -- |
| Yes | -- | -- | 45.1 | -- | -- |
| Federal funding held | |||||
| No | -- | NA | 51.7 | 58.2 | NA |
| Yes | -- | NA | 44.8 | 50.9 | NA |
| Private industry involvement | |||||
| No | 98.7 | NA | -- | -- | NA |
| Yes | 96.1 | NA | -- | -- | NA |
| Interactions | |||||
| No private industry involvement | |||||
| Federal funding | |||||
| •No | -- | 88.1 | -- | -- | 81.3 |
| •Yes | -- | 88.5 | -- | -- | 82.3 |
| Private industry involvement | |||||
| Federal funding | |||||
| •No | -- | 93.5 | -- | -- | 87.2 |
| •Yes | -- | 83.3 | -- | -- | 77.1 |
*
NA, interpretation was not appropriate because of the presence of a significant interaction. All models were adjusted for sex, number of years since receipt of highest degree, tenure status, field of study, and minority status.