Related Experiment Videos
Dangerousness: assessing the risk of violent behavior
C E Shaffer1, W F Waters, S G Adams
1Department of Psychology, Louisiana State University, Baton Rouge 70803.
Journal of Consulting and Clinical Psychology
|October 1, 1994
Summary
This study developed accurate discriminant models to identify dangerous individuals in hospital and prison settings. Population-specific models showed higher accuracy in predicting dangerousness.
Area of Science:
- Forensic Psychology
- Clinical Psychology
- Behavioral Science
Background:
- Accurate identification of dangerous individuals in institutional settings is crucial for safety.
- Existing methods for assessing dangerousness may lack specificity for different populations.
Purpose of the Study:
- To develop and validate discriminant models for identifying dangerous inpatients and prison inmates.
- To compare the accuracy of population-specific models versus a combined model.
Main Methods:
- Discriminant analysis was applied to hospital (N=100) and prison (N=100) samples.
- A stepwise discriminant analysis was performed on a combined sample (N=200).
- Models were evaluated based on their accuracy in classifying individuals as dangerous or nondangerous.
Main Results:
- The hospital-specific model (5 variables) achieved 85% accuracy.
- The prison-specific model (6 variables) achieved 72% accuracy.
- A combined model (8 variables) achieved 75% accuracy in classifying dangerousness.
Conclusions:
- Population-specific discriminant models are empirically valid for assessing dangerousness in hospital and prison settings.
- Hospital-specific models demonstrated higher predictive accuracy than prison-specific or combined models.
- These findings support the use of tailored assessment tools for different institutional populations.