Risk Factors
Validity Issues
Tables and Charts
Tests and Samples
Misc.
100

variables that reduce the likelihood of recidivism

Protective Risk Factor

100

The proportion of individuals who have a condition of interest that are correctly classified by a test

Sensitivity

100

A statistical procedure that allows for simultaneous consideration of sensitivity and specificity at various points along a range of scores

Receiver Operating Characteristics (ROC)

100

a risk factor associated specifically with the individual being assessed

Idiographic Risk Factor

100

Statistical procedures for estimating risk of recidivism that are specifically designed to exclude professional judgment

Nondiscretionary risk assessment procedures

200

a dynamic risk factor that is long standing

Stable Risk Factor 

200

The proportion of individuals who do not have a condition of interest that are correctly classified by a test

Specificity

200

a mistaken prediction that an event will not occur

False negative error

200

a statistical table that depicts recidivism rates for subgroups of the derivation sample and their corresponding scores on an actuarial risk assessment instrument

Experience Table

200

a highly individualized approach to risk assessment that seeks to identify personal risk factors for an individual offender

Anamnestic assessment

300

A risk factor for recidivism that cannot be changed by intervention

Static Risk Factor 

300

Occurs when high-risk groups have lower recidivism rates than low-risk groups

Probability reversals

300

An index of the overall classification accuracy of a risk assessment instrument

Area Under the Curve (AUC)

300

The degree to which different raters (evaluators) obtain similar results

Inter-rater reliability 

300

A fully unstructured approach to risk assessment that relies on clinical knowledge and intuition

Unstructured Clinical Judgement (UCJ)

400

a transient, short-lived, dynamic risk factor

Acute Risk Factor

400

the problem wherein rare events are more difficult to predict than commonly occurring ones

Base rate problem

400

A mistaken prediction that an event will occur

False positive error

400

research samples that are used to test the generalizability of recidivism data collected from the derivation sample for ARAIs

calibration or cross-validation sample

400

risk assessment procedures designed to include professional judgment

Discretionary risk assessment procedures

500

A risk factor for recidivism that can be altered by intervention

Dynamic Risk Factor 

500

Occurs when the predictive validity of ARAIs decreases in cross-validation

Shrinkage

500

seek to predict membership in one of two groups. are specific predictions that an offender will or will not reoffend

Binary predictions

500

the sample of offenders that provides the foundation for the statistical prediction formula and experience table that are the core components of actuarial risk assessment instruments

derivation sample

500

A research design used to combine the results of several studies into essentially one large study. Meta-analysis provide a quantitative review of a large body of research

Meta-analysis

M
e
n
u