A Machine Learning Approach to Assess Differential Item Functioning in Psychometric Questionnaires Using the Elastic Net Regularized Ordinal Logistic Regression in Small Sample Size Groups

Figure 2

The average powers and type I error rates of the nonregularized OLR (solid lines), elastic net OLR model with (ridge) (dashed lines), elastic net OLR model with (dotted lines), elastic net OLR model with (dot-dashed lines), and elastic net OLR model with (LASSO) (long-dashed lines) on measures with five and ten items for severe DIF (). (a) and . (b) and . (c) and . (d) and . (e) and . (f) and .

(a)

(b)

(c)

(d)

(e)

(f)