Längsschnittskalierung der Tests zur Überprüfung des Erreichens der Bildungsstandards der Sekundarstufe I im PISA-Längsschnitt 2012/2013

The present article is devoted to the IRT scaling of tests used in the PISA Longitudinal Study 2012/2013 which verify educational standards for lower secondary school certification. It presents analyses investigating the agreement of freely estimated item parameters with the parameters calibrated in the national assessment 2012, and describes the estimation of competence levels. In addition, analyses are presented on the consequences of the unbalanced test design implemented in the retest assessment. Results indicated that the item parameters estimated for non-academic and academic track school types closely matched the pre-calibrated parameters. With few exceptions increases in competence levels during the 10th grade were estimated for both academic and non-academic track school types on basis of ability parameters estimated by means of the plausible value technique. Further analyses suggested, however, that distortions in the growth estimate are to be expected due to the unbalanced test design administered in the second survey period. Implications of these findings for the evaluation of the competence gains are discussed.