Since the development of D scores for the Implicit Association Test, few studies have examined whether there is a better scoring method. In this contribution, we tested the effect of four relevant parameters for IAT data that are the treatment of extreme latencies, the error treatment, the method for computing the IAT difference, and the distinction between practice and test critical trials. For some options of these different parameters, we included robust statistic methods that can provide viable alternative metrics to existing scoring algorithms, especially given the specificity of reaction time data. We thus elaborated 420 algorithms that result from the combination of all the different options and test the main effect of the four parameters with robust statistical analyses as well as their interaction with the type of IAT (i.e., with or without built-in penalty included in the IAT procedure). From the results, we can elaborate some recommendations. A treatment of extreme latencies is preferable but only if it consists in replacing rather than eliminating them. Errors contain important information and should not be discarded. The D score seems to be still a good way to compute the difference although the G score could be a good alternative, and finally it seems better to not compute the IAT difference separately for practice and test critical trials. From this recommendation, we propose to improve the traditional D scores with small yet effective modifications.

Richetin, J., Costantini, G., Perugini, M., Schönbrodt, F. (2015). Should we stop looking for a better scoring algorithm for handling implicit association test data? Test of the role of errors, extreme latencies treatment, scoring formula, and practice trials on reliability and validity. PLOS ONE, 10(6) [10.1371/journal.pone.0129601].

Should we stop looking for a better scoring algorithm for handling implicit association test data? Test of the role of errors, extreme latencies treatment, scoring formula, and practice trials on reliability and validity

RICHETIN, JULIETTE;COSTANTINI, GIULIO;PERUGINI, MARCO;
2015

Abstract

Since the development of D scores for the Implicit Association Test, few studies have examined whether there is a better scoring method. In this contribution, we tested the effect of four relevant parameters for IAT data that are the treatment of extreme latencies, the error treatment, the method for computing the IAT difference, and the distinction between practice and test critical trials. For some options of these different parameters, we included robust statistic methods that can provide viable alternative metrics to existing scoring algorithms, especially given the specificity of reaction time data. We thus elaborated 420 algorithms that result from the combination of all the different options and test the main effect of the four parameters with robust statistical analyses as well as their interaction with the type of IAT (i.e., with or without built-in penalty included in the IAT procedure). From the results, we can elaborate some recommendations. A treatment of extreme latencies is preferable but only if it consists in replacing rather than eliminating them. Errors contain important information and should not be discarded. The D score seems to be still a good way to compute the difference although the G score could be a good alternative, and finally it seems better to not compute the IAT difference separately for practice and test critical trials. From this recommendation, we propose to improve the traditional D scores with small yet effective modifications.
Articolo in rivista - Articolo scientifico
Psychometrics, implicit cognition, scoring, validity
English
2015
10
6
e0129601
partially_open
Richetin, J., Costantini, G., Perugini, M., Schönbrodt, F. (2015). Should we stop looking for a better scoring algorithm for handling implicit association test data? Test of the role of errors, extreme latencies treatment, scoring formula, and practice trials on reliability and validity. PLOS ONE, 10(6) [10.1371/journal.pone.0129601].
File in questo prodotto:
File Dimensione Formato  
10281-88818.pdf

accesso aperto

Tipologia di allegato: Publisher’s Version (Version of Record, VoR)
Dimensione 664.69 kB
Formato Adobe PDF
664.69 kB Adobe PDF Visualizza/Apri
PONE_RCPS2015.PDF

Solo gestori archivio

Tipologia di allegato: Publisher’s Version (Version of Record, VoR)
Dimensione 664.69 kB
Formato Adobe PDF
664.69 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/88818
Citazioni
  • Scopus 60
  • ???jsp.display-item.citation.isi??? 55
Social impact