Limits...
Using item response theory to investigate the structure of anticipated affect: do self-reports about future affective reactions conform to typical or maximal models?

Zampetakis LA, Lerakis M, Kafetsios K, Moustakis V - Front Psychol (2015)

Bottom Line: In the present research, we used item response theory (IRT) to examine whether effective predictions (anticipated affect) conforms to a typical (i.e., what people usually do) or a maximal behavior process (i.e., what people can do).We found that the GRM provided a better fit to the data.The paper also discusses implications for a growing literature on anticipated affect.

View Article: PubMed Central - PubMed

Affiliation: Management Systems Laboratory, School of Production Engineering and Management, Technical University of Crete Chania, Greece.

ABSTRACT
In the present research, we used item response theory (IRT) to examine whether effective predictions (anticipated affect) conforms to a typical (i.e., what people usually do) or a maximal behavior process (i.e., what people can do). The former, correspond to non-monotonic ideal point IRT models, whereas the latter correspond to monotonic dominance IRT models. A convenience, cross-sectional student sample (N = 1624) was used. Participants were asked to report on anticipated positive and negative affect around a hypothetical event (emotions surrounding the start of a new business). We carried out analysis comparing graded response model (GRM), a dominance IRT model, against generalized graded unfolding model, an unfolding IRT model. We found that the GRM provided a better fit to the data. Findings suggest that the self-report responses to anticipated affect conform to dominance response process (i.e., maximal behavior). The paper also discusses implications for a growing literature on anticipated affect.

No MeSH data available.


Scatter-plot comparisons of the theta value estimates from graded response model (GRM) and generalized graded unfolding model (GGUM).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4585228&req=5

Figure 2: Scatter-plot comparisons of the theta value estimates from graded response model (GRM) and generalized graded unfolding model (GGUM).

Mentions: Results suggest that the GRM produces a better model-data fit compared to the GGUM. Although the absolute fit of the two models is above to the recommended χ2/df fit criteria it is suggested that when moderate sample sizes are used (i.e., 1000–1500), a fixed cut-off value is insufficient to ascertain IRT model-data because of high Type I error rates (Tay and Drasgow, 2012a). The theta estimates for the GRM and GGUM showed some correlation (r = 0.47). Examining the scatterplot matrix of the individual estimates (Figure 2) depicted an inverted funnel distribution. The incongruence between the GRM and GGUM occurred at higher levels of theta.


Using item response theory to investigate the structure of anticipated affect: do self-reports about future affective reactions conform to typical or maximal models?

Zampetakis LA, Lerakis M, Kafetsios K, Moustakis V - Front Psychol (2015)

Scatter-plot comparisons of the theta value estimates from graded response model (GRM) and generalized graded unfolding model (GGUM).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4585228&req=5

Figure 2: Scatter-plot comparisons of the theta value estimates from graded response model (GRM) and generalized graded unfolding model (GGUM).
Mentions: Results suggest that the GRM produces a better model-data fit compared to the GGUM. Although the absolute fit of the two models is above to the recommended χ2/df fit criteria it is suggested that when moderate sample sizes are used (i.e., 1000–1500), a fixed cut-off value is insufficient to ascertain IRT model-data because of high Type I error rates (Tay and Drasgow, 2012a). The theta estimates for the GRM and GGUM showed some correlation (r = 0.47). Examining the scatterplot matrix of the individual estimates (Figure 2) depicted an inverted funnel distribution. The incongruence between the GRM and GGUM occurred at higher levels of theta.

Bottom Line: In the present research, we used item response theory (IRT) to examine whether effective predictions (anticipated affect) conforms to a typical (i.e., what people usually do) or a maximal behavior process (i.e., what people can do).We found that the GRM provided a better fit to the data.The paper also discusses implications for a growing literature on anticipated affect.

View Article: PubMed Central - PubMed

Affiliation: Management Systems Laboratory, School of Production Engineering and Management, Technical University of Crete Chania, Greece.

ABSTRACT
In the present research, we used item response theory (IRT) to examine whether effective predictions (anticipated affect) conforms to a typical (i.e., what people usually do) or a maximal behavior process (i.e., what people can do). The former, correspond to non-monotonic ideal point IRT models, whereas the latter correspond to monotonic dominance IRT models. A convenience, cross-sectional student sample (N = 1624) was used. Participants were asked to report on anticipated positive and negative affect around a hypothetical event (emotions surrounding the start of a new business). We carried out analysis comparing graded response model (GRM), a dominance IRT model, against generalized graded unfolding model, an unfolding IRT model. We found that the GRM provided a better fit to the data. Findings suggest that the self-report responses to anticipated affect conform to dominance response process (i.e., maximal behavior). The paper also discusses implications for a growing literature on anticipated affect.

No MeSH data available.