The Model And Advantages On Comparative Study Of Vocational Interest Test

Unfolding model was a hot issue in psychological testing recent years. some researchers found that GGUM fit the data better than dominance models in personality and attitude tests. However, we didn’t get a clear insight into the extent to what unfolding model fit the vocational interest data. Unfolding Model and Dominance Model in scale development had different item location distribution, with adding more neutral item in unfolding model. But there wasn’t an explicit conclusion that adding neutral item can result the improving of test reliability and validity.This research investigated the differences between the Unfolding Model and Dominance Model in vocational interest measurement and sale development on1250college students..First, study1comparing the GGUM to the GRM was conducted to compare the accuracy of latent trait estimates from these models, using chi-square ratios divided by degrees of freedom for singles, pairs, and triples of items. Second, the author developed scales using traditional classical test theory and ideal point IRT methods. The merits of each method were examined in terms of measurement precision, and criterion-related validity.Result shows that:(1) GGUM can fit the interest data better and provide more information than cumulative IRT models,GGUM can provide more accurate0estimate for person with extreme high interest;(2)By adding more moderate-level Items, scale developed by GGUM have an advantage over scale developed by Likert techniques in higher test reliability. But two way showed no difference in criterion-related validity.In sum:Unfolding model generally performed as well as dominance model in a majority of cases. Unfolding model can provide more accurate0parameter for individuals who hold extreme vocational positions. If a researcher want to identify those individuals who like or dislike a given vocation type the most, then unfolding model was more useful than dominance model, If a researcher was interested in the majority of cases in the middle of the0distribution, then the more simple dominance model was a better choice. For scale construction, unfolding model has advantage in reliability, unfolding model and dominance model produce no different validity. If a researcher want to get the most accurate0parameter of individuals, then unfolding model was more useful than dominance model, If a researcher was interested in the relative location of individuals, then the more simple dominance model was a better choice.