Search

The purpose of the study was to explore the validity of a novel computer-analytic developmental scale, the Writing Ability Developmental Scale. On the whole, collective results supported the validity of the scale. It was sensitive to writing ability differences across grades and sensitive to within-grade variability as compared to human-rated sensitivity. It predicted scores on the human-rated continuous scale very well. As compared to the human-rated continuous scale, it demonstrated similarly constant mean differences across ethnicity and socioeconomic status-although not for all rising grade differences or for gender. There was no genre differential across the two scales, and measurement precision was highly similar for the two scales.

The article focuses on the multiple item model specifications for the measurement of dichotomous items. It explores the role of Rasch model and its important features for dichotomous data compared to other psychometric models and which is a specification that facilitates conjoint measurement. It states that measurement for dichotomous items should involve replication which can only be attained with more number of items with features included in the model with multiple items.