Exact(6)
Each item exhibited distributions reflecting sensitivity to variations in the attributes measured.
Two criteria were used to attribute each item to one of the dimensions: like other authors [ 33], factor loading >0.60 with one principal component was chosen, instead of factor loading >0.40 [ 34], and when an item exhibited factor loading across several dimensions, it was attributed to the one for which it maximized internal consistency assessed by Cronbach's α-coefficient [ 34].
Among the total of 64 items, only two had an RTF smaller than 0.94 (i.e., about 6% of the students taking either item exhibited rapid-guess behavior), ten had an RTF lying between 0.96 and 0.98, and the rest had an RTF greater than 0.98.
Only one item exhibited DIF between the two modes.
This item exhibited quite a low loading on its respective factor (β = 0.19) but it was included in the model since the item was considered to measure a very important aspect in the ABR scale and this item was shown to have statistically significant association with its factor (βABR4: 0.186, p < 0.001).
Furthermore, the ordered natures of the category boundary threshold parameters were estimated as -3.06, -2.36, 0.44, and 4.97 under the rating scale model, indicating that no item exhibited disordering of the step difficulty and the 5-point rating scale was appropriate [ 36, 37].
Similar(2)
Items that exhibited differential item functioning between countries were also removed from the measure.
However, the item 28 exhibited values at the upper threshold for misfit according to Rasch analysis and presented DIF.
Write better and faster with AI suggestions while staying true to your unique style.
Since I tried Ludwig back in 2017, I have been constantly using it in both editing and translation. Ever since, I suggest it to my translators at ProSciEditing.

Justyna Jupowicz-Kozak
CEO of Professional Science Editing for Scientists @ prosciediting.com