To sum up, the factors affecting item difficulty are different depending on the proficiency of the test takers.
More recently, Susanti et al. (2016) conducted an investigation of several potential factors affecting item difficulty in the vocabulary questions used in TOEFL.
Beinborn et al. (2014) worked on predicting the gap difficulty of the C-test3 using a combination of factors such as phonetic difficulty and text complexity, while Hoshino and Nakagawa (2010) and Susanti et al. (2016) investigated factors affecting item difficulty on multiple-choice vocabulary questions.
Although the data reflect in situ assessment practices, there is a need for more systematic investigations that look across all cognitive skill levels to further elucidate relationships between cognitive skill level and item difficulty and identify the multiplicity of factors affecting item difficulty, including student epistemologies.
Hence, the similarity between the correct answer and distractors is considered to be a factor affecting item difficulty.
We analysed the collected data against to answer three research questions: (1) whether the item difficulty can be controlled using the investigated factors, (2) which factor contributes the most to item difficulty and, (3) how these factors affect the item difficulty across test takers with different proficiency.
This factor structure eliminates the items of 9, 10, 13, 15, 17, and 19, and includes 3 structural factors of negative affect (items 3, 6, 14, 18), anhedonia (items 4, 8, 12, 16), and somatic complaints (items 1, 2, 5, 7, 11, 20).
How do these factors affect the item difficulty across test takers with different proficiencies?
How do these factors affect the item difficulty across test takers with different proficiencies? .
In this phase, we aimed to draw on the NPT to develop a comprehensive set of general items –TARS items - reflecting factors affecting the routine use of e-health ready for application in specific settings.
Our results were overall consistent with the meta-analytic work by Cosco et al. [ 6], indicating that a bifactor model with a general factor affecting all items and two orthogonal group factors, accounting for a specific anxiety and depression variance, was the best fitting one compared to six alternative factor structures reported in the literature [ 1, 7, 10, 13, 21, 31].
