The research assessed the clarity and ease of use of the research instruments and any sensitive, biased items were identified and modified.
Based on findings from Gajewski et al. [ 20], the subject experts tend to overestimate the relevancy of items, resulting in highly biased item-to-domain correlations.
Confirmatory factor analysis for ordered-categorical measures (CFA-OCM) uncovered two biased items.
Further, when Filipinos were compared to Hawaiians, there were five significantly biased items.
Therefore, it is not recommended that any of the biased items be removed, or even reworded.
This supports the stated hypothesis that biased items would exist for Filipinos when compared to Caucasian and Japanese groups.
Psychological science should seek the source of this bias and carefully examine the appropriateness of diagnosing and describing cultural minorities using biased items.
Sex bias was significant in two items (one relating to feeling "anxious" and the other to feeling "unattractive" because of PH), and age bias was significant in three items; all five of these biased items were removed.
The purification process for finding a DIF-free anchor item set identified item 9 (χ2 (4) = 92.1, p <.001), item 3 (χ2 (4) = 27.2, p <.001), and item 4 (χ2 (4) = 15.8, p =.003) (results not tabulated) as potentially biased items.
However, if the groups being compared are not equal in the underlying construct being measured, items exhibiting DIF are not necessarily biased and biased items will not necessarily show DIF.
