بررسی سفارش آیتم ثابت در بهداشت روانی: مثالی از استفاده از روش های مختلف

Invariant item ordering is a property of scales whereby the items are scored in the same order across a wide range of the latent trait and across a wide range of respondents. In the package ‘mokken’ in the statistical software R, the ability to analyse Mokken scales for invariant item ordering has recently been available and techniques for inspecting visually the item response curves of item pairs, have also been included. While methods to assess invariant item ordering are available, there have been indications that items representing extremes of distress in mental well-being scales, such as suicidal ideation, may lead to claiming invariant item ordering where it does not exist. We used the Mental Health Inventory to see if invariant item ordering was indicated in any Mokken scales derived and to see if this was being influenced by extreme items. A Mokken scale was derived indicating invariant item ordering. Visual inspection of the item pairs indicated that the most difficult item (suicidal ideation) was located far from the remaining cluster of items. Removing this item lowered invariant item ordering to an unacceptable level.

Invariant item ordering (IIO) is a property of scales whereby items are scored in the same order by all respondents at all levels of the latent trait being measured (Ligtvoet, 2010). As stated by Ligtvoet (2010, p. 8) ‘IIO is a strong requirement in measurement practice’ and that researchers ‘do not realize that an ordering relationship that holds at the aggregation higher level of mean item scores does not automatically generalize to the lower level of individual subjects’ (p. 1). As such, IIO can be considered an exacting but important property of scales (Sijtsma & Junker, 1996) and the extent to which IIO holds in a set of items can be examined using methods that fall under item response theory (IRT) including parametric methods such as Rasch scaling (Meijer, Sijtsma, & Smid, 1990) and the non-parametric method of Mokken scaling analysis (MSA). IRT methods are able to relate, meaningfully, the score on a scale with the score on the latent trait as that score can be related to a specific set of items (Watson et al., 2011).

Two Mokken scales were derived from the present data incorporating 33 items in total; only one scale showing initial IIO is considered here; 8 items were excluded from both scales by the automated item selection procedure because they did not scale due to Hi < 0.30; this is below the lowerbound value and their inclusion in a scale would have lowered the overall Hs to <0.30. No items included in the scales violated monotone homogeneity. The Mokken scale considered here includes 20 items mainly related to emotional stability and is moderate (Hs = 0.44) in strength with weak IIO (HT = 0.34). The scale can be considered unidimensional, according to the criteria for Mokken scaling, because of the automated item selection procedure allocation to the scale and the fact that Hs > 0.40. Five items may be excluded on the basis that their 95% CIs included the lowerbound c = 0.30. Several Hij had 95% CI < 0 and in all cases these item pairs included an item where the 95% CIs included the lowerbound c = 0.30. In addition, two of the above items violated IIO, therefore, it is advisable to remove these items and re-analyse the data to see if the IIO improves. However, removing these items did not improve IIO. There is a sensible hierarchy of items running from very general and relatively mild feelings related to relaxation (e.g. “How much of the time, during the past month, were you able to relax without difficulty?”; “How often during the past month did you find yourself trying to calm down?”), through more specific feelings of anxiety (e.g. “How much of the time, during the past month, have you been a very nervous person?”; “During the past month, have you been anxious or worried?”), to frank expressions of depression and suicidal ideation (e.g.” Did you feel depressed during the past month?”; “How often have you felt like crying, during the past month?”; “During the past month, how often have you felt that others would be better off if you were dead?”; “During the past month, did you think about taking you own life?”). With specific reference to the topic of their paper, inspection of the item pair plots (Fig. 1) shows that item 28 (“During the past month, did you think about taking you own life?”) is located far away from all of the remaining items at the most difficult end of the scale suggesting that it may be responsible for the apparent IIO in this scale. In fact, removing item 28 and re-analysing lowers HT to 0.23, which indicates very weak IIO.