Figure 1 Probability curves for the Indonesian Hope Scale
Related Figures (8)
Table 1. Demographic Profile of Participants Table 2. Reliability of the Indonesian Hope Scale Table 3. Unidimensionality of The Indonesian Hope Scale Rating Scale Functioning The unidimensionality test evaluates the developed instrument to measure what it is supposed to measure. Table 3 demonstrates the calculation of the unidimensionality test of the Indonesian Hope Scale (IHS). Table 3 shows that the raw data variance measurement result is 39.1%. This result indicates that the unidimensionality test requirements are met (Sumintono, B., & Widhiarso, 2014). Meanwhile, Table 3 also known that the variance value moves from 3.1% to 12.8%. It means that the condition for an instrument can be used if it has a variance value below 15% fulfilled. This hope scale instrument can measure the construct of hope and estimate what should be measured (so that there is no variance outside the construct). Table 4. Calibration Scaling Analysis of IHS Table 4 shows that the increment of the Threshold index between each score is less than 1.4 logit. From category 2 to category 3 is 0.11 logit, from category 3 to category 4 is 0.9 logit, from category 4 to category 5 is 0.5 logit, from category 5 to category 6 is 0,15 logit, from category 6 to category 7 is 0.35 logit, and category 7 to category 8 is 0.35 logit). It is known that the scale with eight choices cannot be understood well by respondents, and the answer scale options must be simplified. Figure 1 illustrates the probability curves for the Indonesian Hope Scale. An assessment scale analysis was conducted to determine whether the rating scale could be understood well. If some items are less than the respondents or questionable items are answered, they must be corrected. The requirement that must be fulfilled is the Andrich Threshold index value between 1.4-5.0 logit. If the value index is less than 1.4 logit, the rating scales are combined, but the rating scales are separated if the value index is more than 5.0 (Sumintono, B., & Widhiarso, 2014). Table 4 summarizes the category structure on the gradation scale and the structure of the intersection size. Table 5. Summary Misfit Order and Item Difficulty The analysis begins by analyzing the item's difficulty by combining the mean score and standard deviation, then comparing the item measurement with the 2SD value (Boone, Yale, & Staver, 2014). Thus, the default logit should be between -0.82 and +0.82. After the analysis, there are no items outside the logit value range, which means that there are no outliers. Furthermore, the criteria used to check fit and misfit items (outliers) are to compare the Outfit MNSQ value (0.5-1.5), the Outfit ZSTD value (-2.0-+2.0), and PTMEA Corr. value (0.4-0.85). However, if the items found were the MNSQ and PTMEA Corr values, it does not meet the criteria, but if the Table 6. DIF Analysis based on Gender Figure 2. Analysis Differential Item Functioning (DIF) based on Gender Categories Analysis using the information function test is used to see the level of information exposure and the ability of respondents to complete the items in the instrument (Sumintono, B., & Widhiarso, 2014). The information generated in this test depends on the relationship between the given scale and the individual's abilities. The X- axis shows students’ ability to explain the amount of information obtained on the Y-axis. Figure 3 shows that the information level of this instrument offers a high category that is 78 points. It can be concluded that respondents get the information representing their experiences through this instrument. The findings support the aim of this study in assessing the development and validation of the HIS among ndonesian students. With the validity instrument, the researchers conclude that several items that are: it stated that it does not meet the minimum limits of the valid and reliable scale. This condition requires that some items be removed and repaired to decrease the number of items. Based on the test results, there have been described previously. Of the 25 items designed after testing, 14 items meet the overall validity and reliability requirements. Furthermore, six items were corrected based on field trial analysis and expert suggestion results, and five items were removed. Thus the total number of items used is 20 items (ten items favorable and ten items unfavorable).
