Page 181 - BUKU SYNOPSIS

from one rater to the next is a crucial reliability aspect in rating scale

development. Knoch suggests a number of different types of reliability

estimates and factors contributing to reliability were measured as part of

the facets analysis by analyzing the rater measure report. The assessment

of rater consistency in MFRM analysis can be observed through the fit

values, particularly the outfit values, produced. The character of the

raters can be observed through the fitness values produced by the raters.

Table 21. Rater Measure Report

The rater measurement table in Table 21 demonstrates that the inter rater

measurement findings are acceptable because all score meet the criteria

of fitness (infit MnSq = 0.90 and 1.05, outfit Mnsq = .93 and 1.03, outfit

ZStd = -.9 and (.4), and point measurement correlation values of 0.72 and

.68) since they meet the criteria for “outfit MnSq = 0,5 – 1,5, ZStd = -2 -

+2, and point measurement correlation = 0,4 - 85). Aside from that, the

study results reveal a "separation" score of 4.77, signifying "good." A

decent separation score, according to standards, is a score of 3. A

reliability score of 0.96 is regarded satisfactory. Meanwhile, the

percentage of agreement among raters is 62.5%, which is close to 45.2%.

Good inter-rater agreements, according to Sumintono (2022) the range is

176