Page 127 - BUKU SYNOPSIS
P. 127

criteria in the table ((Sumintono & Widhiarso, 2015). However, if one


                                    of the three criteria is met, the item can be considered fit (Chan et al.,

                                    2014; Rahman, 2023; Yudiarso, 2023).



                                    Table 6. Standard Criteria for Item Fit of Rasch Model
                                       Items                                   Standard Criteria
                                       Outfit MNSQ                             0.5 – 1,5
                                       Outfit ZSTD                             -2.0 – (+) 2.0
                                       Point Measure Correlation (PT Mean      0.4 – 0.85
                                       Corr)




                                      3)  Scoring Quality by the Raters


                                      Two raters rated the students' writing scripts. Hence, the inter-rater


                                      reliability coefficient estimation for the assessment model in writing

                                      script  analysis  is  conducted  using  the  Intraclass  Correlation

                                      Coefficient (ICC). This approach is employed to evaluate the level


                                      of  agreement  among  raters  while  providing  an  assessment.  The

                                      findings of the calculation of the reliability coefficient demonstrate


                                      a  contrast  between  the  fluctuations  induced  by  the  measured

                                      qualities and the overall fluctuations in measurement (Fleiss, 1975;


                                      Widiarso, n.d.). The Intra-Class Correlation (ICC) is a statistical

                                      measure that spans from 0, indicating no agreement, to +1, indicating


                                      perfect agreement. The reliability coefficient between raters in this

                                      study was calculated using the SPSS version 29 software, with a


                                      minimum  accepted  agreement  level  of  0.70.  Koo  and  Li  (2016)








                                                              122
   122   123   124   125   126   127   128   129   130   131   132