Intraclass correlation coefficient or commonly called ICC is the ratio between variance between groups and total variance which is used for the ratio between variance between groups and total variance. The ICC coefficient value can be calculated manually using the formula, but it requires a long process. To simplify the testing can be done using SPSS software. The following is an example of ICC reliability testing using SPSS:
An assessment sheet for assessing the question/instrument grid with 10 indicators was given to 3 experts and each indicator was assessed for suitability of the indicators with the grid, suitability of answers and suitability of language. The result is as below:Test step:
Open the SPSS software
Create Variables by clicking Variable view and entering each expert as a variable, because there are 3 assessors, create 3 variables with the names A1, A2, and A3. Set Decimal to 0. Name labels A1 as Expert 1, A2 as Expert 2, and A3 as Expert 3
Enter the assessment results data by clicking the data view
Click Analyze-Scale-Reliability Analysis
Enter all the variables in the item column by clicking variable then clicking the arrow
the final result
Click button statistics
Select F Test and tick on ICC
Click Continue, then click OK, the test results will output
SPSS Calculation Results
Conclusion
Based on the SPSS calculation results, it can be seen that the ICC value is 0.699. This value is then compared with the ICC score criteria.
So it can be concluded that the grids are assessed by experts Fair or reliable.
No comments:
Post a Comment