Appendix 5: Statistical reliability of the SNSA for academic year 2017 to 2018
To evaluate whether a set of items is psychometrically sound, it is necessary that each item discriminates effectively between learners who have the knowledge or skill that underpins the item and those who have not, and that the set of items as a whole measures a unified latent trait (for example, numeracy). One key summary statistic that is usually used to indicate whether items in an assessment are working well as a set is a reliability statistic, such as Cronbach’s Alpha. According to this metric, reliabilities above 0.75 are typically considered as satisfactory and above 0.80 as excellent.
Table 8 shows Cronbach’s Alpha reliability statistics for the 11 SNSA administered in the academic year 2017 to 2018.
Table 8: Assessments’ reliability results
The summary of the results presented in Table 8 shows that the reliabilities of all assessment forms were excellent (more than 0.80).