Reliability analysis of data obtained by the Concerns-Based Adoption Model (CBAM) Stages of Concern Questionnaire (SOCQ) on the PEARL Reading Program Monitoring project

The main aim of this report was to record findings of reliability and validity analyses performed on Pacific Early Age Readiness and Learning (PEARl) data obtained from a monitoring team using the Concerns Based Adoption Model(CBAM) Stages of Concern Questionnaire (SOCQ)in primary schools in the Kingdom of Tonga. We found that alpha coefficients were marginal (ranging from 0.6. to 0.7), while both discriminant validity coefficients and convergent validity coefficients did not meet the required levels. Inter-item correlations between items in sub-scales were too high, undermining discriminant validity, while within-scale convergent validity coefficients were too low, suggesting that these items were not measuring the stage-related constructs as predicted by the theory (Weiner, 2007). On these grounds, we recommended changes and eventual discarding of the data-collection tool. The project was supported by the Global Partnership in Education (GPID), with technical assistance from the World Bank. This was a background report lodged with the World Bank