Keywords

Our analysis of the PIRLS-2011 data of 41 countries was guided by the analytic framework (Table 3.1), which was based on the general framework (Table 2.1) we used to categorize the studies included in the literature review. The analytic framework matches the available indicators of parental involvement in the PIRLS-2011 data (Table 3.2).

Table 3.1 Analytic framework for parental involvement for secondary analyses of PIRLS-2011
Full size table
Table 3.2 Overview of components and items for secondary analyses of PIRLS-2011
Full size table

The first dimension, i.e. home-based involvement from the perspective of parents, was split into two components or indicators: early literacy activities and help with homework. The early literacy activities component is especially well measured by the PIRLS home questionnaire. In the international reports of PIRLS-2011, early literacy activities is the only component reported as a scale, with Cronbach’s alphas ranging from 0.70 (Czech Republic, Hungary, Italy and Oman) to 0.88 (Romania), indicating high reliability (Martin and Mullis 2012). Although the international report does not report the scale statistics on the items regarding parental help with homework, this component is well addressed, as a total of eight items ask about these practices.

To consider the dimension of school-based involvement and home-school communication from the parent’s perspective (component 3 in the analytic framework), three relevant items were selected from the home questionnaire. The number of items for this indicator is low, but the items do seem highly relevant to this context. The student’s perception of parental involvement and the school’s practices on parental involvement (corresponding to components 4 and 5 in the analytic framework) are measured by five items in the student questionnaire and 15 items in the school questionnaire, respectively. Finally, we established the outcome variable and control variables for the structural multilevel model tested later (Table 3.2): the PIRLS reading literacy achievement scores, gender of the student, and, as approximates for SES, both books at home and highest level of education of the parents.

The benefit of using data from a large-scale assessment study such as PIRLS is not only the richness in data resulting from achievement tests, as well as student, home, teacher and school questionnaires, but also, obviously, the large number of countries for which these data are available. For this study, we initially considered the data from 43 countries participating in PIRLS-2011; countries not meeting the required response rate or for which the average achievement was not reliably measured were excluded. However, two countries, England and the USA, did not administer the home questionnaire and were therefore not included in the scale analyses for components 1–3, using items from the home questionnaire.