Changes in physiological reactivity were estimated on the basis of problem-solving activities and personal characteristics via multilevel regression analyses.
To analyze non-hierarchically structured data, cross-classified multilevel regression analyses are appropriate (Snijders and Bosker, 1999).
The results of multilevel regression analyses supported this interpretation by showing that math competence was a stronger predictor of problem solving competence.
The mixed-method design combines a multilevel regression analyses (of data from 1255 teachers in 67 schools) with focus group discussions amongst teachers and headmasters in ten schools.
Affective factors (i.e., reading attitude, reading motivation, and reading self-concept), peer bullying, home context (i.e., home SES), and school context (i.e., school SES, school bullying, and school safety and order) were used for our later multilevel regression analyses.
The multilevel regression analyses required the calculation of student population estimates within each country and so the authors followed the guidelines set forth in the TIMSS documentation and applied the student sampling weight (TOTWGT) at the student level.
Logistic and multilevel regression analyses showed that tree survival, growth and thinning response (change of tree growth due to a thinning treatment) were functions of tree species, size, age, removed and remaining competition, as well as time since the treatment.
Multilevel regression analyses were conducted to compare changes in the determinants between the TLVM group and the TL group, and between the two intervention groups and the control group.
Multilevel regression analyses were performed to assess the hypothesized relationships.
Table 3 also reports results from the multilevel regression analyses.
We performed separate multilevel regression analyses for each indicator.
