How accurate are teacher and parent judgements of lower secondary school children’s test anxiety?

Abstract

This study examined teacher and mother judgement accuracy of children’s worries and emotionality in the school subjects of German and mathematics. The participants were 59 German language and 58 mathematics teachers, 572 mothers and their children at the end of Grade 6. The analyses for the total sample revealed weak-to-moderate correlations between adult’s and children’s perception of test anxiety. Moreover, we found that teachers overestimated children’s emotionality as well as children’s worry, whereas mothers underestimated children’s worry and overestimated children’s emotionality in both subjects. However, when looking at the highly test-anxious children, we found quite low correlations and an underestimation of children’s worry and emotionality in both school subjects. Thus, it seems important to provide teachers and parents with information about test anxiety and possibly training to improve their judgement accuracy because of moderating effects of test anxiety on achievement.

Keywords:

• judgement accuracy
• test anxiety
• parents
• teachers

Notes

1. Some studies term it ‘physiological hyperarousal’ (e.g. Lowe et al., 2008).

2. To evaluate the degree of measurement and structural invariance of test anxiety for girls and boys, several two-group CFAs (n = 257 boys and n = 315 girls) were computed. The first two-group CFA for the subjects of German as well as mathematics examined the two-factor-structure (worry and emotionality) for boys and girls (baseline models). These two-factor models provided a good fit to the data (German: χ²(16) = 43.81, RMSEA = 0.055, CFI = 0.984 and NFI = 0.975; mathematics: χ²(16) = 66.38, RMSEA = 0.074, CFI = 0.955 and NFI = 0.943). Next, several two-group CFAs with equality constraints were computed. The analyses for test anxiety in the school subject of German indicated that the factor loadings (∆χ²(4) = 6.66, p > .05), the intercepts (∆χ²(6) = 7.56, p > .05), the factor variance (∆χ²(2) = 0.27, p > .05), the covariance (∆χ²(1) = 1.63, p > .05) and the latent factor means (∆χ²(2) = 3.67, p > .05) were invariant for boys and girls. For mathematics, we found that the factor loadings (∆χ²(4) = 3.32, p > .05), the factor variance (∆χ²(2) = 1.23, p > .05) and the covariance (∆χ²(1) = 0.02, p > .05) were invariant in the two groups. However, the intercepts (∆χ²(6) = 35.64, p < .05) and the latent factor means (∆χ²(2) = 25.74, p < .05) were not group-invariant. The girls had a higher mean worry and mean emotionality in mathematics than the boys (boys mean was fixed 0 and girls estimated mean worry: M = 0.46, p < .001; mean emotionality: M = 0.27, p < .01).

3. Students’ grades are reversed: 6 is the best one, 1 is the worst one.