Two MANOVA tests were carried out in order to ascertain both the relative contribution
of each variable to pupils’ attitudes and their various combined ‘effects’. First, a three-
way teachers’ background4 (3) x gender (2) x scales (4) ANOVA on attitude scores was
carried out. This revealed main effects of teachers’ background (F (2, 417) = 11.02, p < .
0001) and gender (F (1, 418) = 30.01, p < .0001). Then a four-way ANOVA test was
used: teachers’ background (3) x perceived ability (2) x gender (2) x scales (3). In this
test the confidence subscale is not used because the variable perceived ability was created
by this subscale. The MANOVA test showed that there were significant effects of the
variables of gender and teachers’ background, already shown from the first MANOVA
test, and an effect of the variable perceived ability (F (1, 418) = 129.34, p < .0001). The
second MANOVA test also showed interactions between the variables a) perceived
ability and gender (F (1, 408) = 5.82, p < .02), and b) perceived ability and teachers’
background (F (2, 408) = 3.26, p < .05). Next, tests that explain the effects of the
variables teachers’ background, gender, perceived ability and the interactions of them
with other variables are presented.
Teachers’ background
Pupils taught by art specialists (named as the AS group) received higher mean scores on
all four subscales, followed by pupils taught by enthusiastic non-specialists (named as
the ENS group) and then by pupils taught by unenthusiastic non-specialists (named as the
UNS group).
Table 2
4 The word ‘background’ is used as a shortcut to the phrase ‘specialisation and attitudes to art/art teaching’.
15