The objective of the present study was to further examine the differential impact over time of single-sex and coeducational school environment on public high school girls’ motivation in two basic academic domains: language Arts and mathematics.

Many authors claimed that a single-sex high school environment is more favourable than a coeducational one for girls' development of positive attitudes toward learning and for the development of their self-concept ((Bornholt and Möller, 2003; Lawrie, & Brown, 1992; Lee, & Bryk, 1986; Lee, & Lockheed, 1990; Watson, Quatman and Edler, 2002). Other researchers have suggested that, when students and school background factors are adequately controlled, there are no academic advantages of single-sex schools for girls (Daly, 1996; Dollison, 1998; Harker, 2000; Lepore, & Warren, 1997; Marsh, 1989). The current lack of uniformity in the results of studies suggests that further investigation is needed on the extent to which single-sex schools can affect girls’ self-confidence and achievement.

Method

Participants from eight coeducational high schools and from two single-sex high schools were followed during a period of three school years in a longitudinal research scheme. Participants, 340 high school girls, were separated in two cohorts. Those in the first cohort (n = 154) participated in the study between their 7^th and 9^th grades while those in the second cohort (n = 186) were followed between 9^th and 11^th grade. Some were under investigation in mathematics, others in language arts.

Data were collected with a self-reported questionnaire including several scales: Parental support (language arts’ alpha = .79/mathematics’ alpha = .74), Teachers’ support (a = .81/.83), Competence beliefs (a = .85/.84), Utility value (a = .85/.80), Mastery-Approach Goals (a = .86/.86), Performance-Approach (a= .76/.77) and Work-Avoidance (a = .73/.71). Participants had to indicate their agreement on a five point Likert-type scale ranging from «Strongly disagree» (1) to «Strongly agree» (5). Data were collected four times: beginning of 7^th and 8^th grades, end of 8^th grade and 9^th grade for the first cohort and beginning of 9^th and 10^th grades, end of 10^th and 11^th grades for the second cohort.

Findings

Because of the restrained number of participating schools in this study, analyses of repeated measures appeared to be a better option than multi-level analyses (Raudenbush & Bryk, 2002). Mixed design repeated measures analyses of variance were conducted according to a factorial design 2 (coeducational or single-sex environment) x 4 (Time of measurement). Each analysis was conducted separately for language arts and for mathematics for both cohorts.

Cohort 1 (grade 7 thru grade 9)

Language Arts

The multivariate test (Pillai’s Trace) indicated a significant main effect of the environment, F_{(2, 63)} = 6.95, p = .002, but no significant effect of the interaction between factors, F_{(6, 59)} = 1.69, p = .14. However, the subsequent univariate analyses (Bonferroni correction applied) did not reach the level of significance. Data on competence beliefs, utility, and achievement goals were examined in another multivariate test. The omnibus test showed no significant effect of the environment, F_{(5, 59)} = 1.66, p = .16, nor for the interaction, F_{(15, 49)} = 1.45, p = .16.

Mathematics

The multivariate analyses on the perception of support from social agents indicated non significant results for the type of environment, F_{(2, 85)} = 2.38, p = .10, but a significant interaction effect, F_{(6, 81)} = 2.33, p = .04. Nevertheless, the adjusted level of significance was not reached. The analyses of the motivational variables revealed no effect of the environment, F_{(5, 89)} = 2.15, p = .07, nor for the interaction, F_{(15, 79)} = .79, p = .13.

Cohort 2 (grade 9 thru grade 11)

Language Arts

            The multivariate analyses indicated a significant simple effect of the environment on the support from social agents F_{(2, 87)} = 4.22, p = .028. However, the interaction effect did not reach significance, F_{(6, 83)} =.54, p = .78. The univariate analyses showed that the participants who attended a single-sex school received more support from their parents. The multivariate analyses conducted on motivation indicated a non significant effect of the environment, F_{(5, 88)} = 1.01, p = .42,, nor of the interaction, F_{(15, 78)} =.86, p = .61.

Mathematics

  Analyses indicated no significant effect of the environment, F_{(2, 93)} = 1.17, p = .32, nor of the interaction, F_{(6, 89)} = .69, p = .66, on support from social agents. The multivariate analyses revealed once again a non significant effect of environment, F_{(5, 94)} = 1.53, p = .19, and of he interaction, F_{(15, 84)} =.70, p = .78, on motivational variables.

Overall, significant time effects on several variables, for both cohorts, indicated a decline of achievement motivation towards mathematics and language arts over time.

Theoretical and educational significance