Causality Studies

One of the main objectives of the school is to enhance children’s affective development and increase their academic performance and achievement. However, the question here is how students’ academic achievement can be increased. Since the late 1960s in America, much research has been done on self-esteem (self-concept) and academic achievement. As a result it is believed that to increase students’ academic achievement and performance, enhancing students’ self-concept is the best approach and consequently, schools have a responsibility in this area. Since then a great deal of research has been done, mainly in the USA, and some in Britain and Turkey, on the effects of either self-concept or self-esteem upon academic achievement and, as can be seen in the review section, a considerable number of these studies have found a positive relationship between the two variables. During the twentieth century, in industrialised countries, improvement of students’ academic performance has been a principal aim of education. However, ideas on how best to achieve this are constantly changing; whether to put the emphasis on better teaching skills or to put it on enhancing students’ self-esteem, in order to improve academic performance.

- 67 -

The controversy about whether to concentrate on basic skills or on the enhancement of self-esteem is still being debated today. In the search for better achievement, researchers and educational psychologists have mainly joined one of these two groups. One group claims that the school curriculum should be designed to enhance students’ self-esteem in order to improve academic achievement (Self-Enhancement Model), while the other group suggests that intervention programs should employ procedures that will help children to develop basic academic skills (Basic-Skills Model).

According to the self-enhancement theorists, self-concept is an important variable influencing both students’ performance and academic achievement. They also argue that the self-esteem variables are the primary causes of academic achievement. Therefore, they argue that in order to promote better achievement, self-esteem must be changed first. Accordingly these theorists and researchers emphasise that considerable time and effort should be spent in trying to enhance the self-esteem of students by means of the school curriculum. Specially designed self-enhancement intervention programs to help children develop a general feeling of self-regard have been designed to promote the feeling that they are good and worthy people. In contrast to the self-enhancement model, the basic-skills model defends the structured teaching of specific skills to improve academic achievement. Basic-skill theorists make no explicit effort to increase the self-esteem of their students. For them, however, self-self-esteem is the consequence of achievement rather than the cause.

In the 1960s and 1970s several educators reported that self-esteem caused academic achievement. Reviewing the studies on the relationship between self-esteem and academic achievement, Purkey (1970) claims that self-esteem is causally prior to academic achievement. He reports that: “it gives us reason to assume that enhancing the self-concept is a vital influence in improving academic performance” (p. 27). Irwin (1967) also makes a similar conclusion. According to his findings he suggested that “a

- 68 -

positive conception of one’s self as a person is not only more important than striving to get ahead, and enthusiasm for studying and going to school, but that it is a central factor when considering optimal scholastic performance” (p. 271). Hansford & Hattie (1982) in their meta-analysis of 128 studies, which examined the relationship between various self-measures of achievement and performance, indicated that 95% of the studies they reviewed used zero-order product moment correlation coefficients. It has to be pointed out that the statements of Purkey (1970) and Irvin (1967) on the causality of the relationship between self-esteem and academic achievement have also been primarily based on correlational data. Pottebaum et al. (1986) point out “any studies of self-esteem and achievement have used designs that showed a single significant correlation and imply causation is a well-known error in such methodology” (p. 141). They maintained that neither correlation nor simple ANOVA provide any information about the direction of possible causation: whether it is from self-esteem to achievement or from achievement to self-esteem. In her comprehensive review of the self-esteem literature, Byrne (1984) suggested that studies which address the direction of the causality question follow these prerequisites: 1) a statistical relation must be established, 2) longitudinal designs should be used with a clearly established time precedence, and 3) a causal model must be tested.

To find out the causality direction, several studies have been specifically designed.

Some studies indicate that academic achievement is causally related to self-esteem (Calsyn & Kenny, 1977; Newman, 1984; Byrne, 1986; Rosenberg et al., 1989). The findings of these studies support the skill development theory in which self-esteem is thought to be a consequence, rather than a cause, of achievement. In his investigation to find out the causal relationship between self-concept of ability and academic achievement, Calsyn & Kenny (1977) examined the data of 556 adolescents and used cross-lagged panel correlation methods. His findings were more supportive of a skill development model than a self-enhancement model. GPA was causally predominant

- 69 -

over both self-concept of ability and perceived evaluations of others, and these patterns of causal predominance were much stronger in females than males. The study of Newman (1984) also supported the findings of Calsyn & Kenny (1977). Newman’s study revealed that maths achievement predicted maths self-concept across grades 2, 5, and 10. His findings indicated that maths achievement was causally related to maths self-concept, although self-concept was not in the position of predicting academic achievement. Rosenberg et al., (1989) also report similar findings. A panel of 1886 adolescent boys was used to explore the reciprocal relationship between self-esteem and school performance. The results indicated that grades had a decidedly stronger positive effect on self-esteem than self-esteem had on grades. The significant effect of school marks on self-esteem, in this research, lends support to the skill development model. As a result, these studies provided evidence that academic achievement is causally related to self-esteem and supported the skill development model.

In comparison, the findings of other research support the self-enhancement theory in which self-esteem is assumed to determine academic achievement (Bridgeman &

Shipman, 1978; Shavelson & Bolus, 1982; Song & Hattie, 1984; Marsh, 1990; Owens, 1991). One of the first empirical results showing causality of self-concept variables over academic achievement is a study of Shavelson & Bolus (1982). In order to examine the self-enhancement theory, the authors tested ninety nine seventh and eight grade junior high school students in February, 1980. In June of the same school year they used cross-lagged panel analysis to examine the causal predominance of self-concept and achievement. The results of this study indicated the causal predominance of self-concept over academic achievement. Marsh (1990) in his investigation examined more than 1400 students’ academic self-concept and GPA scores at four specific times. He found that GPA of students in grades 11 and 12 were significantly affected by academic self-concept measured the previous year, while grades had no effect on subsequent measures of academic self-concept. The results obtained by Marsh (1990) are in

- 70 -

opposition to the findings of Newman (1984). In a recent study, Owens (1991) investigated the direction of the causal relationship between global self-esteem and academic achievement, and between academic self-concept and academic achievement.

The study attempted to determine the direction of this relationship through use of the cross-lagged panel correlation technique, applied to a longitudinal study of community college students during the fall and spring semesters of their freshmen year. The results suggest that even though global self-esteem is not causally predominant in academic achievement, academic self-concept is causally predominant in academic achievement.

Another consideration is that self-esteem and academic achievement effect each other in a reciprocal manner. Some studies consider this approach (Chapman et al. 1981; Marsh, 1984; Skaalvik & Hagtvet, 1990); Liu, Kaplan, & Risser, 1992; Kurtz-Costes &

Schneider, 1994). The study of Chapman et al., (1981) reveals evidence for a reciprocal relationship between self-esteem and achievement across a 12-month interval. They used cross-lagged panel analyses, and obtained correlations ranging from 0.41 to 0.52 between self-esteem and achievement and vice versa for two groups in grades three and four, and five and six. Skaalvik & Hagtvet (1990), in their investigation among grades three and six, obtained two different results. They found that achievement measured in grade three affected self-concept measured one and half years later in grade four. They reported a reciprocal causal relationship operating from grade four to grade six between achievement and self-concept for the older group. The authors concluded that the initial causal predominance of achievement over self-concept diminished over time, leading to reciprocal relationship between the two concepts. The research of Liu, Kaplan, & Risser (1992) was another study which showed a reciprocal relationship between self-esteem and academic achievement. The authors examined general esteem, academic self-concept, and GPA score of 242 students, ranging from grade 7 to 12. The results of this study confirmed that general self-esteem both influences and was influenced by academic achievement. In another recent study Kurtz-Costes & Schneider (1994)

- 71 -

investigated the same problem. In this study forty-six children’s academic self-concept and achievement scores were measured at the ages of 8 and 10 but they obtained mixed results. Their results overall neither supported the skills development theory nor the self-enhancement theory. The results of this investigation indicated that self-concept at Time 1 predicted achievement at Time 2, and that achievement at Time 1 predicted concept at Time 2. These results supported a bidirectional relationship between self-concept and achievement.

Apart from these studies and approaches, some studies have suggested that general self-esteem and achievement may not be causally related (Maruyama, Rubin, & Kingsbury, 1981). Pottebaum, et al., (1986) in their investigation to determine the presence and direction of the causal relation between self-esteem and academic achievement, used a cross-lagged panel correlational technique applied to a longitudinal study of U.S. high school students. The results of their study suggest that there is no significant causal relation between self-esteem and academic achievement, but rather that the observed relation is the result of one or more uncontrolled and unknown third variables. Using path-analytic models, Watkins & Gutierrez (1990) examined self-concept, attributions, and grades of 194 Filipino high school students. The authors obtained no direct path between grades and academic self-concept, the relationship between which, was mediated by attributions for success through ability and effort.