Abstract
Concern regarding the prevalence in primary school teachers of poor attitudes towards and low achievement in mathematics has received some empirical support. But the common assumptions (i) that attitudes towards, and achievement in, mathematics are substantially related, and (ii) that teachers' attitudes and achievement in mathematics affect pupils' attitudes and achievement, await empirical verification. While studies relating teachers' attitudes toward pupils' attitudes and achievements are sparse, those relating attitude and achievement within teachers or within pupils have typically shown a low positive relationship not always reaching statistical significance. Some writers use this as evidence to discount the importance of attitudes in achievement, but the present paper contends that a partial explanation for the lack of the expected result may arise from the measurement assumption that attitude to mathematics is a unidimensional phenomenon. Two multidimensional attitude instruments were constructed and administered to 317 final year prospective primary school teachers. Although the two instruments employed widely differing techniques for tapping attitudes to mathematics and mathematics teaching, dimensions from both instruments showed substantial correlations with mathematics achievement. The findings gave support for the contention that attitude to mathematics is not a unidimensional phenomenon. The consistency of findings within and between both attitude instruments suggests a good validation of each. The contention that teachers' attitudes (and achievement) affect pupil attitudes and achievement is yet to be tested.