5th and 8th grade pupils’ and teachers’ perceptions of the relationships between teaching methods, classroom ethos, and positive affective attitudes towards learning mathematics in Japan

78

to what extent that statement reflects the self. Marsh (1988, 1990a, 1990b) noted that
this complicated response format not only posed difficulties for young children but also
led to negative item effects, because of the low validity of responses to negatively
worded items. SDQ-I and SDQ-II employ Likert scales. The Likert scales measure
respondents’ attitudes from the total scores for a series of items (Oppenhein, 1996). This
item format was simpler than the PCS. The Likert scales provide more precise
information about the respondents’ degrees of agreement or disagreement than when
simple agree∕disagree responses are used. One disadvantage of measurement with the
Likert scale is the lack of reproducibility, whereby the same total score may be obtained
in many different ways (Oppenheim, 1996). However, this did not cause problems for
this study, because the study required an ordering of the participants with regard to their
self-assessment using self-concepts and mathematics concepts rather than distinct
differences in their perceptions of individual items as in the Likert scales.

The advantage of making use of standardised assessment is that it can provide a high
level of reliability and validity (Robson, 1993). For instance, SDQ-I and SDQ-II clearly
support the multifaceted, hierarchical conceptualisation of the self-concept in
Shavelson’s (1976) model: other self-concept instruments fail to do so (Marsh, 1988,
1990a, 1990b). In this sense, the construct validity of SDQ-I and SDQ-II, in terms of
within-network studies, is supported. The results of SDQ-I and SDQ-II support the
construct validity of these tests in terms of between-network studies, in relation to the U-
shape of self-concepts on age and academic achievement indicators. The internal
consistency coefficient alpha has been reported as .89 for the Mathematics Scale and
.81 for the General-Self Scale in SDQ-I, and as .90 and .88, respectively in SDQ-II.
Stability over time has also been reported as generally high, although not as high as the
internal consistency estimates. This is due to the systematic change in the students’ self-
concept over time (Marsh, 1988, 1990a, 1990b). Attempts were made to keep high
reliability and validity in the standardised instructions in the process of translation. All
instruments were translated into Japanese by the researcher. Afterwards, another
Japanese researcher back translated them into English for verification of the translation.
This process was repeated until compatibility was achieved in the back-translated
English version. This process is popularly adopted in studies applying instruments
devised in one language for use in another language to make them accessible to the
participants and also to ensure reliability and validity (e.g. Ho, et al., 2000).

78

<<   89   90   91   92   93   94   95   >>

More intriguing information