The aims of this study were to develop and validate elementary school students’ attitudes toward assessment questionnaire (SATA-Q) and to explore the structural relationships among students’ attitudes toward assessment, the incremental belief of intelligence, the perceived teachers’ formative feedback, behavioral intention, and classroom engagement in either formative or summative assessment situations. In Study 1, the developed items were classified based on sub-factors of the SATA and the psychometric properties of the SATA-Q items were reviewed with the reliability and validity analyses. In Study 2, the multi-level SEM approach was used to identify the structural relationships among the SATA and its related variables using Mplus.
The SATA-Q comprises four sub-factors through the theoretical frameworks based on the Expectancy-Value model (Atkinson, 1958; Eccles & Wigfield, 2002; Fishbein, 1975) and the Control?Value model (Pekrun, 2006). Expectation from and value attributed to the assessment were included in the assessment of cognitive attitude; positive and negative affect were included for assessing the affective attitude. Furthermore, the two sets of SATA-Q were developed for the assessment situations such as formative and summative assessments. The initial item pool comprised items from previous research about students'' or teachers'' conceptions of assessment or affective experiences in the assessment questionnaire. From this pool, 21 items were selected and reviewed by the experts and criteria of semantic redundancy. To validate the SATA-Q, 165 fifth and sixth graders of elementary schools were participants in the pilot survey and 610 fifth and sixth graders were participants in the main survey.
The results of the survey are summarized as follows:
First, SATA-Q with 12 items was accepted for both types of assessment(formative and summative assessment). For assessing the cognitive attitude, the expectation factor has 4 items including importance, necessity, usefulness, and improvement learning; the value factor has 2 items regarding expectancy from the assessment result. To assess the affective attitudes, positive affect factor had 2 items with regard to enjoyment of learning and challenging through assessment while the negative affect factor had 4 items on anxiety, pressure, shame, and hopelessness.
Second, the criterion-related validity of the questionnaire was verified by analyzing the correlations of the SATA with beliefs about intelligence and achievement-goal orientation. Expectancy, value, and positive affect had a positive correlation with the mastery/performance approach of goal orientation and incremental beliefs about intelligence. However, negative affect showed a positive correlation with mastery/performance avoidance and performance-approach goal orientation, and the entity beliefs about intelligence.
Third, before comparing the latent means of SATA across the different types of assessment, I examined the measurement equivalence of SATA by a Longitudinal Confirmatory Factor analysis (LCFA) using Mplus. From the results, factor variance equivalence was found to be partially supported. Analysis of the latent mean differences across the types of assessment revealed that students had higher levels of expectancy from and value for summative assessment than for formative assessment. However, students had higher levels of positive affect toward formative assessment than summative assessment.
Fourth, to explore the structural relationship among SATA variables, incremental belief of intelligence, recognition of teachers’ formative feedback, behavioral intention, and classroom engagement, the goodness-of-fit of the measurement model was examined and direct and indirect effects and mediating effects of the variables were assessed using multi-level SEM. The findings revealed that incremental belief of intelligence and formative feedback had a statistically significant positive effect on expectancy and value and an indirect effect on learners’ participation through value, positive affect, and behavioral intention. Specifically, formative feedback had a positive direct effect on behavioral intention and learners’ participation.
The educational implications based on the results of this study are as follows.
First, the SATA-Q was developed based on theoretical considerations as well as empirical data from elementary school students. This questionnaire is expected to help teachers understand students’ attitudes toward classroom assessment.
Second, teachers’ formative feedback and students’ incremental beliefs of intelligence are variables with a positive influence on SATA. Therefore, teachers need to provide formative feedback on students’ learning regardless of the type of assessment and help students develop incremental beliefs of intelligence in their learning.
Third, students ascribe a greater value to and have better expectations from summative assessment rather than formative assessment. They consider summative assessment results to be more important. Although the mean of value and expectation in formative assessment was low, value showed the greatest positive effect on positive affect while expectation reduced negative affect in formative assessment. Therefore, to help improve students’ learning, the role of formative assessment is important. There is a need to emphasize that teachers should use student-centered assessment to consider students’ individual differences in learning and enhance students’ learning by providing formative feedback.