 |
Asia-Pacific Forum
on Science Learning and Teaching, Volume 12, Issue 2, Article 1 (Dec., 2011) |
A questionnaire was designed to assess children's perceptual thought, and their conceptual understanding of sharks and its ecological significance. About 16 items of perceptual and conceptual-based issues have been developed for the asessment. The demographic variables including gender, age, academic performance and the ways of acquiring environmental information were investigated. The context of the questionnaires was developed on the basis of the curriculum framework of science education covering in the Strand 2, addressing the aims at promoting students' concern for the environment and equipping their basic understanding of nature and the relationships between the environment and human. The learning objectives 'to recognize the interdependence of living things and their environment', 'to respect and care for all living things' and 'to understand people's responsibility of environmental conservation and recognize the importance of environmental conservation' (Curriculum Development Council, 2011) for key stage 2 (Primary 4-6 levels) were specifically highlighted.
The questionnaire comprised two main parts. The first part included questions about the perceptions of sharks and cognitive understanding of ecological issues. The five-point Likert-type scale (Likert, 1932) has been adopted to measure the perceptual and conceptual issues of the children participants. The Likert scale applied in this study was adjusted (in the range from point 5 = 'very agree' to point 1 = 'very disagree' in positive statements and the 5 to 1 point scores were reversed for negative statements; the score for 'uncertain' response was set at 3). The participants were allowed to select only one option from a pre-defined list of each conceptual item. Full mark reflected the participant having the most positive perception towards the issues and the highest confidence to give the correct answer of the conceptual items. The inclusion of 'uncertain' response was adopted to minimize the random trial. The participants were allowed to select more than one options from multiple non-conflicting options in liberal multiple choice items for denoting extra-dimension of choices made by the participants (Ng & Chan, 2009). In the second part, the demographic information, the acquired information sources, the reading preference and habit were investigated. Five options ranging from 'less than 1 hr/day' to 'more than 1 hr/day' were provided to assess the reading habits of the participants.
The issues adopted in the questionnaire have been pilot tested for validity, precision and reliability. Instrument content validity has been qualitatively evaluated by five expert reviewers. The reviewers are the academic and teaching scholars of the relevant expertise in science education, chemistry, biology and environmental science. The Content Validity Index (CVI) of all the 16 items were 1 whereas the the Kappa coefficient for all these items achieved 0.88, reflecting the strong agreement about the relevance of this set of items to the instrument among the experts. External validity has been checked by 8 children randomly selected from a third primary school (excluding from the studied schools) and there was no pre-determined factor applied for the selection so that the participants chosen fell in the range between 7-11 of the age. The results were analyzed and the participants were interviewed for giving comments of the questionnaire. After the interview, minor adjustments have been made to ensure the participants clearly understood the questions in the intended ways (see Summers et al., 2000). The internal consistency was assessed by the item statistics from each question. In the reliability analysis, a Cronbach's alpha coefficient α = 0.739 revealed that the measuring tool was reliable.
The study was conducted between November and December 2010. A total of 140 participants (55 male, 39%; 85 female, 61%) within the age range of 8 to 10 year-old (mean = 8.99, SD = 0.42) from two local primary schools (79, 56% from School A; 61, 44% from School B) participated in the study. The schools were randomly selected from different regions of Hong Kong (one in Tai Po and another one in Diamond Hill). Two classes were randomly selected from primary 4 level of each school and all pupils of these two classes were investigated. One-lesson time (more than half an hour) was allowed for the participants to complete the questionnaire.
The data of perceptual thought and conceptual performance of the participants were rated by using the 5-point Likert scale measure. The performance was presented in both the frequency and total score independently, yet most of the data analyses were based on the score values. The overall performance of the participants was presented by the frequency distribution of the scores. The descriptive statistics were calculated by using a software PASW Statistics 18.0.0 (SPSS, Inc.). The mean scores for each conceptual item or in grouped realm (perceptions and conceptions) were analyzed based on corresponding analytical methods provided by the previously mentioned software, including F-test of equality of variances, Student t-test, variance analysis (e.g. 2-factor ANOVA, Tukey's range test and MANOVA), PCA, linear regression and correlation analysis. Tukey's range test was performed as a post-test in analysis of variance. Pearson's correlation coefficient and the Spearman's rank correlation coefficient were calculated to determine the strength of relationship between two variables. Pearson's correlation analysis was performed under the assumptions that the data were normally distributed and measured on an interval or ratio scale while the Spearman's rank correlation analysis was based on the ordinal rank of the data and no assumption of variables approximating multivariate normal distribution. Both methods are mathematically similar but different in ways of measurement and thus a more comprehensive view on the strength of the association between the variables were obtained from both analyses. PCA was conducted on the variance of the scores in the different variables of acquired information sources. All variable variance were reorganized and presented in new groups of orthogonal components (equal to the number of original sources) which represented the dominant trends in the data set. Most noise falling into the later principal components would be discarded by the reductive nature of PCA. The statistical significance for all tests was set at p < 0.05.
