Caleon, Imelda S.

The impact of two cryogenics-based enrichment programmes (CBEPs) held in out-of-school setting on the learning of science concepts and attitudes towards science was evaluated using both quantitative and qualitative methodologies. The findings presented in this study are based on a sample of 531 Primary Five and Six students from six schools in Singapore who had their class enrichment lessons at the Snow City.

In general, students reported large significant learning gains immediately and two-weeks after their CBEP experience. The learning gains in relation to the CBEP involving liquid nitrogen remained stable after around two weeks but those for the participants of the CBEP involving liquid oxygen slightly waned. About 93% of the participants of the former and 88% of the latter reported cognitive test scores above pretest level around two weeks after their respective CBEP experience. Greater learning gains due to participation in either CBEP were recorded for EM1 than for EM2 students. CBEPs did not have any differential impact on the learning gains of boys and girls.

Regardless of gender and stream, CBEPs produced fairly significant improvements in the attitudes towards science of the participants, specifically in relation to enjoyment of science and motivation to pursue science careers; however, no significant change was detected in the participants' perceptions of the social implications of science. Although satistical tests revealed that the retention of the said attitudinal gains did not occur for most if the students who took retention test, 13 to 20% of such students were found with at least a three-point improvement in scores in the given attitude areas. A substantial proportion of the participants shifted from negative or neutral responses to positive responses for at least two items in both enjoyment of science and career preference subscales two weeks after participating in CBEP. The contents of such items point to the possible "ripple effects" of CEBPs in the participants' future activities that could eventually lead to better achievement and more stable attitude change.

Learning gains derived from CBEPs were found to be basically determined by academic ability while attitudinal gains were predominantly influenced by initial attitude towards science. EM1 students have greater propensity to learn more from CBEPs than EM2 students. These who have initially more positively attitude towards science tend to maintain or improve such level of attitude after experiencing CBEPs. Gender did not have any significant in determining the cognitive and affective benefits that students derived from CBEPs.

Additionally , positive association bewteen post-CBEP attitude towards science and learning of cryogenics-related concepts was detected. The causal direction of the association appears to be from attitude towards cognitive: meaning positive attitude leads to greater learning gains and not the other way around.

At least 92% of the subjects liked the CBEPs covered in this study and expressed willingness to attend similar programmes in the future. The students expressed recognition of both entertainment and educational value of the programmes. The CBEPs appeared slightly more appealing for females than males for EM1 than EM2 subjects of the study. The complexity of the topic and the degree of students participation are two factors that are of importance in relation to the likeability of a particular CBEP.

This study served as an appraisal of Snow City's CBEPs, as well as a pioneering work on the evaluation of cryogenics-based programmes. Its results have provided empirical evidence that Snow City can be venue for a novel learning experience that can generate positive cognitive and effective impact on students. The findings of this study have shown that CBEPs can benefit the majority of its participants, at least for upper primary students, regardless of gender and academics ability.

This study reports on the development, validation and application of a four-tier multiplechoice (4TMC) diagnostic instrument, which has not been reported in the literature. Each 4TMC item has a stem that is followed by the content tier, reason tier, and confidence tiers. The content tier and reason tier measure a respondent’s content knowledge and explanatory knowledge, respectively. The confidence tiers separately measure a respondent’s confidence in the correctness of his or response for the content tier and reason tier. The 4TMC test focused on waves, and was accordingly named the Wave Diagnostic Instrument (WADI). Using several indicators, the reliability of WADI was found to be low to moderate for the content and reason tiers, and moderate to high for the confidence tiers. The validity of inferences about students’ conceptions on waves that were derived from scores on WADI was fair to moderate.

The participants of the study were 931 upper secondary students from the Express and Special stream of 11 co-educational mainstream government schools. They took WADI after they were formally instructed on waves. Mean scores and mean confidence for the content tier were higher than those for the reason tier. The vast majority of the respondents were found to have an inadequate grasp of the topics tested. The students’ mean confidence was slightly above the neutral level. The students who have higher scores in WADI tended to have higher confidence levels (r=.28 to .35, p<.0001), and to have higher academic achievement (r=.28 to .43, p<.001). The students’ academic achievement was found to be weakly correlated with their confidence and confidence bias. The confidence levels of the males were statistically higher than those of the females (t> 4.61, p<.0001), but the test scores of both gender groups were found to be comparable. Confidence and test performance tended to increase when students are familiar with (i.e., formally instructed about) the concepts tested (t>2.95, p<.004).

The students were generally overconfident (i.e., their mean confidence level was beyond what was warranted by the accuracy of their responses), with males tending to be more overconfident than the females. The ability of students to discriminate between what they know and what they do not know, in terms of confidence, was low and was not significantly affected by their academic achievement and gender.

Fifty-eight alternative conceptions (ACs) were expressed by at least 10% of the sample, of which 24 were espoused confidently. Thirteen ACs were applied by more than 50% of the sample at least once. Seven ACs were expressed with high confidence, with the highest confidence being associated with ACs about the role of air in sound propagation and the graphical representation of waves. Six of the nine ACs that were associated with more than one item of WADI were found to be consistently applied by at least 10% of the sample; of these six ACs, two were consistently applied with confidence by at least 10% of the sample.

The students’ confidence ratings obtained using the 4TMC version of WADI was found to be statistically higher than those obtained using a content parallel three-tier version (which requires one confidence rating for both content tier and reason tier responses) of WADI (t=2.83, p<.05).

In the past two decades, the field of Machine Learning has not only greatly expanded in terms of the plethora of increasingly powerful modeling frameworks it has provided, but has also birthed the applied fields of Educational Data Mining and Learning Analytics. Learning Analytics has blossomed as an area in the Learning Sciences, promising impact for various stakeholders working at different educational levels, such as Instructional Designers, Students, Instructors, Policymakers and Administrators. This symposium offers a taste of cutting edge work across each of these levels, with a common emphasis on analytics applied to social processes.