Davide Farchione

Participation in Asynchronous Online Discussion Forums Does Improve Student Learning of Gross Anatomy.

Asynchronous online discussion forums are common in blended learning models and are popular with students. A previous report has suggested that participation in these forums may assist student learning in a gross anatomy subject but it was unclear as to whether more academically able students post more often or whether participation led to improved learning outcomes. This study used a path model to analyze the contribution of forum participation, previous academic ability, and student campus of enrolment to final marks in a multicampus gross anatomy course for physiotherapy students. The course has a substantial online learning management system (LMS) that incorporates asynchronous forums as a learning tool, particularly to answer learning objectives. Students were encouraged to post new threads and answer queries in threads started by others. The forums were moderated weekly by staff. Discussion forums were the most used feature of the LMS site with 31,920 hits. Forty‐eight percent of the students posted at least once with 186 threads initiated by students and a total of 608 posts. The total number of posts made a significant direct contribution to final mark (P = 0.008) as did previous academic ability (P = 0.002). Although campus did not contribute to final mark, there was a trend for students at the campus where the course coordinator was situated to post more often than those at the other campus (P = 0.073). These results indicate that asynchronous online discussion forums can be an effective tool for improving student learning outcomes as evidenced by final marks in gross anatomy teaching. Anat Sci Educ. 7: 71–76.

Do collaborative practical tests encourage student-centered active learning of gross anatomy?

Benefits of collaborative testing have been identified in many disciplines. This study sought to determine whether collaborative practical tests encouraged active learning of anatomy. A gross anatomy course included a collaborative component in four practical tests. Two hundred and seven students initially completed the test as individuals and then worked as a team to complete the same test again immediately afterwards. The relationship between mean individual, team, and difference (between team and individual) test scores to overall performance on the final examination (representing overall learning in the course) was examined using regression analysis. The overall mark in the course increased by 9% with a decreased failure rate. There was a strong relationship between individual score and final examination mark (P < 0.001) but no relationship for team score (P = 0.095). A longitudinal analysis showed that the test difference scores increased after Test 1 which may be indicative of social loafing and this was confirmed by a significant negative relationship between difference score on Test 4 (indicating a weaker student) and final examination mark (P < 0.001). It appeared that for this cohort, there was little peer‐to‐peer learning occurring during the collaborative testing and that weaker students gained the benefit from team marks without significant active learning taking place. This negative outcome may be due to insufficient encouragement of the active learning strategies that were expected to occur during the collaborative testing process. An improved understanding of the efficacy of collaborative assessment could be achieved through the inclusion of questionnaire based data to allow a better interpretation of learning outcomes. Anat Sci Educ 9: 231–237.

Conditional assessment of the impact of a Hausman pretest on confidence intervals

In the analysis of clustered and longitudinal data, which includes a covariate that varies both between and within clusters, a Hausman pretest is commonly used to decide whether subsequent inference is made using the linear random intercept model or the fixed effects model. We assess the effect of this pretest on the coverage probability and expected length of a confidence interval for the slope, conditional on the observed values of the covariate. This assessment has the advantages that it (i) relates to the values of this covariate at hand, (ii) is valid irrespective of how this covariate is generated, (iii) uses exact finite sample results, and (iv) results in an assessment that is determined by the values of this covariate and only two unknown parameters. For two real data sets, our conditional analysis shows that the confidence interval constructed after a Hausman pretest should not be used.

Frequentist and Bayesian Interval Estimators for the Normal Mean

It is well known that a Bayesian credible interval for a parameter of interest is derived from a prior distribution that appropriately describes the prior information. However, it is less well known that there exists a frequentist approach developed by Pratt (1961) that also utilizes prior information in the construction of frequentist confidence intervals. This frequentist approach produces confidence intervals that have minimum weighted average expected length, averaged according to some weight function that appropriately describes the prior information. We begin with a simple model as a starting point in comparing these two distinct procedures in interval estimation. Consider X 1,…, X n that are independent and identically N(μ, σ2) distributed random variables, where σ2 is known, and the parameter of interest is μ. Suppose also that previous experience with similar data sets and/or specific background and expert opinion suggest that μ = 0. Our aim is to: (a) develop two types of Bayesian 1 − α credible intervals for μ, derived from an appropriate prior cumulative distribution function F(μ) more importantly; (b) compare these Bayesian 1 − α credible intervals for μ to the frequentist 1 − α confidence interval for μ derived from Pratts frequentist approach, in which the weight function corresponds to the prior cumulative distribution function F(μ). We show that the endpoints of the Bayesian 1 − α credible intervals for μ are very different to the endpoints of the frequentist 1 − α confidence interval for μ, when the prior information strongly suggests that μ = 0 and the data supports the uncertain prior information about μ. In addition, we assess the performance of these intervals by analyzing their coverage probability properties and expected lengths.