%0 Generic %1 Mujtaba2016Analyzing %A Mujtaba, Abid H. %D 2016 %K teaching %T Analyzing the relative efficacy of student assessment techniques using Conditional Probability %U http://arxiv.org/abs/1611.06583 %X Academics often attempt to analyze problems in pedagogy on the basis of anecdotes when they should be using an evidence based, data driven approach. This paper presents a relatively simple technique for analyzing the relative efficacy of different types of questions when it comes to judging the conceptual understanding of students. The technique is illustrated using a case-study in which a carefully constructed multi-part question (with binary, descriptive, and computational parts) was given to a group of students. The responses were graded and assigned a boolean value to denote success or failure. The boolean values were counted to calculate the empirical probability of success in each type of question and correlations between the different types was analyzed by calculating conditional probabilities. The analysis revealed that while success in answering the descriptive question guaranteed success in the other two types the converse was far from true. Binary and computational questions were revealed to be poor indicators of conceptual competence while conditional probability turns out to be an excellent tool for determining the relative efficacy of different types of questions.

@misc{Mujtaba2016Analyzing, abstract = {{Academics often attempt to analyze problems in pedagogy on the basis of anecdotes when they should be using an evidence based, data driven approach. This paper presents a relatively simple technique for analyzing the relative efficacy of different types of questions when it comes to judging the conceptual understanding of students. The technique is illustrated using a case-study in which a carefully constructed multi-part question (with binary, descriptive, and computational parts) was given to a group of students. The responses were graded and assigned a boolean value to denote success or failure. The boolean values were counted to calculate the empirical probability of success in each type of question and correlations between the different types was analyzed by calculating conditional probabilities. The analysis revealed that while success in answering the descriptive question guaranteed success in the other two types the converse was far from true. Binary and computational questions were revealed to be poor indicators of conceptual competence while conditional probability turns out to be an excellent tool for determining the relative efficacy of different types of questions.}}, added-at = {2019-02-23T22:09:48.000+0100}, archiveprefix = {arXiv}, author = {Mujtaba, Abid H.}, biburl = {https://www.bibsonomy.org/bibtex/22cc27cb68a070e34a82ade256cad9b54/cmcneile}, citeulike-article-id = {14212840}, citeulike-linkout-0 = {http://arxiv.org/abs/1611.06583}, citeulike-linkout-1 = {http://arxiv.org/pdf/1611.06583}, day = 20, eprint = {1611.06583}, interhash = {b55002fc31fdcf3a92720790acaf0d31}, intrahash = {2cc27cb68a070e34a82ade256cad9b54}, keywords = {teaching}, month = nov, posted-at = {2016-11-22 10:39:37}, priority = {2}, timestamp = {2019-02-23T22:15:27.000+0100}, title = {{Analyzing the relative efficacy of student assessment techniques using Conditional Probability}}, url = {http://arxiv.org/abs/1611.06583}, year = 2016 }