Predicting introductory programming performance: A multi-institutional multivariate study

Abstract

A model for predicting student performance on introductory programming modules is presented. The model uses attributes identified in a study carried out at four third-level institutions in the Republic of Ireland. Four instruments were used to collect the data and over 25 attributes were examined. A data reduction technique was applied and a logistic regression model using 10-fold stratified cross validation was developed. The model used three attributes: Leaving Certificate Mathematics result (final mathematics examination at second level), number of hours playing computer games while taking the module and programming self-esteem. Prediction success was significant with 80% of students correctly classified. The model also works well on a per-institution level. A discussion on the implications of the model is provided and future work is outlined.

Predicting Performance in an Introductory Programming Course by Logging and Analyzing Student Programming Behavior

“…However, it is important to consider how our algorithm performs, in terms of accuracy and explanatory power over the duration of a course. Interestingly, previous work [2][3][4][5][6][7][8][9] [ [13][14][15][16], used all available data to drive their predictive models. But predicting a student's failure at the end of a course leaves little time for an instructor intervention.…”

“…Unlike prior work [2][3][4][5][6][7][8][9] which mainly used indirect criteria to predict performance, our approach is based upon analyzing directly logged, quantitative data describing aspects of a student's ordinary programming behavior. This allows prediction of performance to evolve over timereflecting changes in the student's learning progress without the need to use multiple tests that often yield inconsistent results.…”

“…Given potentially high enrollment numbers, the use of tests to gather predictive data can take a considerable amount of time for an instructor to process. Even if a test was indicative of performance, by the time it was processed, it may be too late for students to withdraw, or for instructors to intervene to prevent students from failing [7]. The criteria used for prediction is the main limitation of prior studies.…”

“…As programming became more widespread, researchers began to explore a greater range of cognitive [3], psychological [4], and demographic [5] predictors. Researchers over the past two decades (1990-2010) extended prior work by exploring similar factors [6] [7] and the predictive potential of new innovations in pedagogy [8] [9].…”

Teaching the novice programmer: A study of instructional sequences and perception

“…These factors have been shown by Bergin and Reilly (2006) to be important predictors of success in programming courses. Table 1 shows the demographics for the sample as a whole and for each learning sequence.…”