[back to main ratings page ]
Through correlation coefficient analysis, it is possible to isolate the goodness of fit between the scores assigned in any single evaluation criterion and overall evaluation scores. The scatter charts below illustrate how such results explain which factors our scorers perceived as most important in determining whether a question should receive an overall high score. It is notable that those questions that received the highest “demanding” scores tended to be the questions that received the highest overall scores, suggesting that “demanding” questions play well in this format. Conversely, audiovisual quality was the least powerful predictor of overall high scores, suggesting that slick production may be a less important ingredient of a high quality YouTube question in this format.

Here, R2 represents the proportion of variation in overall score that can be attributed to the “demanding” factor. In other words, the data show that clarity ratings accounted for 81.3% of variance in overall scores - our scorers perceived that “demanding” was the most important variable in determing the overall quality of questions.

Here, R2 represents the proportion of variation in overall score that can be attributed to the “interesting” factor. In other words, the data show that “interesting” ratings accounted for 79.6% of variance in overall scores - our scorers perceived that “interesting” was the second most important variable in determing the overall quality of questions.

Here, R2 represents the proportion of variation in overall score that can be attributed to clarity. In other words, the data show that clarity ratings accounted for 69.5% of variance in overall scores - our scorers perceived that clarity was the third most important variable in determing the overall quality of questions.

Here, R2 represents the proportion of variation in overall score that can be attributed to the assessments of audiovisual quality. In other words, the data show that “audiovisual” ratings accounted for 53% of variance in overall scores - our scorers perceived that audiovisual quality was the least most important variable in determing the overall quality of questions.

Data compilation: Kurt Zemlicka
Data analysis: Gordon Mitchell
Web presentation: Delphine Masse