I'm planning to do a survey with students (age around 16) with questions/answer-possibilities that have been 'proven' to be significant indicators of student success/good grades. So, I'm not really interested in the relationship of the variables to the outcome, instead, I want to feed the data of the survey to a machine learning algorithm to check the algorithm's correctness.

Instead of coming up with the questions of the survey myself, I would like to ground them academically, i.e.: based on academic literature. I don't need very 'original' ones, the significance of the variables themselves is not the focus of my research (actually, I want 'proven' significant variables).

Best case scenario would be to find a canonical list of survey questions and answer-possibilities which have been 'proven' to be influential (significant) indicators of student success/grades, from which I can take the ones appropriate for my research.

Another thing that would be help me is a source which discusses good and bad practices for setting up a survey such as mine (which might also point me to the questions I want to ask).

I have found some papers, such as (Sykes & Kuyper, 2009) about predicting student success/grades, from which I can pick some variables/questions, but none of them were close enough to my case to be fully satisfactory for my purpose. If anyone can help me with a good paper that investigates roughly the same as my research, this would also help, as I might then base my survey (partly) on this research.

  • 1
    $\begingroup$ Adding some links to the papers you have read would be helpful. (: $\endgroup$
    – Seanny123
    Commented Feb 11, 2016 at 14:41
  • 1
    $\begingroup$ Some types of IQ tests are commonly used to predict academic achievement with some success - is this the sort of thing you had in mind? $\endgroup$
    – Arnon Weinberg
    Commented Feb 17, 2016 at 5:58
  • $\begingroup$ Thanks for your answer. I'm more looking for 'simple' features like motivation (what kind of questions would measure this in a good way for example?) and personalia (which have been 'known' to influence performance?). I'm aiming for a relatively simple survey, purely to get some data on the features to feed to the ML algorithm. An IQ test for example, is too complicated for my purpose. I want to do the survey in approx. 15 mins. $\endgroup$
    – matthijs
    Commented Feb 17, 2016 at 8:52

1 Answer 1


Both Carol Dweck and Angela Duckworth have fairly brief tests meant to measure student success and by some extension grades.

Dweck's work focuses on how having a "fixed mindset" limits a students ability to grow in a subject and is used to explain why students are so apt to give up when met with challenge. Her quiz would have to be formatted for pen and paper and her research is rather sparse on the subject.

Duckworth focuses on both grit and self-control. She has researched heavily the effects of grit over the last 10 years while Walter Mischel works on self-control and has been for the past 40 years. Duckworth was Mischels student. Unfortunately, his test for self-control would likely be too abstract for your needs.

Needless to say her work has been shown to measure student success over the last ten years. Her website offers several quizzes on grit or self control for adults and students. The same links provides access to her publications.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.