Mark Byrne 

Project Description

Previous literature suggested that there was an unfair bias present in negatively marked Multiple Choice Questions (MCQ), so upon looking into this I started to learn about some of the benefits and shortcomings of MCQs and negative marking. MCQs are popular as they produce higher grades than typical essay questions and are far easier to correct than a standard essay question. However, they have also been noted to encourage guessing and rely too much on surface-level learning. When negative marking is then applied to MCQs, they are supposed to cause less guessing and produce more reliable grades, in actuality negative marking manipulates guessing behaviour and causes some to guess and some to skip questions far more often than normal. This is what created the gap, as students that usually achieved good grades would get low marks on MCQs due to being risk-averse or risk-prone. Risk-averse students would usually skip questions more often, preferring to receive no marks rather than risk losing marks. Risk prone students typically guess more often, and risking the loss of marks. Short Answer Questions (SAQs) had less influence on guessing behaviour, and didn’t distract students with multiple options. The current study had participants take either an MCQ or an SAQ, each designed to track guesses made. The idea was that SAQs would have fewer guesses than the MCQ, and thus be a better reflection of knowledge than MCQs.

Project Outcomes

Hypothesis 1: There will be a significant difference on the participants’ guessing behaviour based on assessment type (MCQ, SAQ).

For Hypothesis 1, my research showed a significant difference. To go a bit more in-depth, the participants of my study that were assigned the SAQ assessment did not guess too often, averaging around 2 to 3 guesses per participant. Inversely, the participants that were assigned the MCQ assessment would guess far more often, usually guessing around 7 times per participant.

Hypothesis 2: There will be a significant difference on the participants’ guessing behaviour based on their risk propensity (Low, Medium, High).

For Hypothesis 2, there was no significant difference found during my studies. This means there wasn’t a large enough difference between any/all levels of risk propensity to indicate that it impacts guessing behaviour. It’s possible this result isn’t reliable however, as the majority of participants scored in the medium group for risk propensity.

Hypothesis 3: There will be a significant interaction between assessment type and risk propensity on guessing behaviour.

Hypothesis 3 was also not significant, which just means the two independent variables (assessment types and risk propensity) do not interact with each other in a way that effects the guessing behaviour of participants. To the best of my knowledge, there hasn’t been any research that investigated a possible interaction between assessment types and risk propensity, let alone that this interaction may effect guessing behaviour, so for now I can only say that my study supports a lack of interaction.

Thesis Title

Effect of Assessment Type and Risk Propensity on Guessing Behaviour