My supervisor and I agreed on the final design while discussing the results of the pilot experiment. We both think that it is important to replicate Lew & Howe (2016) results if their test is used, but not if one uses 3AFC test. They used a Y/N location recognition task (“Was this object in this place before?”) and location recall task (2D picture version). I will develop a spatial 3D location recall task for the VR system to replicate the results in the first experiment. I will also use a 3AFC test to show that it is bias at retrieval that is responsible for the congruency effect that they report not a bias at encoding as they argue. Lew & Howe (2016) found that object/location memory benefited from schema-expectancy but not from schema-unexpectancy. I will also let the participants rate the general expectancy and not just object/location pairs at very end of the study. Therefore general outline of the experiment will look like this: encoding (VR) -> n-back -> recall (VR) -> AFC -> ratings.
- Testing u-shaped relationship by testing slopes (unexpected > neutral and neutral < expected) with paired t-test.
- Logistic regression
- DV: accuracy (correct vs. incorrect) and confidence (knew vs. not)
- Random slope and intercept
- Random effect: post encoding ratings.
- Fixed effect: pre-encoding ratings