# A tibble: 6 × 43
# Groups: participant, condsFile, trial, adjustTrip [1]
participant realOrNot condsFile trial trip adjustTrip stage.x d_time calFrame
<chr> <dbl> <chr> <dbl> <dbl> <dbl> <chr> <dbl> <dbl>
1 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16309
2 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16310
3 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16311
4 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16312
5 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16313
6 pilotv2001 1 maintena… 1 1 1 mainte… 2.72e5 16314
# ℹ 34 more variables: calTime <dbl>, tripCalFrame <dbl>, tripCalTime <dbl>,
# event_id <dbl>, device_time <dbl>, left_gaze_x <dbl>, left_gaze_y <dbl>,
# left_eye_cam_x <dbl>, left_eye_cam_y <dbl>, left_eye_cam_z <dbl>,
# left_pupil_measure1 <dbl>, left_pupil_measure1_type <dbl>,
# right_gaze_x <dbl>, right_gaze_y <dbl>, right_eye_cam_x <dbl>,
# right_eye_cam_y <dbl>, right_eye_cam_z <dbl>, right_pupil_measure1 <dbl>,
# right_pupil_measure1_type <dbl>, status <dbl>, location <dbl>, …BIO_Project_proposal_Koolhaas
1. A good description of the questions that will be asked by this project.
From our work and work from other labs, we anticipate (alongside our main effects) that there will be considerable individual variation in how much memory participants decide to use in our memory game. There are two major factors relevant to memory usage that may underlie this: differences in working memory capacity (here, the relative ease with which a certain number of list items can be encoded and maintained in memory) and differences in effort (here, the time and cognitive effort one is willing to devote toward the encoding and maintenance of list items).
If individual differences in memory are driven largely by differences in capacity, with participants exerting similar effort levels, we should see a strong relationship between study time and memory, but a relatively weak association between pupil size (a physiological measure of cognitive effort) and memory.
We hypothesize, instead, that individual differences in memory are driven largely by differences in effort, with participants coming to the task with similar memory capacity. We expect individual participants to differ in their disposition towards completing the task, such that they will tend toward either “loading up” internal memory during their first trip (heavy loaders) or balancing their memory usage out across multiple trips (light lifters).
This predicts a strong relationship between both study time and pupil size with memory. However, since both the pupil and study time measures are likely affected by effort (i.e. greater effort may increase study time to ‘load up’ on more items and, relatedly, increase pupil size as a reflection of the maintenance of that larger load), we expect that pupil size will not account for significant additional variance over that accounted for by study time.
TLDR: Is pupil size (alongside study time) a useful variable in a model describing memory usage, is it redundant with study time, or is it unique from study time yet not significantly important?
2. A general framework for the data that you will be using to answer these questions.
Here is what the data looks like:
3. Preliminary thoughts on the types of analyses you might use to approach the data. These do not have the be excessively detailed, as you still have many tools and techniques to learn!
We have preregistered doing this:
We will use hierarchical regression with linear mixed effects models with memory as the dependent variable and an intercept-only model as the null model, study time as a fixed effect in Model 1, and study time and pupil as fixed effects in Model 2. All models will have a random intercept for participant. We use study time as the first included fixed effect because of its theoretical importance for explaining memory strength (Unsworth, 2016) compared to the relatively novel literature in pupillometry as a measure of memory strength. We hypothesize that study time will be significant (Model 1 will be better than the null model), and that pupil will be significant as well (Model 2 will be better than Model 1). If Model 2 is, in fact, the best model (i.e., with pupil producing a significant increase in variance explained), we may use dominance analysis to determine whether study time or pupil is the dominant predictor. We hypothesize that study time will be the better predictor. If the null model is better than Model 1 and Model 2, then we may use an equivalence test using the ‘negligible’ package in R on study time and pupil with a SESOI of Cohen’s dz = 0.4. This SESOI was chosen based on a survey of task-evoked pupil responses that indicated this as a reasonable lower bound for a meaningful effect size (Laeng & Mathôt, 2024)
I know hierarchical/stepwise regression is looked down upon, but we have theoretical reasons to include study time as the first fixed effect (assuming the problem with stepwise regression is that people don’t always take into account that the order the fixed effects are in is important?).