Metadata
| Creators: | Schofield, Andrew  ORCID: https://orcid.org/0000-0002-0589-4678 |
||||
|---|---|---|---|---|---|
| Keywords: | psychophysical data, height judgements | ||||
| Divisions: | College of Health and Life Sciences > Cognition & Neuroscience Research Group (CNRG) | ||||
| Data Publisher: | Aston University | ||||
| Date Made Available: | 6 July 2022 | ||||
| Collection period: |
|
||||
| Data Collection and Processing Methods: | Raw data represent data from individual trials as collected by PsychoPy with no processing. Fits show collated data for each condition and sub-conditions together with the model parameters and model fits. Psychometric functions were fit with a logistic function of the form, p=L+(U-L)/(1+e^(-k(s-s_0 ) )) where L is the lower limit (range 0 to 0.1), U is the upper limit (range 0.9 to 1), k is the slope parameter (range 0 to 100), s is the ground truth height of the variable step relative to the reference in the range +-64 pixels, and s0 is the Table 1 of s where the logistic is halfway between U and L (range +-64 pixels). Allowing the upper and lower limits to vary allows for participant lapses and the lapse rate of 10% was chosen based on there being only 10 trials for each level tested. Equation 6 was used to calculate least sum of squared errors fits to the data for each individual and condition using the ‘fmin_tnc’ function (truncated Newton gradient descent) from the SciPy Python library. To avoid local minima the model fits were repeated 1000 times for each psychometric function with different starting values and the solution with the lowest sum of squared errors was chosen. Summary files show the estimated Points of Subjective Equality based on the above model fits measures in pixels or as percentage of the standard height. Pcent = 100*pixels/164 Data in summary files were calculated from the fits to a different precision than is available from the fit files and any attempt to calculate PSEs from the fit files should compare the two to 6dp only. | ||||
| Identification Number: | https://doi.org/10.17036/researchdata.aston.ac.uk.00000559 |
Schofield, Andrew (2022) Second-order texture gratings produce overestimation of height in depictions of rectangles and steps. [Dataset] Aston University. Available from: https://doi.org/10.17036/researchdata.aston.ac.uk.00000559
Download
![[img]](https://astondr-prod.leaf.cosector.com/style/images/fileicons/archive.png)