Demo 2: Detailed Choice File → Choice File

This demo aggregates repeated comparisons from a detailed choice file into a choice file with the judgments aggregated.

Run the demo

cd rs-software
python -m src.rs_py.demos.demo_combined_choices.py

Inputs

Prompt	What the parameter is
Path to detailed choices .mat file	Output file produced by Demo 1 - detailed choices
Output directory	Directory where the choice file will be written
Experiment/paradigm name	Used when naming output files
Subject ID	Used when naming output files

To use the sample data included with the repository, enter:

0

for all prompts.

Example terminal output

Combining trial wise judgments.
  Input detailed .mat: /path/to/image_detailed_choices_S4.mat
  Output dir:         /path/to/output
  Exp name:           image
  Subject:            S4

Writing choice file in three-column format (ref, s1, s2).
Saved results to /path/to/output/image_choices_S4.mat

Done.

Output file

image_choices_S4.mat

Contents of the output file

The file contains:

metadata
responses_colnames
responses

The metadata field is carried over from Demo 1.

responses_colnames

ref
s1
s2
N(D(ref, s1) > D(ref, s2))
N_Repeats(D(ref, s1) > D(ref, s2))

responses

For the sample dataset:

5994 rows × 5 columns

Unlike the detailed choice file, there is no trial column. Repeated occurrences of the same comparison have been combined into a single row.

Example:

ref   s1   s2   N(D(ref,s1) > D(ref,s2))   N_Repeats(D(ref,s1) > D(ref,s2))
16     3   13              5                              5

Using metadata.stim_list:

3  = bear
13 = elephant
16 = giraffe

This row corresponds to:

(ref, s1, s2)
(giraffe, bear, elephant)

The final two columns indicate that this comparison appeared 5 times in the experiment and the same outcome was observed on all 5 occasions.

Compared to the detailed choice file, many repeated comparisons have been merged, reducing the number of rows from 31,080 to 5,994.

Next step

The output of this demo becomes the input to Demo 3, which fits geometric models to the similarity judgments.