Termination Probability

Summarize when participants stop recalling.

The termination probability curve summarizes when participants stop recalling. Two modes are available: simple mode reports the raw proportion of trials that terminate at each output position, while conditional mode divides by how many trials reached that position, giving a stopping hazard rate.

Workflow

Code 
import os
import matplotlib.pyplot as plt
import warnings
from jaxcmr.analyses.termination_probability import plot_termination_probability
from jaxcmr.helpers import find_project_root, generate_trial_mask, load_data, save_figure

warnings.filterwarnings("ignore")
Code 
data_path = "data/HealeyKahana2014.h5"
figure_dir = "results/figures"
figure_str = ""
ylim = None
trial_query = "data['listtype'] == -1"
mode = "conditional"
confidence_level = 0.95
Code 
project_root = find_project_root()
figure_dir = os.path.join(project_root, figure_dir)
data_path = os.path.join(project_root, data_path)
data = load_data(data_path)
trial_mask = generate_trial_mask(data, trial_query)
Code 
plot_termination_probability(datasets=data, trial_masks=trial_mask, mode=mode, confidence_level=confidence_level)
if ylim is not None:
    for ax in plt.gcf().axes:
        ax.set_ylim(ylim)
save_figure(figure_dir, figure_str)

Interpretation

The x-axis shows output position and the y-axis shows the probability of stopping at that position. Key patterns:

  • Conditional mode: a rising hazard rate indicates increasing difficulty of retrieval as the recall sequence progresses.
  • Simple mode: the raw distribution of stopping points, which often peaks at early output positions.
  • Comparison: conditional mode is generally more informative because it accounts for how many trials are still active at each position.

API Details

Notebook parameters

  • data_path — path to an HDF5 file containing a RecallDataset.
  • figure_dir — directory for saving figures.
  • figure_str — base filename for the saved figure. Leave empty to display without saving.
  • ylim — y-axis limits as a tuple, or None for automatic scaling.
  • trial_query — a Python expression evaluated against the dataset to select trials.
  • mode — "conditional" for hazard rate or "simple" for raw termination proportion.
  • confidence_level — confidence level for subject-wise error bars.