Table 4 Hierarchical (multi-level) logistic regressions linking learning-phase choice frequency (FreqLearn) to transfer phase choice preference (Choice^HC/HG)

From: Action repetition biases choice in context-dependent decision-making

Dataset	Task	Term	Estimate	2.5%	97.5%	BF₁₀	Sig95
all pairs	Pooled data	b_Intercept	−2.269	−2.404	−2.137	>1000	✓
all pairs	Pooled data	b_Slope	3.879	3.653	4.114	>1000	✓
all pairs	Bavard et al.¹⁶	b_Intercept	−3.249	−3.623	−2.848	>1000	✓
all pairs	Bavard et al.¹⁶	b_Slope	5.107	4.337	5.827	>1000	✓
all pairs	Bavard et al.¹⁸	b_Intercept	−2.176	−2.316	−2.037	>1000	✓
all pairs	Bavard et al.¹⁸	b_Slope	3.682	3.444	3.920	>1000	✓
all pairs	Gaussian tasks	b_Intercept	−3.671	−4.106	−3.254	>1000	✓
all pairs	Gaussian tasks	b_Slope	8.220	7.325	9.130	>1000	✓
all pairs	Probabilistic tasks	b_Intercept	−0.841	−1.254	−0.430	>1000	✓
all pairs	Probabilistic tasks	b_Slope	2.139	1.336	2.949	>1000	✓
equal abs value	Pooled data	b_Intercept	−0.455	−0.696	−0.215	>1000	✓
equal abs value	Pooled data	b_Slope	2.349	1.937	2.774	>1000	✓
equal abs value	Bavard et al.¹⁸	b_Intercept	−0.362	−0.632	−0.099	>1000	✓
equal abs value	Bavard et al.¹⁸	b_Slope	2.244	1.809	2.701	>1000	✓
equal abs value	Gaussian tasks	b_Intercept	0.736	−0.414	1.877	0.9711
equal abs value	Gaussian tasks	b_Slope	0.695	−1.254	2.658	0.9711
equal abs value	Probabilistic tasks	b_Intercept	−0.670	−1.470	0.136	82.9645
equal abs value	Probabilistic tasks	b_Slope	2.377	0.925	3.875	82.9645	✓
equal rel value	Pooled data	b_Intercept	−0.699	−0.981	−0.419	>1000	✓
equal rel value	Pooled data	b_Slope	1.976	1.523	2.442	>1000	✓
equal rel value	Bavard et al.¹⁶	b_Intercept	−2.097	−2.790	−1.424	>1000	✓
equal rel value	Bavard et al.¹⁶	b_Slope	2.783	1.679	3.935	>1000	✓
equal rel value	Bavard et al.¹⁸	b_Intercept	−0.207	−0.522	0.113	>1000
equal rel value	Bavard et al.¹⁸	b_Slope	1.598	1.079	2.128	>1000	✓
equal rel value	Probabilistic tasks	b_Intercept	0.322	−0.868	1.568	1.2414
equal rel value	Probabilistic tasks	b_Slope	0.795	−1.030	2.653	1.2414

Posterior summaries are shown for the fixed-effects intercept (b_Intercept) and the slope of FreqLearn (the normalized frequency during learning). Estimate: Posterior mean of the regression coefficient (on the log-odds scale); 2.5% and 97.5% = lower and upper bounds of the 95% posterior credible interval (CI). BF₁₀ = Bayes factor favoring the model that contains the group and random level slope and intercept over a null model containing only group and random level intercepts. Values “>1000” indicate BF₁₀ ≥1000. A ✓ in the Sig95 column marks coefficients whose 95% CI does not include zero. Choice set indicates the subset of cross-context stimulus pairs that was analyzed (“all pairs”, “equal absolute value”, “equal relative value”). Study/Task names the experiments: “Pooled data” refers to the joint analysis across comparisons and datasets.

Back to article page

Table 4 Hierarchical (multi-level) logistic regressions linking learning-phase choice frequency (FreqLearn) to transfer phase choice preference (Choice^HC/HG)

Search

Quick links