Skip to main content
. 2015 May 29;11(5):e1004301. doi: 10.1371/journal.pcbi.1004301

Table 2. Model family comparison: Relative log-group Bayes factors.

Relative log-group Bayes factors (smaller is better)
Family 1 Family 2 Family 3
Moments without pstarve Rank-dependent utility Moments and pstarve
Model Model Model Model Model Model Model Model Model
1 2 3 4 5 6 7 8 9
EV EV EV EV Prelec-I Prelec-II EV EV EV
Var Skw Var pstarve Var Var
Skw pstarve Skw
pstarve
All 0 -1333 -1984 -2032 -2154 -2148 -2191 -2167 -2126
Foraging 0 -884 -1390 -1437 -1466 -1390 -1511 -1532 -1569
Casino 0 -709 -945 -967 -1040 -1023 -1062 -1044 -970
Foraging-block 1 0 -472 -683 -696 -734 -679 -781 -751 -741
Foraging-block 2 0 -412 -661 -729 -725 -689 -751 -737 -773
Casino-block 1 0 -363 -451 -456 -488 -465 -518 -487 -439
Casino-block 2 0 -342 -519 -507 -525 -463 -565 -537 -487

For a fixed-effects analysis, log-group Bayes factors based on BIC were calculated relative to the simplest model (Model 1). Smaller log-group Bayes factors indicate more evidence for the respective model versus the baseline model. The log-group Bayes factors of the winning models according to fixed-effects analyses are written in bold font. The models included free parameters for the respective variables listed. BIC, Bayesian information criterion; EV, expected value; Var, variance; Skw, skewness; pstarve starvation probability