Contextual influence on confidence judgments in human reinforcement learning
Table 2
Free parameters. ABSOLUTE, absolute value learning model; RELATIVE, relative value learning model (best-fitting model); LL optimization, parameters obtained when minimizing the negative log likelihood; LPP optimization, parameters obtained when minimizing the negative log of the posterior probability. The table summarizes for each model the likelihood maximizing (best) parameters averaged across subjects. Data are expressed as mean±s.e.m. The values retrieved from the LPP optimization procedure are those used to generate the variable used in the confidence glme models.