Figure 6
Schematic overview of the effect course analysis for the response time semantic priming effect in Study 2. Within the experimental session, related (represented by “R”) and unrelated (represented by “U”) prime-target conditions occur randomly, n = 160 trials for each condition, resulting in a total number of n = 320 trials (step 1, T represents trial indices). For this total number of trials, nonword targets were already excluded beforehand. The trials of both conditions are extracted and ordered according to their temporal occurrence, that is the next trial within a condition represents the next occurrence of this trial in the experiment (step 2a and 2b, T represents now ordered trial indices). Subsequently, the ordered single-trial data for related and unrelated prime-target pairs (step 3a and 3b) are smoothed by calculating moving averages (step 4a and 4b). Then, the two conditions are contrasted to depict the time course of the effect, i.e. the effect course of semantic priming. Note that the temporal dimension (trial) now only reflects the per-condition available trials. As not for all subjects data of all trials are available (outliers and incorrect responses were excluded), the trials numbers for this comparison are somehow smaller than n = 160 (step 5). These differences were thresholded at each sample/trial to build a cluster, where the semantic priming effect extends over time. The obtained cluster was compared against clusters in random permutations of the data to assess statistical significance (step 6). Note that while this figure depicts the effect course of priming in Study 2, this procedure can be generalized to every experimental effect contrasting two conditions.
