Statistical prediction of the future impairs episodic encoding of the present

The following submission contains the behavioral data reported from the
manuscript "Statistical prediction of the future impairs episodic
encoding of the present". Data from four experiments are included.
All experiments contained a learning phase, in which participants
viewed a series of trial-unique scenes drawn from one of 12 scene
categories and were asked to report whether or not the on-screen scene
contained a manmade object. Accuracy and response time were measured
during this task. Each scene category belonged to one of three conditions:
A scenes, which were predictive of B scenes (which were thus predictable),
and X scenes which could be preceded or followed by any number of scene
categories and thus were neither predictive nor predictable. All
experiments also contained a memory phase, in which participants were
shown a scene image and were asked to indicate whether or not that
particular scene image was presented in the learning phase. All scenes
from the learning phase, in addition to a subset of images that were not
presented during the learning phase (foils) were presented during the
memory phase. After indicating whether an image was old or new,
participants then indicated their confidence in their response
(Experiments 1a, 1b & 3) or when in time they remember
seeing the image (Experiment 2; old responses only). Experiments 1a
& 2 additionally contained a category pair test. On each trial,
participants were presented with two pairs of images and were
asked to indicate which pair felt more familiar to them. Trials consisted
either of a true A/B pair and a scrambled A/B pair (for which a trial
would be accurate if participants chose the true A/B pair), or two
dummy-coded X/X pairs (to which there was no objectively correct
answer).