The study was approved by the ethical committee of the Department of Psychology of the University of Amsterdam (2017-CP-7836). The tasks scripts and data are available on https://osf.io/crw7f/, where the pre-registration for the hypotheses, methods, and analyses can also be found.
Participants
A total of 165 participants (63.6% female) were recruited for this study through a university portal or through advertisements on social media. Their average age was 24.23 years (SDage = 8.44, range 18–61 years). Participants received course credits or a monetary equivalent as compensation. All participants provided consent before taking part in the study.
Seven participants did not complete the full study and were therefore excluded from data analyses. Twenty-two participants were excluded due to low target accuracy on the CIT (i.e. an error rate of 50% or more on target items, see Kleinberg & Verschuere, 2015). This criterion ensured that only those participants who understood the instructions and took the task seriously were included in the data analysis. All participants who completed the CIT had > 50% trials remaining after excluding errors and outliers (see the “Results” section below).
The final sample for analysis consisted therefore of 136 participants (64.7% female, Mage = 24.70, SDage = 9.01). Seventy-one participants (64.8% female, Mage = 25.48, SDage = 9.76) were randomly assigned to the immediate CIT condition and 65 participants (64.6% female, Mage = 23.85, SDage = 8.11) were randomly assigned to the delayed CIT condition, completing the CIT after a one-week interval. There were no significant differences between the immediate and the delayed condition in age, t (134) = 1.06, p = 0.293, dbetween = 0.18 or gender, X2 (1) = 0.00, p = 0.983, φc = 0.00.
Material
Crime scenario
Participants were told that they were going to plan a mock robbery and would work together as partners in crime. The crime scenario consisted of a coherent story based on eight critical details, of which four presented in their categorical form and four in their exemplar form: Participants encoded that they had met each other in the sports club (exemplar: volleyball club) and planned to rob a bank (exemplar: SNS bank) in their residence of South-Holland (exemplar: Delft) in May (exemplar: May 26th). Because they might not be able to flee the scene without a fight, they would bring a knife (exemplar: butterfly knife). The partners in crime plan to steal expensive jewelry (exemplar: ring) and hide it at home (exemplar: attic). Lastly, they planned to flee the crime scene by car (exemplar: Citroën; please see Additional file 1 for all possible item combinations).
RT-CIT
During the CIT all participants were explicitly instructed to conceal their knowledge of the planned robbery. Participants were required to deny knowledge for trials containing critical details from the plan (i.e. respond “no,” hence lying) while telling the truth to irrelevant items (i.e. respond “no”).
All eight critical items were presented at the categorical or exemplar level: two stimuli were encoded at the category level and were also presented at the category level (e.g. encoded as car, tested as car, congruent with encoding); two stimuli were encoded at the exemplar level and were also presented at the exemplar level (e.g. encoded as Citroën, tested as Citroën, congruent with encoding). In two other instances, the stimuli encoded at the category level were replaced by the corresponding test stimulus in its exemplar form (e.g. encoded as car, tested as Citroën; incongruent with encoding, as no exemplar-level information was made available at encoding) and two of the stimuli encoded on the exemplar level were replaced by the corresponding test stimulus in its category form (e.g. encoded as Citroën, tested as car; incongruent with encoding, as only exemplar-level information was made available at encoding; see also Fig. 1). Lastly, target items (e.g. train) were added to ensure that examinees pay attention to all items. These items have to be answered “yes” and were learned just before commencing the CIT. Targets were always presented and tested at the same abstractness level as the critical items in the RT-CIT.
Follow-up questionnaire
Motivational states were reported in a questionnaire involving five questions that participants had to rate on a 5-point Likert scale (ranging from 1 = not at all to 5 = very much so). This questionnaire measured how well participants were able to focus on the screen during the CIT, how involved they were in the study, how well their memory was for the items of the planned robbery and the learned target items, as well as how much they tried to avoid detection and appear innocent in the CIT.
Recall and recognition
To examine potential differences in memory performance between the immediate and delayed CIT condition, memory for the critical items of the planned robbery were assessed with a free recall followed by a recognition test after the full procedure. Participants first had to freely recall the eight details from the encoding phase, in which they had planned the robbery. In the subsequent recognition test, participants had to pick the correct details from a list of all eight critical items intermixed with irrelevant options (i.e. four irrelevant options per each critical detail, leading to 40 items).
For recall, the number of correctly recalled details were counted, leading to a total score between, this represents the possible range 0–8. For recognition, items were scored as either correct (1) or incorrect (0) and summed, leading to a total score between 0–8.
Procedure
Participants were invited to come to the laboratory in pairs and work as partners in crime while planning a robbery. When only one participant showed up (n = 27), the experimenter took the place of the partner in the planning phase of the mock crime. The experimenter explained that it was important to remember the details from the crime as well as possible and to visualize actually committing the robbery. In order to prevent possible detection, participants were asked to align their stories as much as possible and study the details extensively.
The experimenter first read the plan for the robbery out loud, with its eight critical details presented in either categorical or exemplar level. Participants had to encode all items by writing down the words, reading it out loud and probing each other for the information. During this encoding phase, the experimenter stayed in the room to assure that participants would not accidentally fill in an exemplar-level detail when aligning their story (e.g. inventing a specific date if they had to encode the month May). During the encoding phase, the experimenter asked the pair additional questions to stimulate richer encoding of the crime items and to contextualize the critical details (e.g. discuss how long you have known your partner in crime, and who will be driving the getaway vehicle). Then, the experimenter asked participants one by one to repeat the eight critical details from the robbery, until all items were recalled correctly. Lastly, participants filled in the missing details of the story on paper, followed by a free recall of the items. After this 10-min encoding phase, one participant of the pair was randomly assigned to the immediate testing condition, whereas the other participant was assigned to complete the CIT after a one-week delay (± 1 day). Participants were explicitly instructed not to discuss details of the experiment with each other in the one-week period between the encoding phase and the second session.
In the second phase of the experiment, participants were asked to sit behind the computer for the CIT (programmed in Inquisit 4.0). The experimenter explained that the police had received an anonymous tip about an upcoming robbery and that the participant was taken to the police station to undergo a lie detection test. The participant was instructed to try to convince the police of his innocence and beat the lie detector test by hiding all information about the crime. Upon successful concealment, the participant would receive an additional 0.5 course credit compensation.Footnote 1
Then, participants were asked to encode eight target items to which they should respond affirmatively in the CIT, while denying all other information (i.e. both the critical details and irrelevant options). The eight targets were initially presented on the screen for 2 min and participants were asked to recall all items. Then, they saw the targets for an additional 1 min before recalling them again and continuing to the practice phases of the CIT.
For each of the eight critical details encoded in the crime scenario, the CIT included the correct answer, a target item and four incorrect answers serving as irrelevant options (ratio 1:1:4). For instance, if May was the critical stimulus, the target was July and the irrelevant stimuli were June, August, September, and October (categorical Item Type); if May 26 was the critical stimulus, the target was May 30 and the irrelevant stimuli were May 8, May 12, May 17, and May 22 (exemplar Item Type). All eight critical items, eight target items, and 32 irrelevant items were displayed exactly 14 times, leading to a total of 672 trials in the test. These trials were divided over two blocks, each containing 336 trials, with a self-paced break in between. The sequence of the stimuli within the block was completely randomized, following a multiple-probe-protocol (see also Verschuere, Kleinberg, & Theocharidou, 2015).
During the test, participants had to respond to the question “Does this belong to the crime?” by pressing either the left button (A-key) for YES, or the right button (L-key) for NO on their keyboard (see Fig. 2). The question and the response keys remained on the screen during the entire test as a reminder. Participants were instructed to respond with YES only to the target items and NO to all other stimuli (i.e. both the correct details of the planned robbery and the irrelevant options). Each trial consisted of one answer (e.g. June) being displayed as a word in the middle of the screen for exactly 1500 ms. If the participant did not respond within the maximum response deadline of 800 ms, the message TOO SLOW appeared in red above the stimulus for 200 ms. If the participant’s response was incorrect, that is responding with NO for target items or with YES to critical or irrelevant items, the word WRONG appeared in red below the stimulus for 200 ms. Response latency was measured from the onset of the stimulus on the screen until one of the response keys was pressed. After key-press or after the 1500 ms presentation time, the next stimulus appeared on the screen with an inter-stimulus interval (ISI) of either 250, 500, or 750 ms to prevent response preparation.
In order to ensure proper understanding of the task and instructions, each participant had to pass through a stepwise practice procedure that allowed participants to become used to the speed and the requirements of the CIT. Each of the three practice phases of the memory detection test consisted of 24 trials displaying a random subset of critical, irrelevant, and target items. In the first practice phase, participants could pace the speed of the trial sequence themselves, so that a new stimulus only appeared after a key press. Feedback was given upon an erroneous response (i.e. false denying recognition of the target items or falsely claiming recognition of the critical and irrelevant items), but the TOO SLOW message was not presented. Participants could proceed to the next phase when their target accuracy was at least 50%, otherwise the first practice phase was repeated until this requirement was met. In the second phase, the 1500 ms stimulus presentation was added, so that the next trial would automatically appear upon key press or after 1500 ms. Again, feedback was given upon an erroneous response, but the TOO SLOW message was never presented. Participants could only proceed to the next phase when their target accuracy was at least 50% and as an additional requirement, when their mean response latency was < 800 ms, otherwise this practice phase was repeated until their performance was satisfactory. The last practice phase was identical to the full test, including the WRONG and TOO SLOW feedback. Participants could proceed to the actual test only when their target accuracy was at least 50% and when their mean response latency was < 800 ms.
After completing the full CIT procedure consisting of 672 trials, participants were presented with a questionnaire designed to assess their attention to the tasks, involvement in the experiment, memory for the stimuli, and their motivation to avoid detection in the CIT on a 5-point Likert scale (ranging from 1 = not at all to 5 = very much so). Then, participants were told that the experiment was finished and they did not have to hide any information anymore. Finally, all participants completed the recall and recognition questionnaire assessing their memory of the robbery’s details, before being debriefed and compensated for participation with research credits.