Experiment Authors: Dr Muhammad Hijazy, Dr Laura Smith, and Dr Katherine Button from the Department of Psychology at the University of Bath, UK
The E-Prime experiment that we designed and built for our research is inspired by the Social Evaluation Learning Task described in Button et al. (2015), which in turn was developed from the Self-Referential Learning Task in Button et al. (2012). However, since our research is still underway, we took the decision not to submit our actual E-Prime experiment, but a modified version that closely resembles the original Social Evaluation Learning Task explained in Button et al.’s (2015) paper. We advise familiarising yourself with the Social Evaluation Learning Task before looking at our E-Prime experiment.
The experiment presents the research subjects with 32 pairs of words (negative vs positive) which appear in a random order from a pre-designed list of word pairs. The subject will be asked to choose from one of two options consisting of one negative word and one positive word. A feedback page appears every time the subject chooses a word to advise whether the chosen word is actually the correct one. The feedback presented to the subject is based on the probabilities stipulated in each different session; as explained below:
- In Session no. 1, when the subjects meet a computer persona called ‘Charlie’, 80% of the negative words will be correct, while 20% of the positive words will be correct. 32 word pairs will be presented to the subjects in this session.
- In session no. 2, when the subjects meet a computer persona called ‘Alex’, 20% of the negative words will be correct, while 80% of the positive words will be correct. 32 word pairs will be presented to the subjects in this session.
- In Session no. 3, when the subjects meet a computer persona called ‘Jordan’, 20% of the negative words will be correct, while 80% of the positive words will be correct. Only 21 word pairs will be presented to the subjects in this session. This session is designed to reverse any unwanted learning which the subjects may have acquired from session no.1. Only subjects who have done session no. 2 first and then session no. 1 would be required to do this session at the end.
To distribute the correct and incorrect answers randomly but evenly throughout the 32 word pairs, an array of 32 elements is created and divided into six blocks. One False value is allocated randomly inside each one of these blocks (6 elements ≈ 20% of 32), while the rest of the array’s elements (26 elements ≈ 80% of 32) are filled with True values. Not all the blocks are equal in size; six elements are included in the first and last blocks, while only five elements are included in each one of the other four blocks.
After all the 32 pairs are presented, the subject will be asked to rate how much the computer persona likes or dislikes him/her. The subject will have the ability to choose only one value on a scale of 100 = Like and 0 = Dislike. Again, please read Button et al.’s (2015) paper for a detailed explanation of the Social Evaluation Learning Task.
Counterbalance by Subject:
1. Type of feedback: the experiment divides the subjects into two equal groups based on the type of feedback that they receive: Those who have an odd subject number will receive ‘socially validating feedback’, while those who have an even subject number will receive ‘neutral feedback’.
2. Location of words on the screen: the experiment divides the subjects into two equal groups based on where they see the words on the screen (in WordsSlide): the words in each side of the word pairs will be presented inside the top-left corner box to half of the subjects, while they will be presented inside the bottom-right corner box to the other half of the subjects.
Counterbalance by Session:
This will be done manually by using the session number. Half of the subjects will do session no. 1 first and then session no. 2; while the other half will do session no.2 first, session no. 1 second and session no.3 last (it is explained above why and when session no. 3 should be used). The goodbye message is also customized based on the current session.
- Button KS, Kounali D, Stapinski L, Rapee RM, Lewis G, Munafò MR (2015) Fear of Negative Evaluation Biases Social Evaluation Inference: Evidence from a Probabilistic Learning Task. PLoS ONE 10(4): 0119456. doi:10.1371/journal. pone.0119456.
- Button KS, Browning M, Munafo MR, Lewis G (2012) Social inference and social anxiety: evidence of a fear-congruent self-referential learning bias. J Behav Ther Exp Psychiatry 43: 1082–1087. doi: 10.1016/j.jbtep.2012.05.004 PMID: 22699043.