Skip to content

Unsupervised Learning Algorithms being implemented to detect a liar.

Notifications You must be signed in to change notification settings

MarsRoboters/Liar-Detector

Repository files navigation

NAQ-R Object: Mobbing Number of Item: 22 Likert Scale: 1-5 Participants: Honest: 356 Faker: 356 Clinical: 176

Study Design: Within subject Procedure: It was asked to subjects to answer twice a questionnaire developed to identify possible victims of mobbing, the Negative Acts Questionnaire-Revised (NAQ-R), following different instructions: the first time by answering honestly, while the second time by pretending to be victims of mobbing. Therefore, each participant responded twice, once honestly (H) and once faking (D). We also collected data from people whose questionnaire scores exceeded the cut-off (P). Of these, we have the data in the honest condition. FAKING BAD: we expect liars will give higher responses than truth-tellers because they have to exaggerate bullying symptoms

Item and scale: Item: NAQ-R_WR_1; NAQ-R_PR_2; NAQ-R_WR_3; NAQ-R_WR_4; NAQ-R_PR_5; NAQ-R_PR_6; NAQ-R_PR_7; NAQ-R_PI_8; NAQ-R_PI_9; NAQ-R_WR_10; NAQ-R_WR_11; NAQ-R_PR_12; NAQ-R_WR_13; NAQ-R_PR_14; NAQ-R_PI_15; NAQ-R_WR_16; NAQ-R_PR_17; NAQ-R_WR_18; NAQ-R_WR_19; NAQ-R_PR_20; NAQ-R_WR_21; NAQ-R_PI_22

ITEM NAME LEGEND: NAQ-R = questionnaire WR = work-related bullying PR = person - related bullying PI = physically intimidating bullying