Lie prevalence, lie characteristics and strategies of self-reported good liars

People are generally poor at detecting deception, with accuracy around chance. Meta-analytic work shows that lie detection success depends more on characteristics of the liar (sender) than on the detector (judge), with substantial sender-to-sender variability in detectability. This study focuses on the liar, especially those who self-identify as good liars. The research questions were: (1) How does self-reported deception ability relate to lie prevalence in daily life? (2) Do lie characteristics (type/severity, recipient, communication medium) vary with self-reported deception ability? (3) Which deception strategies are reported by self-identified good liars? Understanding these aspects has practical importance for investigative contexts where skilled liars may evade detection.

Prior work estimates average lie frequency at about once or twice per day, but distributions are highly skewed, with a small minority of prolific liars accounting for a large share of lies. Prolific liars may tell more serious lies and show greater willingness to cheat. Lie types range from inconsequential white lies used as social lubricants to serious fabrications often relevant to legal contexts; other taxonomies include omission and embedding lies within truth. People tend to lie less in close relationships compared to casual ones. Medium can shape deception; some prefer online communication, sometimes based on misconceptions that behavioral cues betray lies. Strategy research suggests deceivers manage impressions and information, often staying close to the truth or withholding information. Strategy-based tools like the Verifiability Approach exploit liars’ tendencies to include unverifiable details. It has been hypothesized that good liars align behavior with lay beliefs about truthful demeanor. Despite this, surveying liars about their strategies remains underdeveloped.

Design: Cross-sectional online survey hosted on Qualtrics. Ethical approval obtained from Maastricht University; informed consent obtained. Participants: N = 194 (97 female, 95 male, 2 prefer not to say); Mage = 39.12 years (SD = 11.43). Recruitment via Amazon Mechanical Turk (Masters Qualification). 175 U.S. citizens, 19 Indian citizens. Compensation: $1.75. An additional 133 incomplete responses discarded; 9 removed for insufficient responses. Measures and Procedure: Participants were provided standardized definitions of lying/deception. They rated self-perceived deception ability on a 10-point scale (1 very poor to 10 excellent). They reported the number of lies told in the past 24 hours and, using multiple-response options, indicated: (i) types of lies (white lies, exaggerations, omission/concealment, commission/fabrications, embedded lies); (ii) recipients (family, friend, employer, colleague, authority figure, other); and (iii) mediums (face-to-face, phone, social media, text, email, other). Deception strategies: Open-ended question on general strategies used when lying. Likert ratings (1–10) of the importance of verbal and nonverbal strategies (definitions provided). Selection from a predetermined list of 10 verbal strategies (e.g., keeping statements clear/simple, telling a plausible story, embedding lies in truthful information, providing unverifiable details, matching amount/type of details, avoidance/vagueness, reporting from previous experience, complete fabrication, other). Qualitative coding: Content analysis transformed open-ended strategies into seven categories (e.g., omitting information, relating to truthful information, behavioral control). Inter-rater reliability established on 20% random subsample using two-way mixed ICCs, Single Measures ICCs .79 to 1.00. Main coder completed remaining data. Analyses: Descriptive statistics for prevalence and characteristics; correlations between self-reported deception ability and lie prevalence/characteristics. Participants categorized as Poor (≤3; n=51), Neutral (4–7; n=75), Good (≥8; n=68) liars. Chi-square tests for associations between strategy categories and ability groups. Between-subjects ANOVAs for importance ratings of verbal/nonverbal strategies, with Bayesian ANOVAs (BF10/BF01). One-way ANOVAs for endorsement of specific predetermined strategies with post hoc tests (Bonferroni). Exploratory chi-square tests for associations between sex and education with deception ability.

Lie prevalence: Mean self-reported lies in past 24 hours = 1.61 (SD = 2.75; range 0–20). Distribution highly skewed (skewness = 3.90, SE = 0.18; kurtosis = 18.44, SE = 0.35). The six most prolific liars (<1% of participants) accounted for 38.5% of all lies; 39% reported telling no lies. Correlations with self-reported deception ability: Higher ability associated with more lies per day, r(192) = .22, p = .002; higher endorsement of white lies, r(192) = .16, p = .023; and exaggerations, r(192) = .16, p = .027. No significant associations for embedded lies, r = .14, p = .051; omission, r = .10, p = .171; commission, r = .10, p = .161. Recipients: Higher ability linked to lying to colleagues, r(192) = .27, p < .001; friends, r(192) = .16, p = .026; and “other” recipients (e.g., romantic partners/strangers), r(192) = .16, p = .031. Not significantly related to lying to family, r = .08, p = .243; employers, r = .04, p = .558; authority figures, r = .11, p = .133. Medium: Higher ability correlated with face-to-face deception, r(192) = .26, p < .001. No significant associations for phone, r = .13, p = .075; text, r = .13, p = .083; social media, r = .03, p = .664; email, r = .05, p = .484; other, r = .10, p = .153. Strategy categories (qualitative): No significant associations across most categories; exception: endorsement of “No strategy” differed across groups, χ²(2) = 8.26, p = .016, V = .206; pairwise: Good < Poor, p = .004 (caution due to small expected frequencies). Importance of strategies: Verbal strategies rated more important by Good liars, F(2,191) = 5.62, p = .004, ηp² = .056; BF10 = 7.11. No group differences for nonverbal strategies, F(2,191) = 0.003, p = .997, ηp² < .001; BF01 = 18.55. Specific verbal strategies (predetermined list): Significant group differences for:

• Embedding the lie: F(2,191) = 11.97, p < .001, ηp² = .111; BF10 = 1438.20 (Good > Poor).
• Matching amount of details: F(2,191) = 4.77, p = .010, ηp² = .048; BF10 = 3.32 (Good > Poor).
• Matching type of details: F(2,191) = 3.56, p = .030, ηp² = .036; BF10 = 1.15 (Good > Poor).
• Keeping statement clear/simple: F(2,191) = 5.07, p = .007, ηp² = .050; BF10 = 4.15 (Good > Poor).
• Telling a plausible story: F(2,191) = 5.48, p = .005, ηp² = .054; BF10 = 5.98 (Good > Poor).
• Providing unverifiable details: F(2,191) = 4.95, p = .008, ηp² = .049; BF10 = 3.78 (Good > Poor).
• Avoidance (vagueness): F(2,191) = 3.79, p = .024, ηp² = .038; BF10 = 1.43 (Poor > Good). Non-significant: Reporting from previous experience (F = 1.32, p = .268; BF01 = 5.96), complete fabrication (F = 0.57, p = .565; BF01 = 11.36), other strategies (F = 0.51, p = .600; BF01 = 11.96). Exploratory associations: Sex associated with deception ability, χ²(2) = 12.31, p = .002, V = .253; among Poor liars, 70% female; among Good liars, 62.7% male. Education not significantly associated, χ²(4) = 9.09, p = .059, V = .153.

Findings support the premise that individuals who view themselves as good liars report higher lie prevalence and specific lie characteristics: more inconsequential lies (white lies, exaggerations), more lies to colleagues and friends, and preference for face-to-face deception. Strategy-wise, good liars emphasize verbal content management: embedding lies within truthful narratives, maintaining clarity and simplicity, ensuring plausibility, matching detail quantity and type to truthful parts, and selectively using unverifiable details. Poor liars more often report avoidance/vagueness. These results align with literature that content plausibility strongly influences deception judgments and that liars tend to stay close to truth where possible. The observed preference for face-to-face deception among good liars may reflect strategic adaptation to expectations of online deception. The reliance of good liars on sophisticated verbal strategies has implications for content-based lie detection approaches, which may be less effective against skilled liars who mimic truthful recall patterns and add unverifiable details (as highlighted by the Verifiability Approach).

This study provides new insights into the prevalence, characteristics, and strategies of self-reported good liars. A minority accounted for a disproportionate number of daily lies, and these prolific liars also tended to rate themselves as good liars. Self-reported good liars mainly told inconsequential lies, particularly to colleagues and friends, typically via face-to-face interactions. They reported strategically managing verbal content by embedding lies within truthful information and by keeping stories clear, simple, and plausible. These observations offer a starting point for further work on skilled liars’ meta-cognitions and may inform investigative interviewing and lie detection methods.

Findings are based on self-reports of lying behavior and ability, which can introduce biases and limit inference; ground truth of actual lie-telling skill was not established. The MTurk sample raises typical concerns about motivation and compensation, although steps were taken to ensure data quality and prior work supports MTurk data validity. Discrepancies between qualitative, data-driven coding and predetermined strategy endorsements highlight potential subjectivity and measurement limitations. Generalizability could be improved by replication in different populations (e.g., students) and by using controlled laboratory assessments of deception ability and strategy use, potentially with cued recall or mixed inductive–theory-driven coding approaches.