Nonverbal communication shapes outcomes more often than people realize, and the evidence keeps getting sharper. Employers weigh communication skills so heavily that 67% say nonverbal signals matter when evaluating candidates, yet many people feel harder to read emotions over video with 62% reporting it is more difficult than in person. This post connects the dots from facial microcues to gaze and posture so you can see exactly where “tone” is actually coming from.
Key Takeaways
- 67% of employers say they consider communication skills (including nonverbal signals) important when evaluating candidates.
- 9 out of 10 people feel more connected to others when they can perceive nonverbal cues like facial expressions and gestures.
- The US Department of Labor’s O*NET defines ‘Social Perceptiveness’ and related competencies that include interpreting people’s reactions and behavior—core nonverbal communication tasks—within occupational requirements.
- In a controlled study, participants judged smiles as more trustworthy after correcting for lighting and image contrast, indicating strong sensitivity to facial expression cues.
- In a large meta-analysis, emotional facial expressions were associated with improved emotion recognition accuracy compared with neutral faces, with effect sizes varying by emotion type.
- A frequently cited meta-analysis found that nonverbal cues alone yield only modest deception-detection accuracy (about 54% correct) for humans.
- Automatic facial expression analysis systems can detect basic emotions with reported accuracies exceeding 80% in benchmarking datasets (e.g., FER datasets).
- Vision-based eye-tracking can estimate gaze direction with typical angular error on standard benchmarks in the ~1–3 degree range depending on dataset and method.
- Machine learning models for facial landmark detection achieve normalized mean error reductions to below 2 pixels on common face alignment benchmarks.
- In virtual meetings, a survey reported that 87% of employees feel video conferencing improves communication effectiveness compared with audio-only calls.
- A Gartner survey found that 74% of organizations plan to shift more meetings to virtual formats in the next 12 months.
- A 2020 report by Microsoft Work Trend Index found that Teams meeting hours increased 50–70% year over year in many organizations, increasing exposure to nonverbal cues on video.
- Under GDPR, biometric data for uniquely identifying a natural person is classified as ‘special category data’ subject to stricter processing conditions.
- The NIST Face Recognition Vendor Test (FRVT) reports false match rates (FMR) and false non-match rates (FNMR) for face recognition systems, measuring nonverbal biometric accuracy.
- The American Psychological Association (APA) ethics materials emphasize avoiding harm and ensuring informed consent when collecting behavioral/nonverbal data in research contexts.
Nonverbal cues strongly shape trust, engagement, and emotion recognition, with evidence from studies and workplaces.
Related reading
01 · Category
Business Impact5 stats
01
67% of employers say they consider communication skills (including nonverbal signals) important when evaluating candidates.
02
9 out of 10 people feel more connected to others when they can perceive nonverbal cues like facial expressions and gestures.
03
The US Department of Labor’s O*NET defines ‘Social Perceptiveness’ and related competencies that include interpreting people’s reactions and behavior—core nonverbal communication tasks—within occupational requirements.
04
The O*NET skill ‘Active Listening’ is tied to understanding spoken content, but the model also rates ‘Speaking’ and ‘Social Perceptiveness’ required for interpreting cues from others, including nonverbal behavior.
05
O*NET lists ‘Social Perceptiveness’ as a Personal and Social Skill for many occupations, reflecting that interpreting others’ behavior is a quantified work requirement.
Interpretation
Business Impact Interpretation
With 67% of employers factoring nonverbal communication into candidate evaluation, and 9 out of 10 people feeling more connected when they can read cues, the business impact is clear: being able to interpret nonverbal signals is a measurable workplace skill reflected across O*NET requirements.
02 · Category
Research Findings14 stats
01
In a controlled study, participants judged smiles as more trustworthy after correcting for lighting and image contrast, indicating strong sensitivity to facial expression cues.
02
In a large meta-analysis, emotional facial expressions were associated with improved emotion recognition accuracy compared with neutral faces, with effect sizes varying by emotion type.
03
A frequently cited meta-analysis found that nonverbal cues alone yield only modest deception-detection accuracy (about 54% correct) for humans.
04
In a meta-analysis of nonverbal communication, effect sizes for gaze and face cues on social judgments were reliably greater than zero across studies.
05
Facial action coding (FACS) decomposes facial movement into 44 action units used in 1960s/1970s research and is still referenced in contemporary nonverbal emotion measurement.
06
The Facial Action Coding System (FACS) includes 44 action units that map to visible facial movements for coding nonverbal expressions.
07
The ‘Reading the Mind in the Eyes’ test has 36 items and is widely used to measure how well people interpret subtle facial cues.
08
A meta-analysis reported that nonverbal behavior is associated with personality judgments (e.g., extraversion impressions), with statistically significant pooled effects.
09
In a study of video-mediated communication, eye gaze cues are associated with improved perceived engagement, with measurable differences across gaze conditions.
10
In a controlled experiment, showing participants more facial-expression intensity increased emotion recognition accuracy by several percentage points versus lower-intensity displays.
11
In a meta-analysis, body posture (e.g., openness, dominance-leaning) influenced impressions with statistically significant pooled effects across studies.
12
In psychotherapy, therapist nonverbal synchrony has been reported to correlate with better relationship outcomes in multiple studies (pooled evidence reported in the literature).
13
A 2019 meta-analysis reported that nonverbal cues can improve person perception accuracy compared with chance levels, though the magnitude depends on cue type (face, body, voice).
14
The Karolinska Directed Emotional Faces (KDEF) dataset contains 70 individuals and multiple posed images across emotional categories for studying facial nonverbal expressions.
Interpretation
Research Findings Interpretation
Across research findings, nonverbal communication reliably improves social and emotion reading, with meta-analytic accuracy modestly above chance and several studies showing measurable gains, such as about 54% deception detection for humans and well-established tools like FACS with 44 action units and the 36-item Reading the Mind in the Eyes test.
03 · Category
Technology Performance10 stats
01
Automatic facial expression analysis systems can detect basic emotions with reported accuracies exceeding 80% in benchmarking datasets (e.g., FER datasets).
02
Vision-based eye-tracking can estimate gaze direction with typical angular error on standard benchmarks in the ~1–3 degree range depending on dataset and method.
03
Machine learning models for facial landmark detection achieve normalized mean error reductions to below 2 pixels on common face alignment benchmarks.
04
Kinesic (gesture) matching studies report that gesture recognition accuracy is significantly higher when observers see full-body motion rather than cropped frames.
05
The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) contains 24 actors and 7356 files for emotion recognition research involving vocal and facial expressions.
06
The CK+ facial expression dataset includes 593 video sequences labeled with 7 basic emotions and is used to evaluate nonverbal emotion recognition models.
07
The AffectNet dataset contains approximately 450,000 labeled images used for facial expression recognition research.
08
The MultiPIE dataset includes 337 subjects and provides a large set of facial images under varying pose, illumination, and expression for studying nonverbal facial cues.
09
The DISFA dataset includes 27,000+ facial action unit frames and is widely used to model facial nonverbal muscle movements.
10
The BP4D+ dataset provides 140 subjects and annotated facial action units for facial nonverbal behavior research.
Interpretation
Technology Performance Interpretation
In the technology performance of nonverbal communication systems, face and gaze analysis can reach over 80% emotion detection accuracy and about 1–3 degrees gaze error while modern landmark models push error below 2 pixels, and large-scale benchmarks like AffectNet’s ~450,000 images and DISFA’s 27,000+ action unit frames are helping drive these gains.
04 · Category
Industry Trends3 stats
01
In virtual meetings, a survey reported that 87% of employees feel video conferencing improves communication effectiveness compared with audio-only calls.
02
A Gartner survey found that 74% of organizations plan to shift more meetings to virtual formats in the next 12 months.
03
A 2020 report by Microsoft Work Trend Index found that Teams meeting hours increased 50–70% year over year in many organizations, increasing exposure to nonverbal cues on video.
Interpretation
Industry Trends Interpretation
Industry Trends show that as virtual meetings expand rapidly, with 87% of employees saying video improves communication and Gartner reporting 74% of organizations plan to shift more meetings online, increased Teams meeting hours by 50–70% in 2020 is likely driving far greater reliance on nonverbal cues on video.
05 · Category
Policy & Ethics4 stats
01
Under GDPR, biometric data for uniquely identifying a natural person is classified as ‘special category data’ subject to stricter processing conditions.
02
The NIST Face Recognition Vendor Test (FRVT) reports false match rates (FMR) and false non-match rates (FNMR) for face recognition systems, measuring nonverbal biometric accuracy.
03
The American Psychological Association (APA) ethics materials emphasize avoiding harm and ensuring informed consent when collecting behavioral/nonverbal data in research contexts.
04
ISO/IEC 30107-3 specifies performance testing of presentation attack detection for biometric systems, relevant to nonverbal face recognition spoofing defenses.
Interpretation
Policy & Ethics Interpretation
Policy and ethics in nonverbal communication are tightening as GDPR treats uniquely identifying biometric data as special category data, while accuracy and safety standards keep expanding through NIST’s FRVT false match and false non-match metrics and ISO/IEC 30107-3 performance tests for presentation attack detection.
More related reading
06 · Category
Industry Benchmarks3 stats
01
20% of US adults report trouble communicating (a social/communication difficulty category that can include nonverbal challenges)
02
15% of people aged 15+ worldwide have a disability that affects functioning (disability can influence perception and use of social/nonverbal signals)
03
During 2020–2021, 28% of call centers adopted or planned adoption of automated emotion/voice analytics (used to infer customer affect from vocal/nonverbal cues)
Interpretation
Industry Benchmarks Interpretation
Industry benchmarks suggest that nonverbal communication is a real and measurable challenge at scale, with 20% of US adults reporting trouble communicating and 15% of people worldwide living with a disability that can affect how social signals are understood, which helps explain why 28% of call centers in 2020 to 2021 moved toward automated emotion and voice analytics to better read affect.
07 · Category
Research Evidence7 stats
01
1.4x increase in the likelihood of accurately identifying emotions when nonverbal cues are available versus not available (cue availability effect reported in a quantitative synthesis of emotion perception studies)
02
2.1x higher odds that therapists’ nonverbal synchrony is associated with better therapeutic alliance outcomes versus low synchrony (pooled effect direction reported across studies)
03
0.24 pooled standardized mean difference for the association between gaze behavior and social judgments across studies (quantitative synthesis reported pooled effects)
04
0.31 pooled standardized mean difference for the effect of facial expressivity on emotion recognition accuracy in a meta-analytic synthesis (facial cues increase recognition above chance)
05
0.18 pooled standardized mean difference for posture/stance cues influencing impressions (meta-analytic evidence pooled across studies)
06
A 2020 randomized study found that participants exposed to therapist nonverbal mirroring delivered more supportive responses, with an effect size reported as Cohen’s d = 0.44
07
The ‘Reading the Mind in the Eyes’ task uses 36 items and is designed to measure how accurately people infer mental states from eye-region cues
Interpretation
Research Evidence Interpretation
Across this research evidence, nonverbal cues consistently improve social and emotional understanding, including a 1.4x increase in accurate emotion identification when cues are available and a pooled effect of 0.31 showing facial expressivity boosts emotion recognition above chance.
08 · Category
Market Size1 stats
01
The global affective computing market is projected to grow from $8.7B in 2022 to $61.8B by 2032 (growth driven by facial/voice/nonverbal analytics)
Interpretation
Market Size Interpretation
The affective computing market tied to facial, voice, and nonverbal analytics is expected to surge from $8.7B in 2022 to $61.8B by 2032, underscoring a major market-size expansion driven by advances in nonverbal communication understanding.
09 · Category
User Adoption2 stats
01
6 out of 10 workers in a 2022 survey reported they use video calls for important conversations, increasing the frequency of nonverbal cue interpretation
02
In a 2021 study of remote work, 62% of respondents stated they felt it was harder to read others’ emotions over video than in person
Interpretation
User Adoption Interpretation
For the user adoption angle, 6 out of 10 workers say they already rely on video calls for important conversations, yet 62% report it is harder to read emotions over video than in person, showing adoption is rising even as the challenge of interpreting nonverbal cues remains.
Reference
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
APA
Henrik Dahl. (2026, February 13). Nonverbal Communication Statistics. Gitnux. https://gitnux.org/nonverbal-communication-statistics
MLA
Henrik Dahl. "Nonverbal Communication Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/nonverbal-communication-statistics.
Chicago
Henrik Dahl. 2026. "Nonverbal Communication Statistics." Gitnux. https://gitnux.org/nonverbal-communication-statistics.
Sources & references
49 datasets cited across this report · attribution is report-level
+23 additional datasets cited (not shown individually)