Mental Health Last reviewed: June 28, 2026

Does power posing increase confidence and testosterone?

Weak Evidence
Confidence Score 20%

NO

Weak evidence. The original 2010 power posing study claiming hormonal changes (increased testosterone, decreased cortisol) failed to replicate in larger, pre-registered studies. A co-author retracted support. Modest self-reported confidence effects may exist but are small and context-dependent.

The Verdict

Weak evidence. The original 2010 power posing study claiming hormonal changes (increased testosterone, decreased cortisol) failed to replicate in larger, pre-registered studies. A co-author retracted support. Modest self-reported confidence effects may exist but are small and context-dependent.

What the Evidence Shows

Power posing—adopting expansive, open body postures for 1-2 minutes—was popularized by Amy Cuddy's 2012 TED talk based on a 2010 study claiming that high-power poses increased testosterone by 20%, decreased cortisol by 25%, and increased risk tolerance compared to low-power poses. The study used only 42 participants. The claim became emblematic of the replication crisis in psychology when Dana Carney, a co-author, publicly stated in 2016 that she did not believe the effects were real. Multiple well-powered replication attempts have failed to find hormonal changes. A P-Curve analysis of the original literature suggested the effects were likely driven by p-hacking or selective reporting. Ranehill et al. (2015) conducted a direct replication with 200 participants and found no significant effects on testosterone, cortisol, or risk-taking behavior. However, several subsequent studies have found small effects on self-reported feelings of power and confidence, though these may be driven by demand characteristics (participants guessing the hypothesis). A 2017 meta-analysis by Gronau et al. found no evidence for hormonal effects but weak evidence for self-reported effects (d=0.22), with considerable heterogeneity. The current scientific consensus is that power posing does not produce the physiological changes originally claimed.

Evidence Quality

2

Meta-Analyses

8

RCTs

3

Observational

Important Caveats

  • ⚠️ The original co-author (Dana Carney) publicly retracted support for the findings in 2016
  • ⚠️ Multiple pre-registered replications failed to find hormonal changes
  • ⚠️ Any self-reported confidence effects are small (d=0.22) and may reflect demand characteristics
  • ⚠️ The original study had only 42 participants—severely underpowered for between-group hormonal comparisons
  • ⚠️ P-Curve analysis suggests the original literature contained selective reporting

Population Studied

University students aged 18-25 in laboratory settings; predominantly Western, educated populations from psychology participant pools

Dosage

Standard protocol: 2 minutes of expansive posing (hands on hips, feet apart, or leaning back with hands behind head) compared to 2 minutes of contractive posing

Duration

Single-session laboratory studies with measurements taken immediately after 2-minute posing protocol; no long-term intervention studies

Supporting Studies (2)

Power Posing: Brief Nonverbal Displays Affect Neuroendocrine Levels and Risk Tolerance

RCT

Carney DR, Cuddy AJ, Yap AJ. · Psychological Science (2010)

In 42 participants, adopting high-power poses for 2 minutes increased testosterone by 20%, decreased cortisol by 25%, and increased self-reported feelings of power and risk tolerance compared to low-power poses.

View paper (DOI) →

Review and Summary of Research on the Embodied Effects of Expansive (vs. Contractive) Nonverbal Displays

Meta-Analysis

Cuddy AJC, Schultz SJ, Fosse NE. · Psychological Science (2018)

Meta-analysis of 55 studies found a significant overall effect of expansive postures on self-reported feelings of power (d=0.22, p<0.001), though no significant effects on hormonal measures when restricted to pre-registered studies.

View paper (DOI) →

Contradicting Studies (1)

Assessing the Robustness of Power Posing: No Effect on Hormones and Risk Tolerance in a Large Sample of Men and Women

RCT

Ranehill E, Dreber A, Johannesson M, et al. · Psychological Science (2015)

In a pre-registered replication with 200 participants (5x the original sample), high-power poses produced no significant changes in testosterone (p=0.62), cortisol (p=0.34), or behavioral risk-taking compared to low-power poses, though a small effect on self-reported feelings emerged.

Why this disagrees:

With 5 times the statistical power of the original study and pre-registered analysis, this well-powered direct replication found no hormonal effects whatsoever, suggesting the original findings were likely false positives from an underpowered study with researcher degrees of freedom.

View paper (DOI) →
Share:
Was this helpful?

Related Claims