Skip to main content
. 2021 Jun 21;7:100068. doi: 10.1016/j.cpnec.2021.100068

Table 2.

Group comparisons of hair variables.


Nondepressed controls (n = 27)
Major depressive disorder (n = 21)


Mdn (IQR) [Range] Mdn (IQR) [Range] Test statistics Effect size
Cortisol [pg/mg]a 9.7 (3.2) [4.6, 65.9] 8.3 (6.3) [3.9, 157.0] ANCOVAc: b[95% CI]d = −0.20 [ −0.36, −0.03], p = .012, pFDR= .040 d = −.49
Cortisone [pg/mg] 24.6 (15.6) [15.6, 87.8] 16.9 (7.8) [10.1, 50.4] U = 824.5, p < .001, pFDR= .005 rrankbiserial = −.57
Testosterone [pg/mg] 6.0 (3.8) [3.1, 23.1] 6.9 (2.4) [4.9, 12.5] ANCOVAc: b[95% CI] = −0.95 [ −0.18, 2.13], p = .050, pFDR = .116 d = .42
Progesterone [pg/mg]b 5.2 (1.0) [4.3, 9.1] 7.2 (2.5) [4.5, 9.1]
2-AG [pg/mg] 246 (90) [166, 483] 256 (107) [152, 453] U = 657.5, p = .934, pFDR = .934 rrankbiserial = .01
PEA [pg/mg] 2522 (1450) [1323, 11815] 2302 (2586) [1069, 31671] U = 645.0, p = .732, pFDR = .854 rrankbiserial = .06
SEA [pg/mg] 723 (634) [84, 1912] 851 (416) [331, 10007] U = 595.0, p = .167, pFDR = .292 rrankbiserial = .23
OEA [pg/mg] 1829 (1168) [690, 7167] 1687 (2175) [927, 25327] U = 633.0, p = .554, pFDR = .775 rrankbiserial = −.10

Note: Substance concentrations could not be measured in all hair samples: an = 23 vs. 15, bn = 5 vs. 4; due to the small number of successful measures, progesterone concentrations were not statistically analyzed. c Robust ANCOVAs were adjusted for age and/or hair washing frequency. Detailed model and predictor statistics are provided in Supplementary Table 2d The outcome variable’s natural logarithm (ln) was used to lower the influence of distant data points.