Overview

This report includes a pre-to-post comparison of the hormones cortisol and estradiol measured at Session 1 (Pre) and Session 5 (Post) in sedentary peri- and post-menopausal individuals (N = 11; Peri n = 4, Post n = 7). Participant M010 was excluded due to missing Session 5 data.

All analyses compare paired observations within each participant. The statistical pathway is determined by a Shapiro-Wilk normality test on the within-person differences: if p > 0.05 (normal), a paired t-test is used; if p ≤ 0.05 (non-normal), a Wilcoxon signed-rank test is used instead. Effect sizes are reported as Cohen’s dz (Negligible / Small / Medium / Large) and η². 95% confidence intervals are provided for each session mean.

1. Normality Testing (Shapiro-Wilk)

The Shapiro-Wilk test was run on the paired differences (Session 5 − Session 1) for each hormone and group. A p-value > 0.05 supports sufficient normality for a paired t-test; all groups passed and paired t-tests were used throughout.

Table 1. Shapiro-Wilk Normality Tests on Paired Differences
Hormone Group N W statistic p-value Decision
Cortisol
Cortisol ALL 11 0.9476 0.6134 Normal -> Paired t-test
Cortisol Peri 4 0.9019 0.4404 Normal -> Paired t-test
Cortisol Post 7 0.9111 0.4035 Normal -> Paired t-test
Estradiol
Estradiol ALL 11 0.9053 0.2145 Normal -> Paired t-test
Estradiol Peri 4 0.8994 0.4282 Normal -> Paired t-test
Estradiol Post 7 0.9293 0.5452 Normal -> Paired t-test

2. Descriptive Statistics & 95% Confidence Intervals

Means, standard deviations, and 95% confidence intervals for each hormone at each session broken down by menopausal status.

Table 2. Descriptive Statistics with 95% Confidence Intervals
Hormone Group N Session 1 M (SD) 95% CI S1 Session 5 M (SD) 95% CI S5
Cortisol (ug/dL)
Cortisol ALL 11 0.2088 (0.2071) [0.0697, 0.348] 0.1831 (0.1812) [0.0613, 0.3048]
Cortisol Peri 4 0.2698 (0.274) [-0.1662, 0.7057] 0.1758 (0.1359) [-0.0405, 0.392]
Cortisol Post 7 0.174 (0.1734) [0.0136, 0.3344] 0.1873 (0.2132) [-0.0099, 0.3844]
Estradiol (pg/mL)
Estradiol ALL 11 31.1185 (5.0098) [27.7529, 34.4842] 30.8189 (5.8042) [26.9196, 34.7182]
Estradiol Peri 4 30.432 (5.0322) [22.4247, 38.4393] 27.392 (3.9345) [21.1314, 33.6526]
Estradiol Post 7 31.5109 (5.3549) [26.5584, 36.4633] 32.7771 (6.0087) [27.22, 38.3343]

3. Inferential Tests, Effect Sizes & η²

Table 3. Inferential Statistics, Effect Sizes, and Eta-Squared
Hormone Group N Test Used Statistic p-value Sig Cohen’s dz Effect Size eta2
Cortisol (ug/dL)
Cortisol ALL 11 Paired t-test 0.4205 0.6830 ns -0.1268 Negligible 0.0174
Cortisol Peri 4 Paired t-test 1.3031 0.2835 ns -0.6515 Medium 0.3614
Cortisol Post 7 Paired t-test -0.1522 0.8840 ns 0.0575 Negligible 0.0038
Estradiol (pg/mL)
Estradiol ALL 11 Paired t-test 0.1080 0.9161 ns -0.0326 Negligible 0.0012
Estradiol Peri 4 Paired t-test 0.6901 0.5397 ns -0.3451 Small 0.1370
Estradiol Post 7 Paired t-test -0.3447 0.7421 ns 0.1303 Negligible 0.0194

Key: ns = not significant  ·  * p < .05  ·  ** p < .01  ·  *** p < .001
Cohen’s dz: Negligible < 0.2  ·  Small 0.2–0.5  ·  Medium 0.5–0.8  ·  Large > 0.8
η²: small ≈ 0.01  ·  medium ≈ 0.06  ·  large ≈ 0.14

4. Results

Cortisol (ug/dL)

Combined sample (N = 11): Cortisol decreased from Session 1 (M = 0.2088 ug/dL, 95% CI [0.0697, 0.348]) to Session 5 (M = 0.1831 ug/dL, 95% CI [0.0613, 0.3048]). A Paired t-test indicated this change was not statistically significant (t = 0.4205, p = 0.683). The effect size was negligible (Cohen’s dz = -0.1268, eta2 = 0.0174).

Perimenopause (N = 4): Cortisol decreased from Session 1 (M = 0.2698 ug/dL, 95% CI [-0.1662, 0.7057]) to Session 5 (M = 0.1758 ug/dL, 95% CI [-0.0405, 0.392]). A Paired t-test indicated this change was not statistically significant (t = 1.3031, p = 0.2835). The effect size was medium (Cohen’s dz = -0.6515, eta2 = 0.3614).

Postmenopause (N = 7): Cortisol increased from Session 1 (M = 0.174 ug/dL, 95% CI [0.0136, 0.3344]) to Session 5 (M = 0.1873 ug/dL, 95% CI [-0.0099, 0.3844]). A Paired t-test indicated this change was not statistically significant (t = -0.1522, p = 0.884). The effect size was negligible (Cohen’s dz = 0.0575, eta2 = 0.0038).

Estradiol (pg/mL)

Combined sample (N = 11): Estradiol decreased from Session 1 (M = 31.1185 pg/mL, 95% CI [27.7529, 34.4842]) to Session 5 (M = 30.8189 pg/mL, 95% CI [26.9196, 34.7182]). A Paired t-test indicated this change was not statistically significant (t = 0.108, p = 0.9161). The effect size was negligible (Cohen’s dz = -0.0326, eta2 = 0.0012).

Perimenopause (N = 4): Estradiol decreased from Session 1 (M = 30.432 pg/mL, 95% CI [22.4247, 38.4393]) to Session 5 (M = 27.392 pg/mL, 95% CI [21.1314, 33.6526]). A Paired t-test indicated this change was not statistically significant (t = 0.6901, p = 0.5397). The effect size was small (Cohen’s dz = -0.3451, eta2 = 0.137).

Postmenopause (N = 7): Estradiol increased from Session 1 (M = 31.5109 pg/mL, 95% CI [26.5584, 36.4633]) to Session 5 (M = 32.7771 pg/mL, 95% CI [27.22, 38.3343]). A Paired t-test indicated this change was not statistically significant (t = -0.3447, p = 0.7421). The effect size was negligible (Cohen’s dz = 0.1303, eta2 = 0.0194).

Notable finding: Although no comparison reached statistical significance (all p > 0.05), cortisol in the perimenopause group showed a medium effect size (Cohen’s dz = −0.65, η² = 0.36), indicating a potentially meaningful decrease that likely failed to reach significance due to the small sample (n = 4). Estradiol was highly stable across all groups and sessions.

5. Figures

Figure 1 — Individual Trajectories

Each line represents one participant’s change from Session 1 (Pre) to Session 5 (Post). The thick diamond line is the group mean. Red = increase; blue = decrease. Statistical annotations (test used, p-value, effect size, and η²) are shown per panel.

Figure 2 — Group Means with 95% Confidence Intervals

Error bars represent 95% confidence intervals around each group mean. This is the uncertainty in each estimate and is more informative for inference than standard deviation bars.

Figure 3 — Distributions: Box Plots with Individual Data Points

Shows the full spread of values per group and session. Each dot is one participant. The box shows the median and interquartile range. Significance brackets compare Pre to Post within each group.

Figure 4 — Effect Size Summary

Horizontal bars show absolute Cohen’s dz for each hormone and group. Dashed lines mark conventional thresholds (small = 0.2, medium = 0.5, large = 0.8). η² is annotated on each bar. Bars coloured red are from the perimenopause group; blue from postmenopause; purple from the combined sample.

6. Conclusions

No statistically significant pre-to-post changes were detected for either cortisol or estradiol in any group (all p > 0.05). Key observations:

  • Cortisol — Perimenopause: The most notable finding in the dataset. A medium effect size (Cohen’s dz = −0.65, η² = 0.36) indicates a potentially meaningful decrease in cortisol from Session 1 to Session 5. Statistical significance was not achieved, most likely due to the small perimenopause sample (n = 4). This trend needs follow-up with a larger cohort.

  • Cortisol — Postmenopause & Combined: Effect sizes were negligible (dz < 0.13, η² < 0.02), reflecting inconsistent individual changes with no overall directional trend.

  • Estradiol: Highly stable across all groups and sessions. The largest effect was a small dz = −0.35 in the perimenopause group (η² = 0.14), with no meaningful directional change observed in postmenopause or the combined sample.

  • Clinical relevance: Effect sizes of medium magnitude are considered clinically meaningful even when p-values exceed α = 0.05, particularly in small pilot studies. The cortisol finding in perimenopause should not be dismissed on the basis of p-value alone.

Note: For unevenly spaced visits, time-series models (ARIMA, state-space models) or functional data analysis can capture longitudinal patterns subject-specifically and should be considered for future analyses involving more sessions.

Analysis conducted in R. Figures produced with ggplot2 and patchwork.