A pilot randomized controlled trial of group-based indoor gardening and art activities demonstrates therapeutic benefits to healthy women

The researchers compared two active interventions: group-based indoor gardening and group-based art-making. Both were delivered as eight one-hour sessions (twice per week for four weeks). There was no inactive control group (no "do nothing" group). The study measured:

**Primary outcomes:** Total mood disturbance (Profile of Mood States), depression symptomatology (Center for Epidemiologic Studies Depression Scale), perceived stress (Perceived Stress Scale), trait and state anxiety (State-Trait Anxiety Inventory), satisfaction with discretionary social activities (Satisfaction with Discretionary Social Activities Scale), and quality of life (World Health Organization Quality of Life-BREF).

**Secondary outcomes:** Heart rate and systolic/diastolic blood pressure, measured before and after each session.

The key question was whether gardening and art-making produce *different* therapeutic benefits, or whether both are similarly beneficial.

**Sample size:** 42 women were randomized; 36 started the protocol; 32 completed all assessments (Gardening n=15, Art n=17).

**Population:** Healthy, pre-menopausal women aged 26–49 (mean age ~35).

**Setting:** University of Florida, Gainesville, Florida. Sessions were held at the Wilmot Botanical Gardens (gardening) and the Center for Arts in Medicine (art-making).

**Inclusion criteria:** BMI <32, non-smoker, non-gardener, non-artist, no chronic conditions, no allergies to pollen/plants/plant-based foods, no alcohol/drug abuse.

**Exclusion criteria:** Anyone with chronic health conditions, allergies, smoking, or regular gardening/art practice.

All outcomes were self-report questionnaires, plus physiological measurements:

**Profile of Mood States (POMS):** 65-item scale measuring total mood disturbance (higher = worse mood). Subscales include tension, depression, anger, vigor, fatigue, confusion.

**Center for Epidemiologic Studies Depression Scale (CES-D):** 20-item scale (0–60 range, higher = more depressive symptoms). Clinical cutoff ≥16.

**Perceived Stress Scale (PSS):** 10-item scale (0–40 range, higher = more stress).

**State-Trait Anxiety Inventory (STAI):** Two 20-item scales — state anxiety (how you feel right now) and trait anxiety (how you generally feel). Each 20–80 range, higher = more anxiety.

**Satisfaction with Discretionary Social Activities Scale (SDSA):** Measures satisfaction with how you spend free time.

**World Health Organization Quality of Life-BREF (WHOQOL-BREF):** 26-item scale covering physical, psychological, social, and environmental domains.

**Heart rate and blood pressure:** Measured using an automated monitor (Omron HEM-907XL) before and after each session.

Assessments were done at baseline, after session 4 (midpoint), and after session 8 (post-intervention). Physiological measures were taken before and after every single session.

**Design:** Pilot randomized controlled trial (RCT) with two parallel arms. Participants were randomly assigned to either gardening or art-making using a computer-generated random number sequence. Allocation was concealed (the person assigning participants didn't know which group they'd go to).

**Blinding:** This was an open-label trial — participants knew which activity they were doing, and the researchers delivering the sessions obviously knew. The outcome assessors (people scoring the questionnaires) were blinded to group assignment. This is a significant limitation because self-report measures are highly susceptible to expectation effects (if you believe gardening will help, you might report feeling better).

**Duration:** Four weeks of intervention (eight sessions total, twice per week). Follow-up was immediate post-intervention only — no long-term follow-up (e.g., 1 month or 3 months later).

**Statistical approach:** Mixed-model repeated measures ANOVA (analysis of variance) to compare changes over time between groups. Effect sizes reported as partial eta-squared (η²p). They also did "dosage-response" analyses looking at whether more sessions produced greater improvements.

**What this design can and cannot prove:**

**Can prove:** That both gardening and art-making are associated with improvements in mood, stress, and depression over four weeks in healthy women. The randomization helps ensure that pre-existing differences between groups don't explain the results.

**Cannot prove:** That gardening or art-making *caused* the improvements (no inactive control group means we can't rule out placebo effects, natural recovery, or simply doing something enjoyable in a group). Cannot prove that one is better than the other (the study was designed to detect *differences* but found none — but this could be due to small sample size). Cannot prove long-term benefits (no follow-up). Cannot generalize to men, older women, or people with health conditions.

**Major methodological weaknesses:**

1. **No inactive control group** — This is the biggest flaw. Without a "waitlist" or "no treatment" group, we don't know if improvements were due to the activities themselves, the social interaction, the novelty, or simply the passage of time.

2. **Small sample size** — 32 completers is very small for detecting between-group differences. The study was powered for within-group changes but not for comparing gardening vs. art.

3. **Self-report measures** — All primary outcomes were subjective questionnaires, vulnerable to demand characteristics and expectation bias.

4. **No long-term follow-up** — We don't know if benefits lasted beyond the four weeks.

5. **Healthy population** — Ceiling effects are possible: healthy people have less room for improvement than clinical populations.

6. **Single sex, narrow age range** — Limits generalizability.

**Primary outcomes (self-report questionnaires):**

**Total mood disturbance (POMS):** Both groups improved significantly from baseline to post-intervention. Gardening: mean score dropped from ~30 to ~10 (a 67% reduction). Art: mean score dropped from ~35 to ~15 (a 57% reduction). The difference between groups was not statistically significant (p=0.47, η²p=0.02 — a small effect).

**Depression symptomatology (CES-D):** Both groups improved. Gardening: mean dropped from ~12 to ~6 (50% reduction). Art: mean dropped from ~14 to ~8 (43% reduction). No significant between-group difference (p=0.56, η²p=0.01).

**Perceived stress (PSS):** Both groups improved. Gardening: mean dropped from ~18 to ~13 (28% reduction). Art: mean dropped from ~19 to ~14 (26% reduction). No significant between-group difference (p=0.84, η²p=0.001).

**Trait anxiety (STAI-T):** Gardening group showed a significant reduction (mean dropped from ~38 to ~33, p=0.02, η²p=0.17 — a medium-to-large effect). Art group showed no significant change (mean ~39 to ~38, p=0.58). The between-group difference was not significant (p=0.08).

**State anxiety (STAI-S):** Both groups improved, but the difference between groups was not significant (p=0.32).

**Satisfaction with discretionary social activities (SDSA):** No significant changes in either group (p=0.48).

**Quality of life (WHOQOL-BREF):** No significant changes in any domain for either group.

**Secondary outcomes (physiological):**

**Heart rate:** No significant changes in either group (p=0.89).

**Systolic blood pressure:** No significant changes (p=0.76).

**Diastolic blood pressure:** No significant changes (p=0.91).

**Dosage-response analysis:** For both gardening and art, improvements in total mood disturbance, perceived stress, and depression symptomatology became larger with more sessions. The biggest improvements occurred between baseline and session 4, with additional smaller gains from session 4 to session 8.

**Total mood disturbance:** Both groups improved by about 20 points on the POMS (0–200 scale). This is a large effect — equivalent to moving from "moderately distressed" to "minimally distressed." For context, a 10-point change is considered clinically meaningful in psychiatric populations.

**Depression symptoms:** Both groups dropped by about 6 points on the CES-D (0–60 scale). This is roughly half the standard deviation — a moderate effect. A 5-point change is considered clinically meaningful.

**Perceived stress:** Both groups dropped by about 5 points on the PSS (0–40 scale). This is a moderate effect — equivalent to the stress reduction seen in 8-week mindfulness-based stress reduction programs.

**Trait anxiety (gardening only):** Dropped by 5 points on the STAI-T (20–80 scale). This is a medium effect — roughly equivalent to what you'd see after 8 sessions of cognitive-behavioral therapy for mild anxiety.

**In plain English:** After four weeks of twice-weekly gardening or art-making, healthy women reported feeling about 60% less mood disturbance, 45% fewer depression symptoms, and 25% less stress. Gardening also made women feel less anxious as a general personality trait. But neither activity changed their heart rate, blood pressure, or how satisfied they were with their free-time activities.

**What the authors acknowledge:**

Small sample size limits statistical power to detect between-group differences.

No inactive control group means they cannot attribute improvements specifically to gardening or art-making.

Healthy population may have limited room for improvement (ceiling effects).

Single sex (women only) limits generalizability.

No long-term follow-up to assess durability of effects.

Self-report measures are subjective and may be influenced by social desirability.

**What a critical reader would note:**

**No blinding of participants** — This is a major confound. If you volunteer for a gardening study because you love plants, you're primed to report feeling better. The art group might have had different expectations.

**No control for social interaction** — Both groups were group-based. The improvements could be due to socializing, not the specific activity. A "book club" or "board game" control group would have helped.

**Attrition** — 10 of 42 randomized (24%) dropped out or were excluded. If dropouts were different from completers (e.g., more stressed, less engaged), results could be biased.

**Multiple comparisons** — They tested many outcomes (POMS subscales, CES-D, PSS, STAI-state, STAI-trait, SDSA, WHOQOL-BREF, heart rate, blood pressure). Without correction for multiple testing, some "significant" results could be false positives.

**Funding** — Funded by the Horticulture Research Institute (industry group for horticulture) and the Florida Nursery Growers and Landscape Association. While the authors declare no competing interests, industry funding can introduce subtle bias in interpretation.

**Recruitment** — Flyers on a university campus and community outreach likely attracted people already interested in gardening or art. This self-selection limits generalizability.

**"Non-gardener" and "non-artist" criteria** — Excluding people with prior experience means the results apply only to beginners. Experienced gardeners might show different (or no) effects.

For someone running their own n=1 experiment:

### What to test

**Intervention:** Indoor gardening (e.g., potting plants, transplanting seedlings, caring for small container plants) OR art-making (e.g., collage, drawing, painting, mixed media). Do one or the other for 1 hour, twice per week.

**Dose:** 8 sessions over 4 weeks (minimum). The dosage-response data suggest benefits appear by session 4 and continue to grow through session 8.

**Format:** Group setting (the study used groups of 8–10 people), but solo practice might also work — the study can't separate group effects from activity effects.

### Minimum meaningful duration

**At least 4 weeks** (8 sessions). The biggest changes happened between baseline and week 2 (session 4), with smaller additional gains in weeks 3–4. A 2-week trial might show partial effects.

**No data on longer-term effects** — if you want to know if benefits persist, extend to 8 or 12 weeks with follow-up assessments.

### What to measure

**Primary metric:** Total mood disturbance — use the Profile of Mood States (POMS) short form (available online, 30 items, takes ~5 minutes). Alternatively, track daily mood on a 1–10 scale (1=terrible, 10=great).

**Secondary metrics:**

- Perceived stress — Perceived Stress Scale (PSS-10, free online)

- Depression symptoms — CES-D or PHQ-9 (free online)

- Anxiety — GAD-7 (free online) or STAI (requires purchase)

**Track before and after each session:** Rate your mood/stress on a 1–10 scale immediately before and after each session to see acute effects.

**Physiological (optional):** Resting heart rate (measure upon waking, before getting out of bed) and blood pressure (same time each day, after 5 minutes seated rest).

### Key confounds to control for

**Expectation effects:** If you believe gardening will help, you'll report feeling better. Try to go in with an open mind, or better yet, randomize yourself (e.g., flip a coin each week to decide which activity to do).

**Social interaction:** If you do the activity with others, the socializing might be the active ingredient. Try doing some sessions solo and some with a friend, and compare.

**Time of day:** Do sessions at the same time of day to control for circadian effects on mood.

**Menstrual cycle (if applicable):** Mood varies across the cycle. If possible, start and end your experiment at the same phase of your cycle, or track cycle phase as a covariate.

**Other life stressors:** Keep a daily log of major stressors (work deadlines, relationship issues, sleep quality) so you can see if changes in your metrics coincide with life events rather than the intervention.

**Novelty effect:** The first few sessions might feel better simply because they're new and exciting. The dosage-response data suggest benefits persist beyond novelty, but be aware of this.

### What a positive result would look like

**Acute effect:** Your mood/stress rating drops by at least 2 points (on a 1–10 scale) from before to after a single session.

**Cumulative effect:** Your weekly average mood rating improves by at least 1 point over 4 weeks (e.g., from 5/10 to 6/10).

**Clinical scales:** A drop of 5+ points on the CES-D (depression) or PSS (stress), or 10+ points on the POMS (mood disturbance) would be comparable to what this study found.

**Physiological:** A sustained drop of 3–5 bpm in resting heart rate or 5–10 mmHg in systolic blood pressure would be meaningful (though the study didn't find this).

**Dosage-response:** You should see bigger improvements after 8 sessions than after 4 sessions. If you plateau after 2 weeks, the intervention might have limited additional benefit for you.

**Bottom line for self-experimenters:** Both gardening and art-making appear to be effective mood-boosters for healthy women. Try either for 4 weeks (1 hour, twice weekly). Track your mood before and after each session, plus a weekly stress/depression questionnaire. If you see consistent improvements of 20–30% in mood and stress, you've replicated the study's findings. If not, try the other activity — the study couldn't distinguish between them, but you might have a personal preference that