Randomized or Quasi-Experimental Designs Isolating the Causal Effect of GAH/PB on Anxiety

What the evidence says

A meta-analysis of randomized controlled trials found that positive psychology interventions increase subjective well-being relative to control conditions, with a pooled effect of d = 0.34 (95% CI [0.22, 0.45], p<.01) — a small-to-moderate effect. This estimate carries moderate heterogeneity across studies (I² = 49.5%), publication bias was detected in the pooled sample, and some included studies were rated low quality, which lowers confidence in the precise magnitude while leaving the direction well supported.

The same body of trials found that positive psychology interventions decrease depressive symptoms relative to control conditions, d = 0.23 (95% CI [0.09, 0.38], p<.01) — a small effect, smaller than the well-being effect above. Heterogeneity here is comparable (I² = 47.0%) and publication bias was again detected.

Within the depression outcome, effect size varies by how the intervention is delivered and to whom. Longer interventions produce larger effects than shorter ones: those lasting more than 8 weeks show d = 0.68 versus d = 0.15 for interventions of 4 weeks or less. Delivery format also moderates the effect: individually delivered interventions show d = 0.88, group-delivered interventions d = 0.47, and self-help formats d = 0.15. And targeting matters: interventions offered to people with specific psychosocial problems show d = 0.78, versus d = 0.16 when offered without such targeting. All three moderator breakdowns come from small subgroup samples (as few as 2 studies in one arm), so while the meta-regression slope for duration reached p=.02, confidence in these subgroup magnitudes is rated medium-to-low even though the overall depression effect itself is rated high-confidence in direction.

Effects on subjective well-being partly persist over time but shrink: at 3-6 month follow-up, the well-being effect was d = 0.22, down from d = 0.34 at post-test. This follow-up estimate rests on only 6 studies with high attrition rates, though heterogeneity across those follow-up studies was low (I² = 1.1%). The depression effect, by contrast, became non-significant at 3-6 month follow-up according to the same source, though no follow-up magnitude for depression is given in the claims.

No claims here address outcomes beyond subjective well-being and depressive symptoms — there is no measured effect reported for anxiety, functioning, or other mental-health outcomes, and no claim addresses effects beyond the 6-month follow-up window.

Where the evidence splits

No contradictions in the current sources.

Thresholds and dose-response

Duration is the one dose-response relationship the sources quantify: for depressive symptoms, interventions lasting more than 8 weeks produce a pooled effect of d = 0.68, compared with d = 0.15 for interventions of 4 weeks or less — a roughly four-and-a-half-fold difference in effect size tied to the longer-duration threshold.

Open empirical questions

• Whether well-being or depression effects persist beyond the 3-6 month follow-up window is not measured; the longest follow-up data reported covers only 6 studies with high attrition.
• What mechanism explains why individually delivered interventions outperform group or self-help formats is not established.
• Whether the duration dose-response continues to increase past 8 weeks, or plateaus/reverses at some longer duration, is not measured.
• No claims address outcomes beyond subjective well-being and depressive symptoms (for example anxiety or functional outcomes), so their effect, if any, is unmeasured here.