A meta-analysis synthesizes 30 published studies on a new anxiety treatment. The pooled effect size is $d = 0.52$ (95% CI: [0.43, 0.61]). However, a funnel plot reveals asymmetry: small studies cluster above the pooled estimate, while large studies cluster closer to zero. Additionally, studies reporting near-null effects ($d < 0.15$) are conspicuously rare despite being feasible.

What does this pattern suggest?