# I have done a Hayes Mediation analysis with 2 mediators. My indirect effect is very small but significant. How do I interpret this?

I looked at the Effect of ADHD on Global Seasonality Score (GSS), mediated by both Sleep onset time on Free days and Sleep duration on Free days (N= 1738).

The indirect effect of ADHD on seasonality through mediator Sleep onset time (free days) = c, c =0.003 (0.001 - 0.007)

Effect of second mediator sleep duration (free days) on GSS was not significant.

c =0.000 (-0.003 - 0.002) Sleep duration (free days)

My question is, even though effect of ADHD on seasonality through mediator Sleep onset time (free days) is very small, it is significant. c =0.003 (0.001 - 0.007) .

How do I interpret this? What can I say about the effect size of this mediator?

The value of the unstandardized indirect effect $I_u$ depends on the scales of the predictor (ADHD) and the dependent variable (GSS). It is hard to tell whether 0.003 is large or small. It is better to calculate the standardized indirect effect $I_s$ (e.g., Cheung, 2009). Specificity, $I_s = I_u\frac{s_x}{s_y}$, where $s_x$ and $s_y$ are the standard deviations of $x$ and $y$, respectively.