# Mediation analysis with JASP reporting

Hello JASP-ers,

I have a question related to performing mediation analyses in JASP. I have a very sim,ple mediation model (1 predictor, 1 mediator, 1 outcome (allof these are mean accuracies) and 1 "counfounding" (age in moths)). Previsously I have used PROCESS in SPSS (I have a very limited experience with mediation in general), but I find the one in JASP much more pleasing :) . However, I have some difficulties understating what is reported. For example, the z-value in the table. Is this the result of the Sobel test? When I plot the path analysis I clearly see the values for the direct effects, but there are certain values over the arrows next to the variables themselves. I can not seem to understand what these values refer to. I see the arrows are dashed and solid, and I assume this is something related with direct and indirect effect, but could someone may be clear these questions for me or just point me to a documentation, where I can read about it?

Many thanks,

Mila

## Comments

Hi Mila,

For some reason the png files do not work for me (I'll ask Sebastiaan what's up). This makes it difficult to answer your question. Could you specify exactly which analysis option you are using? As soon as I know I will pass on your question to the relevant team member.

Cheers,

E.J.

Hi E.J.,

Thank you for your answer. I have attached a pdf now, so may be it will work. Otherwise, I for this mediation, I am using the standard method option (the results are in attached to)

Thank you in advance,

Mila

Hi Mila,

the "estimate" column in the tables represent the effect estimate. For example, the total effect shows that a one point change in WoDi_DiWo_mean leads to, on average, a change of 1.005 in DiDo_DoDi_mean. The standard error is a measure of uncertainty about this effect size, which decreases as the sample size gets bigger. This results in the z-value (which for the standard method is simply Estimate / Std. Error) and the p-value comes from comparing this value to a normal distribution.

You can interpret the indirect effect table similar to the Sobel test. If the indirect effect is significantly different from zero, you can reject the hypothesis that there is no indirect effect. I do recommend that you assess this not with the standard s.e. method, but with the bootstrap method as it is known to result in more accurate standard errors and p-values.

Hope this helps!

Erik-Jan

Hi,

Sorry for my late response, but thank you very much for these clarifications.

Best

Mila