2
$\begingroup$

I want to build a linear mixed-effects model.

Condition was changed within subjects, so as fixed effect, I entered condition and as random effects subjects, items and a by-subject random slope for condition.

In a second step, I wanted to include scores of a questionnaire that were collected from participants before the experiment and are believed to influence the dependent variable as well.

Thus, is this a fixed or a random effect?

Improve this question
$\endgroup$
2
  • $\begingroup$ Is there just one or one for each item? Do you mean item by pre-experiment value (influence presumably differs by item?)? Subject by condition may be unusual, but perhaps not in your field. $\endgroup$ Commented Apr 25, 2017 at 19:30
  • $\begingroup$ @Björn There is one questionnaire score for each subject (indicating how stressed they were before the experiment). $\endgroup$ Commented Apr 27, 2017 at 11:34

1 Answer 1

Reset to default
5
$\begingroup$

In a second step, I wanted to include scores of a questionnaire that were collected from participants before the experiment and are believed to influence the dependent variable as well.

There is one questionnaire score for each subject (indicating how stressed they were before the experiment)

is this a fixed or a random effect?

The questionnaire score should be a fixed effect in your model, unless it is a mediator, in which case (assuming you are interested in the total effect of condition, and not just the direct effect) it should not be in the model at all. It is always a good idea to draw a DAG (ie., a causal diagram) which informs the set of variables to include in a model. See my answer in the following thread for details:

How do DAGs help to reduce bias in causal inference?

Improve this answer
$\endgroup$

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.