Should I have a variable as random?

Asked Oct 03 '19 at 13:55

Active Oct 04 '19 at 18:25

Viewed 28 times

I have put individual ID as a random variable in my model, but some individuals only have 1 record and others have 19 records.

Does this mean that individual ID shouldn't be a random variable?

Here's the context: I have birds visiting a feeder and want to know if the length of a visit is determined by species (3 options), site (3 options), and indiviudal ID.

The model I'm running currently is

lenght ~ species + site + (1|id)

edited Oct 03 '19 at 14:57

asked Oct 03 '19 at 13:55

Rachael

No, it does not mean that. – user2974951 Oct 03 '19 at 14:01
1

Unless you tell us more about what is your data and what are you trying to achieve, this is impossible to answer. – Tim Oct 03 '19 at 14:15
@Tim, sorry! I'll add some context. – Rachael Oct 03 '19 at 14:56
@user2974951 thank you for your response – Rachael Oct 03 '19 at 14:56
1

Thanks, I think it is already answered in here: https://stats.stackexchange.com/questions/242821/how-will-random-effects-with-only-1-observation-affect-a-generalized-linear-mixe is it? – Tim Oct 03 '19 at 15:35
Thanks @Tim for the help! – Rachael Oct 04 '19 at 18:19

Should I have a variable as random?

0 Answers0