Do I need a two-way MANOVA test?

I am testing for differences between 3 groups. I have 2 dependent variables which are continuous data. However, I am confused about my independent variables and if a two-way MANOVA test would be suitable.

My three groups are Rural, Semi-rural and Urban (3 groups). For each group I am testing for differences in two ecoacoustic metrics, NDSI and BIO (2 dependent variables). My independent variables are the time of day (testing for differences in these metrics between dawn and dusk) and season (testing for differences in these metrics between summer and winter). I am unsure if I have 2 independent variables (season, time of day) or 4 (dawn, dusk, summer, winter).

I am unsure if I need to run a two-way MANOVA for this, or run 2 two-way MANOVAs, one for season and one for time of day. Or is there another statistical test which would be better suited to what I am doing?

Thank you in advance


TS Contributor
Who or what are the subjects of your study (persons?)? Are there repeated measures (e.g. the same
study subjects were measured in summer and in winter)? And how large is your sample size? Do
you expect an interaction between the time-of-day effect and the season effect?

With kind regards

The 3 groups consist of 3 sites each (total of 9 sites). Yes, the same sites were surveyed in the summer and winter. The sample size is large but varies considerably between groups and between seasons. In total there are around 2000 observations. I think I expect time-of-day and season to have independent effects.

EDIT: Perhaps I have been thinking about this wrong. I am looking for the difference in effect of habitat type, time of day and season on 2 dependent continuous variables, NDSI and BIO. Does this mean that I have 3 independent variables?

Habitat types: Rural, Semi-rural, Urban (3)
Seasons: Summer, Winter (2)
Time-of-day: Dawn, Dusk (2)

Dependent variables: NDSI, BIO (2)

Hope this makes it a bit clearer.

What kind of statistical test would you recommend for this?

Thank you
Last edited: