I have a dataset which consists of location ID's (1 to 29 different locations) and each location has a couple of repeated measures (max n = 780, min n = 50). Each measurement consists of a number of fish per dive (n = 0, 1, 2,...7 for example). Now I want to test if the location influences the amount of fish that are seen each dive, how can I test this?
$\begingroup$
$\endgroup$
0
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
It depends on if you think your data is best characterised by linear or non-linear tests.
The most simple of tests is most likely a Repeated Measures ANOVA, you could also average the number of fish per location -but any linear statistical test will lose out on a lot of the variability (that I assume) exists in your data.
Depending on what your data actually looks like you could probably run either RQAs (checks for self-similarity for one location) or CRQA which would check one location against another over time.
-
$\begingroup$ Since counts can't go below 0, the relationship can't truly be linear, but they may approximate it well enough for some purposes. $\endgroup$ Commented Jan 20, 2016 at 17:05
-
$\begingroup$ I don't think that's a strict requirement of linearity. Linear entails, roughly, that you get an output from an input that follows a linear trajectory. Put in 1, get 2, put in 2, get 4 (following an exponential equation for example). So it depends on assumptions about data. One way would be to explore the data for linearity but that is outside the scope of the question... $\endgroup$– PsyPhiCommented Jan 20, 2016 at 19:15