Seasonal migration in sea trout - What test to use?

#1
Discription:
Approximately 200 sea trout will be caught in the estuary outside a given river in the spring. During the capturing of the fish, a small low-frequency tag will be injected into the abdomen of the fish. That way, every individual is separated by an ID-number. When these fish return over the course of the summer they will swim past a pre-installed antenna while migrating up the river. This antenna will read the ID-tag we we installed. Now, for the statistical analysis:

The fish are separated into two groups in the dataset. Those who return (1) during the summer, and those which for some reason do not (0). Then we would like to investigate if some of the measured variables (ref. attachment picture) had a significant impact on the fish choosing to return or not.

Our educated guess.
This if the first time we try to apply statistics in such a way. We believe some the independent cartegorical variable to (return/not return). From what the investigation we have conducted, some sort of a Multiple-way ANOVA-test would be suitable for analyzing these variables.

Our challanges:
If anyone could point us in the direction of which test and perhaps some sources on the theory behind that test that would be very helpful.
Thanks in advance :)

OBS! The dataset is just for reference and does not necessarily reflect realistic values. If more info is needed I will try to provide that ASAP.
 

Attachments

Karabiner

TS Contributor
#2
Return/no return is the dependent variable. And you want to predict
whether a fish returned or did not return, using predictors such as...?

-- If you want to use several predictors jointly in one analysis, you have
to account for the binary level of measurement of your dependent
variable, i.e. you will possibly have to consider mutliple binary logistic
regression analysis.

About how large are the respective proportions of returners/non-returners?

Wth kind regards

Karabiner
 
#3
Return/no return is the dependent variable. And you want to predict
whether a fish returned or did not return, using predictors such as...?

-- If you want to use several predictors jointly in one analysis, you have
to account for the binary level of measurement of your dependent
variable, i.e. you will possibly have to consider mutliple binary logistic
regression analysis.

About how large are the respective proportions of returners/non-returners?

Wth kind regards

Karabiner
Sorry for a slighly inadequate explanation. We would not like to predict if a fish return or not, but rather if there were any siginificatnly impacts from the varibles that determined if the fish did decide to return or not.

E.g. A significantly higher weight in the returners. A significantly lower amount of sea lice on the returing fish.

We will tag approximetly 200 fish, and 70-80 of those are expected to return based on previous studies.
 

Karabiner

TS Contributor
#4
Technically you "predict" return/non-return using the variables measured in the past.

If you only want find out whether a single predictor is associated with return,
then you can use t-test (comparison of means for interval scaled variabkes,
such as weight), or, alternatively, U-tests. And you can use Chi² tests for categorical
predictors (such as sex etc.).

But if you want to use several predictors at the same time, within one model,
you should consider binary logistic regression, with return yes/no as dependent
variable.

With kind regards

Karabiner
 
#5
Technically you "predict" return/non-return using the variables measured in the past.

If you only want find out whether a single predictor is associated with return,
then you can use t-test (comparison of means for interval scaled variabkes,
such as weight), or, alternatively, U-tests. And you can use Chi² tests for categorical
predictors (such as sex etc.).

But if you want to use several predictors at the same time, within one model,
you should consider binary logistic regression, with return yes/no as dependent
variable.

With kind regards

Karabiner
Alright! Thank you so much for your answer. I will look into those tests.