What kind of statistical test would I use for this experiment?

#1
Hi there, I'm sorry I'm still very new to this platform and am generally clueless about statistics. This may be a little long winded, but I'd really appreciate any help!

I have two groups of organisms, all of the same species but from two different habitats, labelled "group 1" and "group 2" as well as a control group for both habitats "c1" and "c2." All organisms are female. I have a separate group of males.

I would like to see if dietary vitamin C levels effect the number of eggs produced by the females, so I have set up a hypothetical experiment as follows:
The control group is given food containing maximum Vitamin C. Individuals in group 1 and 2 are started off with food containing little Vitamin C, allowed to breed and the number of eggs recorded. They are then placed on a diet containing moderate Vitamin C, allowed to breed again and the number of eggs recorded. This is then done once more with an even higher dosage of Vitamin C.

I am quite clueless when it comes to statistical tests; I am not sure how many factors this experiment would be, I thought perhaps 2 (level of Vitamin C and habitat type). I assume it has three levels (little Vit C, moderate and high levels). And also then which statistical test I would use to check significance. I have been suggested to use some type of ANOVA, but unsure which and for what reason...

Any help at all would be greatly appreciated!
 

fed2

Active Member
#2
thanks for your eggs-cellent question. I would be thinkin' that it was repeated measures ANOVA? But maybe issue with normality assumption given what i know about how eggs. Also check it out: Time is confounded with Vitamin C dose, which may be bad or not depending on whether one believes the number of vitamin C doses of any concentration might have an effect.

Good luck, and may god have mercy on these experiments.
 
#3
thanks for your eggs-cellent question. I would be thinkin' that it was repeated measures ANOVA? But maybe issue with normality assumption given what i know about how eggs. Also check it out: Time is confounded with Vitamin C dose, which may be bad or not depending on whether one believes the number of vitamin C doses of any concentration might have an effect.

Good luck, and may god have mercy on these experiments.
I have never heard of this type of ANOVA but after searching it up, it seems about right! :) Thank you for your response! Do you by any chance know how many factors this experiment would be? I know it's not really a statistics question, but I'd like to mention it somewhere, I'm just not sure if it'd be a one or two factor study considering I have two independent variables (except the habitat variable is there as more of a control to see if there are any differences, not REALLY something that is being investigated....)
 
Last edited:

fed2

Active Member
#4
Yeah, it may not fit exactly into the Repeated measures ANOVA box, but it is something like that. You usually have a time or 'within subjects' factor and a between subjects or 'treatment' factor. For you, i think vitamin-C is basically the time factor, even though it is not time ,it is identified with it. habitat is probably best to treat as the between groups factor. The only thing is the control group is only at 1 vitamin C level, which is fine, essentially this will give a group of subjects measured at only one 'time' point. Alot of software will balk at this, but many will flip to 'linear mixed effect' model, which is modern version of repeated measures ANOVA. I know prism does this.

To simplify, essentially what you get is paired t-test to compare within the animals with multiple vit c groups, and 2 sample t-test to compare to the control, although not exactly.

good luck.
 

noetsi

Fortran must die
#5
You have threats to validity since events that occur can bias your results (formally this is called history). Something that occurs during the period in question accidently can distort your findings.