I would say that studies on small sample sizes are not without any worth, however obviously they are very limited in what they can do. Obviously the ideal is a robust study looking at as many individuals as possible to create a representative sample, but case studies can still highlight interesting issues that are worth further investigation. The main issues you need to consider are time and funding - is focusing on a small group really likely to yield any notable results (would studying it highlight any information not already known say? Or perhaps something rarely documented?) or is it a waste of time and money that could more effectively be spent elsewhere?

Sabhrina -

None of my expertise was in animal studies, so perhaps I am missing what would be obvious to you, but I wonder what kind of information you are gathering. If it is for characteristics that would describe a theoretically infinite population - say a superpopulation from which the animals that exist in nature are just examples - then a small sample size (n) is more of a problem than if you are describing something only about the finite population (of size N) that actually does exist. If you are trying to gather information just applicable to a finite population, then you might need to consider the finite population correction factor. (See the link for a definition.)

Thus, if your finite population of size N is small, it takes a smaller n to know a great deal about that population. In the extreme, if N is so small that n = N, then you have a census of that finite population.

I assume you are using continuous data. ???

Sorry if this is not what your question is about, but I thought it worth mentioning, just in case it might pertain to your question.

Cheers - Jim

Article FINITE POPULATION CORRECTION (fpc) FACTOR

Please also note that the expression "a representative sample" has different meanings for different people in different contexts.  Please note the discussion found at the link attached.    

https://www.researchgate.net/post/What_is_a_representative_sampling

Small sample sizes are in certain cases inevitable, especially when dealing with elusive or rare species, but provided your experimental design is carefully developed they can still provide biologically meaningful results, with important consequences for aspects such as species conservation or management.

It is very important to try and design your experiments or observations so that you will make the most out of them. In fact reducing sample size may be critically important when working with endangered species and/or subjects to be kept in captivity, and in such cases the 3Rs guiding principles (replacement, reduction and refinement) should be carefully taken into account.

Dear James R Knaub,

We never sure how the population exist in nature, actually. We only know that some species were detected more often than other, and it doesn't even always mean that the species not detected more often is less abundant. Therefore, I could not really sure if I could use the term "finite" (though they are surely not infinitive!)

However, it is actually interesting what you gave me, the finite population correction factor. I have not yet really understood how it could be used but I got the idea from your brief explanation and considering it.

Sorry for the late response.

Hypothetically, I get a manuscript to review that describes biological data from 6 mammalian predators. I have no way to really evaluate the results, so what I will do is look very carefully at the methods. Did the authors go above and beyond to try and get more samples? Did they use appropriate collection methods? Is there any reason to suspect that the outcome was due to incompetence or ignorance? If the methods were appropriate, then the data are the data.

I would also consider what other research organizations are doing. If there is an existing program that is tracking several hundred Siberian Tigers, and your program is able to track a single tiger, then it will be difficult to justify your effort.

I also would consider what data are gathered. If the big research organizations are only gathering minimal data, and you can offer more complete data (but on fewer individuals) then it might be worth doing (subject to cautions raised by Danilo). In this case, I would also question how my observations could influence the subject being observed.