Greetings!

Sample size for a study is ideally calculated by a person qualified in Statistical Analysis, depending on the sample size which was chosen for similar studies earlier. Hence we have to procure published articles very similar to the new work considered for research and then the Statistics person would determine if the sample size is appropriate for the new work - this in turn also depends on the study objectives and the proposed hypothesis - in short, the results derived from this sample size should be possible to be extrapolated to the actual target population (from which the sample size was derived). Hence, depending on the size of actual population the sample size should be a good representation of the same.

Now considered the number 45 was derived at, after taking into account all these, the sample size for pilot work can be ten percentage of this sample size (roughly). The purpose of the pilot work would be to validate the research methodology and or the tools involved in the same - for example in order to validate a tool we may have to do a semi validation first and then full validation subsequently, for which you need two pilot works - all the samples involved in the pilot work cannot be involved in the actual work.

To sum up, I feel that five as sample size maybe sufficient for one pilot work - but all these depends on actually what the study is upon, the research objectives and so on and forth

Basically, you should first start with a smaller sample in order to test the reliability of your methodology, than, you adjust accordingly and get to share it with a larger sample, for the size, in order to be representative usually it's between 10 to 15% of the population.

You can conduct pilot study by considering a minimum of 10% of total sample size

Conventionally, 10% of your sample size (after adjusting for non-response).

.. in this case, 4.5: approximately 5.