how to deal with small sample?

[Tapas Paul]

In a pilot survey, I have about 200 samples and want to run an analysis for a large sample with these 200 samples. What are the possible procedures to deal with this small sample size?

[Felix Bittmann]

200 is not per se a "small" size IMHO. What model do you want to estimate? How many independent variables do you want to include? What is the outcome variable? I would suggest to estimate the desired model and then report effect sizes and confidence bands transparently. For robustness, you can compare various methods to estimate standard errors (e.g. regular, hc2, hc3, bootstrap). After having these first results one can think about additional checks.