Suppose that I made a survey with multiples questions where I have a target population of size nearly $N=100$ and a sample of size $n=8$. Is possible to create any type of confidence interval for the overall results in this survey?

How is the overall margin of error reported in surveys? I've seen some material where the maximum margin of error of the survey is reported as

$$MOE=\frac{0.98}{\sqrt{n}}$$

but in this case I have a really small sample so the normal approximation doesn't seem correct to be used. Is there anything I can do about it?