hello there! I think that the only way to properly define the correct sample size is through preliminary experiments. Different surveys/experiments all need differennt sample sizes, depending on the variables measured. when working with human responses to a certain situation, say, the variance is likely to be much larger than when working with human variances in body temp for example (only a few degrees compared to many different opinions...). so, im not sure what exactly ur project is, but when u get to ur stats, if ur variance is large, then u need to continue the surveys, increase ur sample size and this will reduce variance making any results more significant and reliable.
imnot sure if this makes sense exaclty, the point is use whatever sample size is necessary for minimal variance, and always do prelim trials first! :)