I will assume the sample is random. In general, the larger the sample, the smaller the percentage error will be (the difference between percentages in the sample, and the percentages in the universe from whence the sample is taken). The percentage error tends to go down as the square root of the size of the sample.
Chat with our AI personalities
A small sample is sufficient provided that it is representative: that is, the sample units are chose randomly.
Because then you can assess how valid your results are =D
The answers are usually always valid. What may or may not be valid are your assumptions about the underlying model. Also, the number of times the results should be similar depends on the number of possible outcomes and the variability in the outcomes. For example, if you spin a fair spinner with 12 equal segments, then the probability of similar results is less than likely.
Because without representative sample, your results will not be valid.
validity is whether the results are valid so the data has no mistakes of as such in it whereas reliability is the dependability; when the results you have are accurate and are of enough quality.
If you documented all your results, had a partner, had a witness, completed the experiment many times with the same results, and tested the experiment on the proper things then this would be good validation.