A researcher has a report for an experiment with a sample size equal to 100, the resulting 95% prediction interval was to large. If the researcher wishes to run a more precise experiment that will results in 95% prediction interval half the size of the earlier study, then the researcher will require what size of the sample?