reply to this post I think this is really important to note, especially when applying null hypothesis testing to research. Larger datasets are also better for representing a population and eliminating bias. As more data is collected, statistical significance is "easier" to determine or be observed. A visual comparison of this could be to take a classroom of students and ask about a genetic trait like tongue rolling. If the null hypothesis were that there is no difference in tongue rolling and the sample size was 5 we might have to accept the null hypothesis because we can't observe the difference within the small population. If this hypothesis test was applied to a classroom of 100 we could then clearly see there is a difference in tongue rolling ability, and therefore we can easily reject the null hypothesis