Statistical significance

From Conservapedia
Jump to: navigation, search

In statistics, it is required to know whether two groups of data, both subject to random variation, should be reqarded as distinct or not in order to draw meaningful conclusions. That is, are the observed differences between the groups due to chance or not? This is tested by initially assuming the null hypothesis (typically that there is no difference between groups A and B), then computing a test statistic that has a known probability distribution under the null hypothesis (called a p-value). If the calculated p-value falls under the given α-level (typically .05, but sometimes .01), the difference is statistically significant. This means that the data gathered provides significant evidence that the two groups are different, as the results couldn't have happened by chance.