Some statistics, called descriptive statistics, are primarily used to describe one or more samples of data. Other statistics, called inferential statistics, are used to infer whether variations in one variable predict variations in another variable. But by themselves, they cannot tell whether variations in one variable caused variations in the other variable. The probability values given by the statistical tests indicate the likelihood that the size of the effect could have occurred by chance. If that probability is less than 5%, written as p < 05, the result is said to be statistically significant. With very large samples, even small correlations can be statistically significant. Correlations smaller than r=.10 explain less than 1% and are considered trivial correlations in this book.