Test statistic

A test statistic compresses a whole sample into one number whose behavior under the null hypothesis you already know, which is what lets you convert it into a p-value. The common ones in an introductory course are z, t, and $\chi^2$ (chi-square), and the first two share the single recipe described in standardized test statistic. For example, $z = 2.10$ says the sample proportion sits 2.10 standard errors above the value the null hypothesis claims. The larger the magnitude, the further the data are from the null, and the smaller the p-value.