Test Statistic

A test statistic is a single number computed from your data that measures how far the data stray from what the null hypothesis predicts, scaled by the noise. A common form is (observed − expected) ÷ standard error, so it answers “how many standard errors away from ‘no effect’ are we?”

Its value is only meaningful next to its reference distribution — the values the statistic would take if the null were true. A test statistic of 2.5 “standard errors out” is unusual under a normal reference; that unusualness is exactly what gets converted into a p-value.