Test Statistic

 

What Is a Test Statistic?

A test statistic is a numerical value calculated from sample data in order to decide whether to reject or not reject a statistical hypothesis.

  • It is computed from sample statistics (sample means, variances, proportions, etc.).
  • Often it is a transformation or “second processing” of those basic sample statistics, designed so that its distribution under the null hypothesis is known (e.g., t, z, F, chi-square).

Formally:

A test statistic is a function of the data used to summarize the evidence against the null hypothesis.


T-value (t Test Statistic)

The t-value is a type of test statistic used mainly to compare means.

Typical use:

  • To test the difference between two sample means.
  • Or to test whether a single sample mean differs from a hypothesized population mean.

When comparing two groups, we are interested in:

  1. The difference in sample means, and
  2. The uncertainty (standard error) of that difference.

The t statistic combines these into a single number:

\[t = \frac{\overline{X}_1 - \overline{X}_2}{\,S_{\overline{X}_1 - \overline{X}_2}\,}\]

where

  • (\overline{X}_1), (\overline{X}_2): sample means of group 1 and group 2,
  • (S_{\overline{X}_1 - \overline{X}_2}): standard error of the difference in sample means.

Intuition:

  • Numerator: how large is the observed difference in means?
  • Denominator: how much random variation do we expect in that difference?

A large absolute t-value means that the difference in means is big relative to the noise, suggesting stronger evidence against the null hypothesis (e.g., “the two population means are equal”).


What Counts as a “Large” T-value?

The t-value by itself is not enough; we need to know its sampling distribution under the null hypothesis.
Under typical assumptions (normality, equal variances, etc.), the t statistic follows a t distribution with some degrees of freedom.

For a two-sided test at significance level (\alpha = 0.05):

  • We look at the upper 2.5% and lower 2.5% tails of the t distribution.
  • These cutoffs (critical values) define what is “sufficiently large” in magnitude.

In other words:

  • If (t) falls in the top 2.5% or bottom 2.5% of the t distribution (i.e., ( t ) is greater than the critical value),
  • Then we reject the null hypothesis at the 5% significance level.

So:

“Sufficiently large t-value” = a t-value that lies in the rejection region (extreme tails) of the t distribution for the chosen significance level and degrees of freedom.