T-statistic

  • allows researchers to use sample data to test hypothesis with an unknown population mean, unlike the z test
  • standard error of mean: distance between sample and population mean
  • sample data can now estimate the distribution
  • as the sample size increases, the t-distribution approaches the z-distribution
  • with small df, the t distribution is smaller and more spread out
    • critical values increase
  • on t-table, always look at the two-tailed column T hypothesis tests
  1. state the hypothesis and select a value for a
  2. locate the critical region
    • the sample must be independent observation
      • no relationship between observations or the sample
    • the underlying population from which the sample is drawn must be normal