PALS0046: Intro to Statistical Methods - Hypothesis testing

Research hypothesis

States the actual prediction of a relationship - describes the predicted effect

Statistical hypothesis

Split into null hypothesis and alternative hypothesis

Null hypothesis (H0)

Expected result if there was no effect - number of correct responses is no different to what we would expect by chance if guessing

Alternative hypothesis (H1)

Expected result if there is an effect - number of correct responses is greater than or less than we would expect by chance

Type 1 error (alpha)

Rejecting null hypothesis when it is true (false positive)

Type 2 error (beta)

Accepting the null hypothesis when it is false (false negative)

Test statistic

Statistical value that expresses our data

Effect size

Strength of observed effect, e.g., the size of the difference between two groups or the strength of the association between 2 variables

P-value

probability of observing this particular value of our test statistic assuming that H0 is true

Alpha level

Threshold that p-value must reach for us to reject H0 (e..g, when alpha = 0.05, we need to observe p<0.05 to declare the results statistically significant)

Beta level

Type II error rate (false negative = failure to detect true effect)

Statistic power

Test's capacity to detect a true effect (power = 1 - beta)

Bonferroni correction

Correction we must make to our critical threshold (alpha) to account for multiple comparisons - corrected alpha = 0.05/number of comparisons