by Albert Major
HYPOTHESIS TESTING
• A hypothesis is a clear statement about a relationship or difference you expect to see in the population.
• It is not a question. It is a claim that your statistical test will evaluate.
• Example hypothesis types:
• “More hours spent studying increases grades.”
• “Playing video games predicts stronger social relationships.”
• “Word frequency predicts reading time.”
• In CSAI research, hypotheses appear everywhere: comparing groups, testing learning effects, checking model performance, or studying cognitive processes.
NULL AND ALTERNATIVE HYPOTHESES
• Null hypothesis (H0): there is no effect, no difference, no relationship in the population.
• Alternative hypothesis (H1): the expected effect exists.
• Statistics evaluates whether the observed data are too unlikely under the null.
THE GOAL OF HYPOTHESIS TESTING
• You try to rule out chance as the explanation for your findings.
• Because you sample from a population, sampling error will always exist.
• Statistical testing helps you decide whether your observed result is compatible with the null.
TEST STATISTICS