Correlational research & NHST

Large correlation (r close to 1) = small p value

• Large correlation = less agreement with the null hypothesis = small p value

Statistically significant: H₀ is rejected (large p value)

Type I error: incorrectly rejecting the null hypothesis - concluding that there exists a relationship between two variables when there is none

• alpha = 0.05 → acceptable cutoff point for social sciences. Meaning there is a 5% chance of making a type 1 error. May decrease as consequences of the error get more serious

• p > a → non-supporting data leads to revision of the hypothesis or improved research design

• p <= a → the data supports the theory

Type II error: incorrectly not rejecting the null hypothesis → believing there is no relationship when there actually is one

• This may be due to obtaining a large p value

Difference between r and rho:

• rho is the relationship within the population, r is the relationship within the sample