chapter 9 - null hypothesis testing

correlation size

general rules of thumb for significance are r=0.20 is small, r=0.50 is medium, and r=0.80 is large

hypothesis

a predictive statement regarding relationships within or among populations (not samples)

null hypothesis (H₀)

no relationship between x and y, generally trying to reject this, H₀: p=0

alternative hypothesis (H₁)

yes relationship between x and y, what the study is investigating, H₁: p≠0

rho (ρ)

population correlation coefficient

sampling distribution

data or samples you have

central limit theorem

as sample increases above 25, the distribution will take on a normal shape

alpha (α)

the probability with which are willing to reject the null, the significance level (p value), if α<0.05 we reject the null, if α>0.05 we fail to reject the null

degrees of freedom (df)

df = n-2, use critical value to compare test statistic: r<r_crit reject null, r>r_crit fail to reject null

summary of hypothesis testing

1. write null and alternative hypotheses in words, symbols, or both

2. determine alpha for rejecting null / find r_crit

3. compute test statistic (pearson)

4. compare r to r_crit and reject/fail to reject null

5. state conclusion in words

   1. direction of correlation, nature of relationship, and how it applies to problem