stats final exam

two sample t-test

• two separate groups

• means

• ASSUMPTION: normality

correlation

• numeric x numeric

• measuring relationship between two variables

one sample proportion test

• binomial test

• one sample, one proportion

• ASSUMPTIONS: none

chi-squared goodness of fit

• comparing proportions to each other simultaneously

• fair v not fair

• ASSUMPTIONS: normality, expected value of at least five successes

chi-squared test of independence

• proportions

• categorical x categorical

• ie no religion has higher MNS proportion than others

• ASSUMPTIONS: normality

• null hypothesis: two categorical variables are independent

analysis of variance (ANOVA)

• numeric x categorical

• football conference and GPA

• ASSUMPTIONS: normality, equal variances across groups

• null hypothesis:

  • all population means are the same

  • numeric independent of categorical

  • model does not improve prediction ability

• can’t tell which mean is different

hildebrand rule

• symmetry test for populations

Shapiro test

• normalcy test

• sufficiently normal if p value is greater than 0.05

• null hypothesis = data from sufficiently normal distribution

one sample t-test code

t.test(x,mu)

two sample t-test code

t.test(x,y)

fligner.tes(x ~ g)

• tests for equal variances

how to reject null hypothesis

• p value below alpha