What does a Chi-Square test do?
1) Tells us if the difference between cell values on a table are large enough to be statistically significant
2) tells us if two variables are independent or dependent
What LOM are chi-square tests typically used for?
Nominal and ordinal variables
What is the observed frequency?
The cell frequency actually observed
What is the expected frequency?
The cell frequency you would expect to see if two variables are independent
What is the chi-square obtained statistic?
The test statistic that summarizes the differences between the observed and expected frequencies in a table
What is the sampling distribution of chi-square?
A family of
1) asymmetrical
2) positively skewed
Distributions who’s exact shape is
3) determined by their degrees of freedom
What are 3 notable characteristics of the sampling distribution of the chi-square?
1) It never approaches normality
2) As its DF increases, its distribution does become more symmetrical (but never normal)
3) Since it's positively skewed, the alternative hypothesis is always one-tailed (its values are always positive)
What are the 2 big limitations of a chi-square test?
1) Is very sensitive to sample size, with large enough samples a small relationship can seem significant
2) Only tells us if two variables are related, does not imply a causal effect
Explain the steps for running a chi-square test
1. State the null and alternative hypothesis and the alpha level
2. Calculate the expected frequency of EACH cell using the expected frequency formula
3. Using those values, use the chi-square formula for EACH expected frequency and add them together
4. Find the critical statistic using the chi-square table
5. State your conclusion