Term
|
Definition
Tests the frequency of outcomes in all categories (groups) against a hpothesized distribution which is either
Equal distribution OR
pre-defined distributions |
|
|
Term
| What does chi square GOF test? |
|
Definition
| whether the observed frequency of observations n each category match what we would expect |
|
|
Term
|
Definition
1. state hypothesis
2. calculate expected counts under H0
3. check assumptions
4. calculate test stat
5. compare test stat to critical value, compare p to alpha
6. make a decision about the null
7. make a conclusion and state it in terms of alt. hypo. |
|
|
Term
|
Definition
1. data is from random sample
2. independent observations
3. all expected counts are greater than 1
4. at least 80% of expected counts are greater than 5 |
|
|
Term
| discribe the X^2 test statistic distribution |
|
Definition
| X^2 is a family of distributions that take only posive values and are skewed to the right |
|
|
Term
what do you do if:
test stat > critical value |
|
Definition
|
|
Term
| How would you state a conclusion for a chi-sq GOF test in terms of the alt. hypo? |
|
Definition
| "There is/isn't evidence to show that the distributions of "groups" does/doesn't correspond to "pre-defined distributions OR equal distribution" |
|
|
Term
| what does contingency analysis allow us to do? and what test do we run for it? |
|
Definition
| contingency analysis allows us to look at the relationship between two categorical variables. We run chi-sq independence to test for it |
|
|
Term
| what does the chi-sq independence/homogeneity do? |
|
Definition
| the chi-sq test of independence tests for association or independence between two categorical variables. It is used to test the frequency of outcomes in a contingency tabel to see if two categorical variables are independent or not. |
|
|
Term
| what are the hypo's for a chi-sq test of independence? |
|
Definition
Null: the two variables are not associated/independent
alt: the two variables are associated/dependent |
|
|
Term
| what are the assumptions for chi-sq indepn? |
|
Definition
1. data is from random sample
2. independent observations
3. all expected counts >1
3. at least 80% of expected counts are greater than or equal to five |
|
|
Term
| how do you calculate expected frequency for chi-sq indpt? |
|
Definition
| expected frequency = (row total X column total)/Grand total |
|
|
Term
|
Definition
|
|
Term
|
Definition
| df = (# of rows -1) X (number columns -1) |
|
|
Term
| How would you state a conclusion for a chi-sq indpt in terms of the alt hypo? |
|
Definition
| "We have/don't have evidence that 'variable 1' and 'variable 2' are not/are independent" |
|
|
Term
| if you conclude what you observed is differenet from expected, what can you do next and how? |
|
Definition
odds ratio: compare conditional probabilities acrosss groups
odds = p/(1-p)
odds ratio = odds a/odds b |
|
|
