Task 6 Flashcards
What are the assumptions of a z-test for 1 proportion?
- Variable = categorical and dichotomous
- Normality
What assumption is always violated for a z-test for 1 proportion?
normality therefore ensure samples are large enough -> CLT
What is the Ho of a z-test for 1 proportion?
Ho: π = πo (i.e. population proportion equals a certain value
What is the unbiased estimator in a z-test for 1 proportion?
sample proportion (p aka p^)
How do we find the standard deviation involved in a z-test for 1 proportion?
If we know the mean proportion then we automatically know the standard deviation (simple- if you know one proportion you know the others since we have 2 so you can draw the distribution in a bar chart )
How to calculate the standard deviation of a z-test for 1 proportion?
top line of standard error formula
How do you calculate a sample proportion?
p = x/n e.g. = 0.14 -> 14%
What is the z-test for a population of 1 proportion?
p - πo / SE
How to find a confidence interval for a z-test of 1 proportion?
p +- z* (SE)
-> In SE formula use sample proportion to estimate
What are the assumptions for a z-test with 2 proportions?
- dependent and independent variables = categorical
- normality
- groups are independent
What is the Ho of a z-test with 2 proportions?
Ho: π - πo
How to calculate z-test for 2 proportions?
p1 - p2 / SE
What is the most efficient estimator of the population proportion?
weighted average
x1+x2 / n1+n2
What is the expected counts rule for X^2 goodness of fit test?
all counts in the frequency table must = 5 and taken together = 20
Why is X^2 goodness of fit non-parametric?
It doesn’t use parameters
What is the Ho of a X^2 goodness of fit?
follow known a distribution (25% 50% 25%
Is the Ha of X^2 of goodness of fit?
two-tailed, does not follow distribution or no association
How do you calculate expected counts?
N x πi
What is the Ho of X^2 for contingency tables?
there is no association between the 2 variables x and y
X^2 for contingency tables aka
X^2 for independence
Expected counts assumption for X^2 for contingency tables
are at least = 5
Why does the normality assumption not apply to X^2 tests?
it is non-parametric!
What is the difference between the 2 X^2 tests in relation to variables?
X^2 for contingency tables -> 2 categorical variables
X^2 for goodness of fit -> 1 categorical variable
