Test 3 Flashcards
Stats tests useful for normally-distributed data are called
Parametric tests
When is a distribution considered skewed
When the median differs from the mean
Negative skew- mean < median
Positive skew- mean > median
Kurtosis, positive vs negative
A measure of the extent to which observations cluster around the mean
Positive kurtosis= more cluster
Negative kurtosis= less cluster
What test can you use to calculate if groups are normally distributed and have equal variances
Levenes test
Possible measures in nominal data
Mode, range, frequency distribution
Possible measures in ordinal data
Same as nominal plus-
Median, percentile, semiquartile range
Possible measures in interval data
All ordinal tests
Mean
Standard deviation
Addition/subtraction
Possible measures in ratio data
All measures possible
Can make comparisons
Type one error with null hypothesis
Not accepting the null hypothesis when it is actually true
Type two error with null hypothesis
Accepting the null hypothesis when it is actually false
What type of test is he looking for if he says “I want to see changes over time”
Survival test
What type of test is he looking for if he says “I want to PREDICT the likelihood of some outcome”
Regression test- he will have the word “predict” in question stem
Test used for 2 groups of independent nominal data
Pearsons chi square test
Test used for 3+ groups of independent nominal data
Chi-square test of independence
Test used for 2+ groups of nominal data with expected cell count LESS THAN 5
Fishers exact test
Chi square test will not work with expected counts less than five
When do you used post-hoc testing
For statistically significant findings in 3 or more comparisons, in order to determine which groups are different
Post-hoc test used for nominal data
Bonferroni test of inequality
Paired/related data
He must state that they collected data throughout the study, AND that they compared those sets of data (data sets must be from same group as well)
“Pre vs Post” “Before vs After” “Baseline vs End”
Test used for 2 groups of paired nominal data
McNemar test
Test used for 3+ groups of paired nominal data
Cochran test
Test used for 2 groups of ordinal data
Mann-whitney test (can be used for interval data not meeting parametric requirements)
Test used for 3+ groups of ordinal data
Kruskal-Wallis test (can be used for interval data not meeting parametric requirements)
Test for 2 groups of paired ordinal data
Wilcoxon signed rank test
Test for 3+ groups of paired ordinal data
Friedman test
When would you use a student-newman-keul test and what does it do
(Post-hoc for ordinal data)
Used when Friedman test gives a statistically significant value
Compares all pairwise comparisons possible
All groups must be equal
When would you use a Dunnet test and what does it do
Post hoc for ordinal. Used when friedman test gives statistically significant value
Compares all pairwise comparisons against a single control
All groups must be equal
(Spoke and wheel picture)
When would you use a Dunn test and what does it do
Post hoc for ordinal. Used when Friedman test gives statistically significant value
Compares all pairwise comparisons possible
Useful when all groups are NOT equal size
Test for 2 groups of independent interval data
Student t test
Test for 3+ groups of independent interval data
Analysis of Variance (ANOVA)
Test for 3+ groups of independent interval data with confounders
Analysis of Co-variance (ANCOVA)
Test for 2 groups of paired interval data
Paired T test
Test for 3+ groups of paired interval data
Repeated measures ANOVA
Test for 3+ groups of paired interval data with confounders
Repeated measures ANCOVA
Correlation test for nominal data
Contingency coefficient
Ordinal correlation test
Spearman correlation
Interval correlation test
Pearson correlation
Nominal survival test
Log-rank test
Ordinal survival test
Cox-proportional hazards test
Interval survival test
Kaplan-Meier test
Nominal regression test
Logistic regression
Ordinal regression test
Multinomial logistic regression
Interval regression test
Linear regression
