types of data and tests Flashcards
(32 cards)
Nominal/Categorical
- does not represent a numeric value
- discrete
- not ordered or ranked
Discrete Variable
Counting, Integers, non-negative integers, a positive integer, or only 0-1
continuous variable
measuring, can be uncountable sets of values
Non parametric statistics
Does not rely on data’s assumptions of distribution
How is this possible?
Chi-Square
- test for independence
- checks if the categorical variables are related
- for large samples
- non parametric
Fisher Exact
- checks if the two categorical variables related
- small samples
- non parametric
Binomial (test)
-Checks to see if the proportions of sample population same as measured population
-non parametric
(Reminds me of stratification)
Cochran Q Test
testing mean of an experimental unit/one participant that was tested under multiple variables,
non parametric
what test category do nominal stats get analyzed by
non parametric
Phi Coefficient
correlation coefficient between variables that can only have 2 meanings
Contingency Coefficient
degree of association between 2 categorical variables
what are the possible results of a contingency coefficient
zero to one
zero is total independnece
Tetrachoric r
- correlates nominal data
- index of two variables that could be ranking, but are now dicontimized
What are the characteristics of ordinal data
discrete
rank/order
no true zero
no set space/distance between variables
what type of measurement is temperature: nominal, ordinal, interval, or ratio
interval
Mann-Whitney U Test
compare two 2 independent samples of ordinal data or contrast significant differences between 2 sets of rankings
what is the non parametric equivalent of the t test
Mann-Whitney U Test
Kolmogorov-Smirnov
-non parametric
-compare two ordinal sets of data against each other
OR
-compare test data against expected or known data
Wilcoxon Matched-Pairs Signed-Ranks Test
group correlation
tests if two independent samples were taken from same poulation
what test is done for more than 2 independent rank groups
Kruskal-Wallis One-Way Analysis of Variance (ANOVA)
Kruskal-Wallis One-Way Analysis of Variance (ANOVA)
the medians of multiple samples of ranked data
What test is done for Ordinal data with Related samples
Friedman two-Way ANOVA
Friedman two-Way Anova
define
whether 2 categorical independent variables have a consistent relationship with a dependent variable that involves ordinal data
What tests measure relationships for correlating ordinal data
Spearman Correlation Coefficient and Kendall’s Tau
