Correlation Analysis Flashcards
(22 cards)
when to use a t-test
2 nominal or categorical treatment variables
interval/ratio dependent variable
when to use an ANOVA
3 nominal treatment variables
interval/ratio outcome variable
when to use a two-way ANOVA
2 separate classes of independent variables
ie - treatment and gender differences
interval/ratio dependent outcome variable
when to use a repeated measures ANOVA
3 or more treatment independent variables assessed at intervals
interval/ratio dependent variable
an R number indicates
closer to 1 = direct linear relationship
closer to 0 = no relationship
closer to -1 = indirect linear relationship
explain how to calculate the % factor than an independent variable plays when given an R number
squaring the r value explains how much of an effect it plays on the dependent variable
r = 0.3 –> 27% factor in the outcome
how to calculate correlation
pearson correlation coefficient
explain level of correlation
low —- r <or= 0.4
mod — r = 0.41-0.7
good — r > 0.71
what test is performed for non-parametric statistics
Chi-Squared Test
define non-parametric statistics
nominal independent variable
ordinal dependent variable
explain what expected values are in chi-squared test
second set of values for comparison
– made on the idea that the null hypothesis is true
explain p values in chi-squared test
same concept
> 0.05 accept the null hypothesis
< 0.05 accept alternate hypothesis
explain degrees of freedom for chi-squared
(r-1) x (c-1)
r = rows
c = columns
how to calculate expected values
column total x (row / grand total)
test of normalcy include
shapiro-wilk test
kolmogorov-smirnov test
explain mann whitney U test or wilcoxon rank sum test
equivalent of t-test for nonparametric data
dependent variable = interval/ratio
– data is skewed and does not follow normal distribution
what can the mann whitney U test or wilcoxon rank sum test be used for
rank data
what is the comparative of an ANOVA for nonparametric data
kruskal wallis one way ANOVA
explain indication of kruskal wallis ANOVA
dependent variable = ordinal or non-normal skewed data
how to further test kruskal wallis ANOVA
P value will tell you a difference between variables, not necessarily which one was better
use a bonferroni correction to adjust p-value for significance
what is the non-parametric equivalent of pearson correlation
spearman’s rank-order correlation
explain use of spearman’s rank order correlation
data is ranked and ordered
denoted by an r value with same numbering
