RDS Flashcards
Discrete comparison
Chi squared
Discrete correlation
chi squared
non-normal continuous correlation
Spearman
normal continuous correlation
pearson
continuous normal one variable more than two groups comparison
ANOVA
continuous normal comparison with two groups
t test, paired or unpaired
Not normal continuous comparison with more than two groups
kruskal wallis
not normal continuous comparison with two paired groups
wilcoxon
not normal continuous comparison with two independent groups
mann whitney
types of data
categorical, ordinal is ordered nominal isn’t
quantitative discrete or continuous
Normal distribution terms
kurtosis = tail heaviness skewness = symmetry gaussian = perfect
assessing normality
more than fifty = kolmogorov-smironov
less shapiro wilks
hierarchy of evidence
meta analysis and systematic review randomised control welldesigned control case control descriptive shit casereport
measures of averages
normal = mean non-normal = median and mode
rrr
relative risk reduction, probability decrease after lifestyle change
RR
Relative risk, used to estimate strength of association between treatments/risk factors and outcomes. Tests significance between groups, ie. disease vs control or drug vs placebo.
Contingency table formulae
RR= (A/A+B)/(C/C+D), over one means positive association, less than one means negative association
OR= odds of something occurring with exposure, vs odds without exposure
odds of exposure with outcome / odds of exposure in people without outcome