Introduction to BIostatistics in Epi Flashcards
3 key attributes of data variables
Magnitude(biger is more, lower is less) Consistency of scale(or fixed intervals) -equal, measurable spacing btw units rational zero Each can be answered with yes or no repsonse
Nominal
Dichotomouse/binary; no order or rank; non ranked categories…. NO magnitude, NO consistency of scale, NO rationonal zero….. examples…. gender, occupationsoal class, party affiliation (ALL dichotomous and non-ranked categories)
Ordinal
Ranked Categories; non-equal-distance… YES magnitude, NO consitency of scale, NO rational zero… Rank of canidates in order of preference from worse to best…( all ranked categories)
Interval/ Ratio
YES magnitude, YES consitency of scale, NO or YES rational zero (no-interval; yes-ratio) ex… number of living sibilings and age… (all numerical scales with ture units)
What type of data is being collected or evaluated?
…
What type of comparison/assessment is desired?
Correlation!!!! Correlations test.
Correlation
Provides a quantitative measure of the strength and direction of a relationship btw variables…. creates a 45 degree angle…. X goes down, Y goes up….
Correlation….. Is the correlation linear? if not significanlty significan then there is NOT linear correlation
ordinal= spearman correlation nominal= contingency coefficent interval= pearson correlation
Contingency coefficent
if more than .05 it is not significantly significant
Time-to event… event-occurrence
survival test… can all be represented by kaplan meier curve
Suvival Test
ordinal=cox proportional hazards
nominal= log-rank
interval= kaplan-meier product-limit estimate
survial tests….
compares the proprtion of, or time to event occurence btw groups….
Outcome prediction/association
Regression
Regression (THE WORD PREDICT!!!0
Provide a measure of the relationship btw variables by allowing the prediction about the dependent, or outcome, variable (DV) knowing the value/rank of others independent variables….. able to calculate measerue of associateions
Regression tests
Nominal=logistic regression
ordnial= multinominal logistic regression
interval= linear regression
Frequencies/ counts/ proportions
HOW MAN GROUPS!
is the datat independent or related
Data from the same( paired) ore different groups(independent)?
Nominal 2 groups independent data
(Pearsons) Chi-square test
nominal greater than 3 independent data
Chi-square test of independence… no cell with expected count of <5…. to predeict 3 more comparisons, one must perform subsequent analysis(post-hoc testing) to determine which groups are different…. Multiple Chi tests NEVER acceptable…. Bonferroni test of Ineqaulity, ver conservative
nominal >2 groups with expected cell count of <5
fisher’s exact test
Nominal 2 groups of paired/related data
McNemar test…. key words… pre vs post… before vs after, basline vs end
nominal >3 groups of paired related data
cochran ….. must do bonferroni test of inequality
Ordinal 2 groups of independent data
Mann-whitney test
Ordinal data >3 groups of independent data
Kruskal wallis test…. compares the mdeian values btw groups…. mujst perform post hoc
If related data
before and after… beginning and end… baseline vs end
Ordnial data 2 groups of Paired data
Wilcoxon SIgned Rank test
Ordinal >3 groups of paired data
Friedman test… do post hoc after
Post hoc tests for 3 or more group comparisons
Student-Newman Keul test - comapres all pairwise comparisons possilbe
Dunnett Test- compares agains single.. equal in size
Dunn Test- Compares all possilbe comparisons
Interval 2 grousp of independent data
student t test… compares mean values btw groups
interval data >3 groups of independent data
ANaysis of Varainace (ANOVA) (1DV)- comapres menas of all groups
Multiple Analysis of Variance (Manova) (>2 DVs) must do post hoc
Interval data >3 groups of independent data w cofounders
ANCOVA
MANCOVA
Interval 2 groups of data paired
Paired t test
interval data >3 groups of paired data
repeated ANOVA Repeated MANOVA(>2)
INterval data>3 paried with cofounders
repeated ANCOVA
Reated MANCOVA