intro to biostats Flashcards
required assumptions of interval data
normally distributed
equal variances
randomly derived and independent
what test can be used to determine equal variance in interval data
levene’s test
handling interval data not normally distributed
use non parametric tests
or transform data to standardized value (z score or log)
if stem has prediction then what comparisson
regression
occurences over time
survival comparison
are there differneces btwn to categories
group comparison
will say want correlation
correlation
nominal correlation test
contingency coefficient
ordinal correlation tewst
spearman correlation
interval correlation test
pearson correlation
special thing about pearson correlation
just assesses for linear correlation, may still be non linear correlations if pearson is non significant
what is surivaval tests usually represented by
kaplan meier curve
nominal survival
ordinal survival
interval survival
how do you survive?
Long Cox Kill
log rank
cox proportinal hazards test
kaplan meier test
nominal regression
ordinal regression
interval regression
Logs multiply linearlly
logistic regression
multinomaial logistic regression
linear regression
nominal, 3 groups or more and related
cochran
johnny cochran has worked for 3 or more seperate normal clients
nominal 3 groups or more and independent
chi sqauare and fisher exact
cameron fiser nominally has 3 independent crews working for him
nominal, 2 groups, related
mcnemar
nominal, 2 groups, independent
(pearsons) chi square and fisher’s exact
cam fiser with 2 independent crews
follow up confirmation test for nominal
bonferroni test of inequality
normal bon fire
ordinal, 3 groups or more, related
friedman
ordinarily 3 related groups are free
ordinal, 3 groups or more, independent
kruskal-wallis
krush the wall
ordinal, 2 groups, independent
whitney is ordinary, independent of mann
mann whitney
ordinal, 2 groups, related
wilcoxon signed rank
mrs wilcox signed rank
she is ordinary and want to be related
follow up test ordinal for 3 groups or more
student newman keul
dunnett
dunn
nominal, 2 groups or more with expected cell count under 5
fisher’s exact test
follow up test interval 3 or more groups
student newman keul dunett dunn tukey or scheffe bonferroni
a correlation test showing relationship or agreement between evaluator (consistency of decisions, determinations)
kappa statistic
kappa interpretation
+1 = observers perfectly classify everyone exact same way
0 = no relationship between observers
-1 = boservers classify everyone exact opposite
null hypothesis
no true diff btwn groups of comparison
type 1 error
rejecting null hypothesis when shouldnt
type 2 error
not rejecting null hypothesis when you should
the ability of a study design to detect a true difference if one truly exists btwn group comparisons, and therefore the level of accuaracy in correctly accepting/rejecting the null hypotheis
power (1-B error)
what increases the power
increasing the sample size
interpretation of p value
4
probability of making type 1 error
probability of claiming diff btwn groups when one doesn’t really exist
probability of obtaining group diff as great or greater if the groups were actually same/equal
probability of obtaining a test statisti as high/higher if groups were actually the same/equal
interval 2 groups, independent
student t
interval, 2 groups, related
paired t test
