BIO exam

Single trial distribution

bernoulli

Multiple trial distribution

binomial

Time period distribution

poisson

Bayer’s expanded

Pr(A|B) = (Pr(B|A)*Pr(A))/((Pr(A)*Pr(B|A))+(Pr(notA)*Pr(B|notA)))

Normality null hypothesis

Data are normally distributed in the population

Normality alt hypothesis

Data are not normally disributed in the population

Wilcoxon rank-sum test

mann-whitney u test, two-sample t-test

Wilcoxon signed-rank test

paired samples t-test

Wilcoxon rank-sum, Asymp. Sig (2-tailed)

when you have at least 20 samples in each group

Wilcoxon rank-sum, Exact. Sig. [2* (1-tailed Sig.)]

when you have no ties

Wilcoxon signed-rank, Asymp. Sig (2-tailed)

number of non-zero differences is greater than or equal to 20

Wilcoxon signed-rank, Exact. Sig. [2*(1-tailed Sig.)]

number of non-zero differences < 20 and there are no tied ranks

Wilcoxon rank-sum, Exact Sig. (2-tailed)

you have small sample sizes in each condition and ties are present

Wilcoxon rank-sum, Monte Carlo Sig. (2-tailed)

you have small sample sizes in each condition and ties are present

Wilcoxon signed-rank, Exact Sig. (2-tailed), Exact Sig. (1-tailed)

number of non-zero differences < 20 and there are tied ranks present

Wilcoxon signed-rank, Monte Carlo Sig. (2-tailed) or Monte Carlo Sig. (1-tailed)

number of non-zero differences < 20 and there are tied ranks present

R²

coefficient of determination, proportion of variability in the dependent variable that can be explained by a linear relationship with the predictor variable

R

pearson correlation coefficient, measure of the strength of the relationship between two variables

Steps of hypothesis testing

state hypotheses, decide test, check assumptions, set alpha and determine critical value, evaluate sample evidence with test statistic, make comparison and conclude

Regression equation

y = a + Bx

Prospective cohort study

groups chosen based on risk factor prior to condition

Retrospective case-control study

group with condition chosen, group without condition and similar characteristics chosen, risk factors assessed

Retrospective cohort study

group chosen with risk factor, group chosen without, medical records searched for who has condition

one sample z-test, t-test null

= x

one sample z-test, t-test alt

does not = x

Homogeneity of variance null

population variance is equal

Homogeneity of variance alt

population variance is not equal

Two-sample t-test null

x1-x2 = 0

Two-sample t-test alt

x1-x2 does not = 0

Paired samples t-test null

x delta = 0

Paired samples t-test alt

x delta does not = 0

Wilcoxon rank-sum null

the two samples have been drawn from two identical population distributions

Wilcoxon rank-sum alt

The two samples have not been drawn from identical population distributions

Wilcoxon signed-rank null

there is no difference between scores in the population

Wilcoxon signed-rank alt

there is a difference between scores in the population

Chi-square null

there is no association in the population between the categorical variables

Chi-square alt

there is an association in the population between the categorical variables

ANOVA regression null

the data does not fit the linear regression model

ANOVA regression alt

the data fits the linear regression model

constant null

a = 0

constant alt

a does not = 0

slope null

B = 0

slope alt

B does not = 0

R null

r = 0

R alt

R does not = 0

risk factor null

RR = 1

risk factor alt

RR does not = 1

odds ratio null

OR = 1

odds ratio alt

OR does not = 1

Result reporting

test used, what was tested, mention values, significance level, mean and SD, results of significance, test(df) =, p=, 95% CI [ ]

Chi square null rejection

X2> X2crit

simple random sampling

SRS, every person in a population has an equal chance of being selected

stratified sampling

divides populations, picks random people in each group proportional to the strata’s representation

cluster sampling

population divided into clusters, defined number of clusters selected, SRS

convenience sampling

makes no effort to increase likelihood of sample being representative, high probability of bias

Z test assumptions

representative, independence, normality, known population SD

one sample t test assumptions

representative, independence, normality

paired t test assumptions

normality, representative, independence

two sample t test assumptions

normality, representative, independence, equal variances

Chi-square assumptions

representative, independence, counted data, expected frequency >5

regression assumptions

independence, normality, linearity, equal variances

relative risk assumptions

representative, independence, counted, prospective

odds ratio assumptions

representative, independence, counted, retrospective

risk ratio

likelihood of condition occurring in the future

odds ratio

likelihood that people with the condition were exposed to a risk factor