Statistical Tests: ANOVA (Test 1)

Analysis of Variance (ANOVA)

is a technique to assess how one or more categorical independent variables affect a continuous dependent variable

F statistic

The test statistic employed in ANOVA is called the ___________, which is actually a ratio of variance estimates.

F distribution

The test statistic will have an ____________ when the stated null hypothesis is true and the basics of hypothesis testing we covered earlier apply

factors

nominal variables that are independent variables

levels

different categories of a factor

One-way ANOVA

-involves a comparison of several (two or more) population means

-The main analysis problem in one-way ANOVA is to determine whether the population means are all equal or not

-is a statistical method for testing for differences in the means of 3 or more groups

overall test (omnibus test)

the test of the null hypothesis

-response variable Y

-single independent variable X

What is required in a one-way ANOVA?

F statistic

a ratio of variance estimates

-normally distributed

-homogeneity of variance

-independent

What are the assumptions for ANOVA?

multiple-comparison procedures (MCPs)

-Tukey's HSD (honestly significant difference) test

-The Bonferroni approach

-Scheffe's method

type I error

Multiple-comparison procedures (MCPs) control the possibility of making a _____________

Two-way ANOVA

used in situations where you want to compare means of a continuous outcome across 2 or more factors

interaction or effect modification

variation (heterogeneity) in treatment effects

effect modification

occurs when the association between an independent variable of interest (i.e., exposure or treatment groups assignment) and a health outcome (i.e., the dependent variable) is "modified" (i.e., is different) depending on the value of one or more other variables called effect modifiers

effect modifiers

these variables modify the relationship between exposure - or treatment - and outcome

subgroup analysis

is a direct examination of whether the treatment effect depends on patient subgroups

factorial design

A research design with two or more categorical independent variables (a.k.a., factors), each studied at two or more levels

repeated-measures ANOVA

like the paired t test, can be used to test the equality of means when all members of a random sample are measured under a number of different conditions

repeated

As the sample is exposed to each condition in turn, the measurement of the dependent variable is ______________

longitudinal research

A research design in which the same individuals are followed over time and their development is repeatedly assessed

parametric

T-test and ANOVA are called _____________ stats techniques.

parametric stats

Assumes that sample data comes from a population that can be adequately models by a probability distribution

nonparametric statistics

Statistical techniques that make little to no assumptions about the sampled population (distribution-free procedures) fall into the general set of techniques

assumption violations

these techniques can be used when the outcome variable is ordinal or when the outcome variable is interval or higher and there are _______________

unpaired t test

What is the analogous test to the nonparametric test called Mann-Whitney Test?

paired t test

What is the analogous test to the nonparametric test Wilcoxon Test?

one-way ANOVA

What is the analogous test to the nonparametric test Kruskal-Wallis test?

two-way ANOVA

What is the analogous test to the nonparametric test Friedman Test?

t-test

continuous dependent variable, comparing two means

one-way ANOVA

continuous dependent variable, comparing means from two or more groups on a single factor

two-way ANOVA

continuous outcome, independent variable and effect modifier (both represented by grouping or "categorical" variables)

repeated measures ANOVA

ANOVA for dependent observations