Chapter 9: Anova

Front 1

Very flexible and general technique, and the principles can be applied to a wide range of statistical tests.

Back 1

ANOVA

Front 2

What variable do we measure in ANOVA?

Back 2

outcome variable

Front 3

What is the other term for the outcome variable?

Back 3

dependent variable

Front 4

This outcome must be measured on a

Back 4

continuous scale

Front 5

It is called dependent because it depends on one or more __________ variables.

Back 5

predictor

Front 6

variables that can be Manipulated or simply measure

Back 6

Predictor Variables

Front 7

In ANOVA, predictor variables are mostly ______?

Back 7

categorical

Front 8

When predictor variables are categorical, they are also called ___________ or _____________

Back 8

FACTORS; INDEPENDENT VARIABLES

Front 9

what statistical test measures the differences between groups

Back 9

ANOVA

Front 10

In ANOVA, differences happen for two reasons: (a) because of the effect of ______ variables (b) because of _________

Back 10

predictor; other reasons.

Front 11

What do we call the difference in ANOVA

Back 11

Variance

Front 12

Variance due to ________ measures differences between Group

Back 12

treatment

Front 13

Variance due to __________ measure differences within Group

Back 13

Error

Front 14

In ANOVA, the variance is conceptualized as sums of SDs from the _________.

Back 14

mean

Front 15

Sum of SDs can be denoted as ____?

Back 15

SS

Front 16

the variance that represents the difference between the groups

Back 16

SS(between)

Front 17

What other term can we use to denote SS(between)?

Back 17

SS(treatment)

Front 18

The ________ variance is the variance that we are actually interested in

Back 18

between-groups

Front 19

also called within-groups sum of squares.

Back 19

Error Sum of Squares

Front 20

people, who have had the same treatment, have different scores and it's because of error. So this is variance is called either _____ , or ________

Back 20

SS(within) ; SS(error)

Front 21

sum of squared differences between the mean and each score

Back 21

SS(total)

Front 22

Effect Size goes under two different names: these are

_______ or _______.

Back 22

R-Squared or eta-Squared

Front 23

df between = ___?

Back 23

(no. of groups) g- 1

Front 24

What is the statistic (symbol) for ANOVA?

Back 24

F or F-ratio

Front 25

How many sets of df we need to find the probability value (p-value) associated with F

Back 25

2 - df(within) and df(between)

Front 26

ANOVA and ____ are exactly the same test

Back 26

t-test

Front 27

is we squared the value of t, what will we get?

Back 27

F-ratio value

Front 28

In ANOVA, what do we assume as normally distributed?

Back 28

the data within each group

Front 29

As with the t-test, we assume that the SD within each group is approximately equal. What states this?

Back 29

Homogeneity of Variance

Front 30

also called as simple-randomized groups design, independent groups design, or the single factor experiment, independent groups design.

Back 30

One way ANOVA

Front 31

one in which the effects of two or more factors or IVs are assessed in one experiment

Back 31

Factorial experiment

Front 32

With __________, conditions or treatments used are combinations of the levels of factors

Back 32

Factorial experiment

Front 33

More complicated ANOVA

Back 33

Two way ANOVA

Front 34

If factor A had three levels, the experiment would be called a 3 x 2, _______.

Back 34

two way ANOVA

Front 35

In a 2 x 4 x 3 design, there would be 3 factors having 2, 4, and 3 levels, respectively – this is called a _______

Back 35

3 way ANOVA

Front 36

here, the levels of each factor were systematically chosen by the experimenter rather than being randomly chosen

Back 36

Fixed effects design

Front 37

what do we call it when we want to determine whether factor A has a significant effect, disregarding the effect of factor B.

Back 37

the main effect of factor A

Front 38

what do we call it when we want to determine whether there is an interaction between factors A and B.

Back 38

interaction effect of factors A and B.

Front 39

In analyzing data from a two-way ANOVA, we determine four variance estimates: (give atleast 2)

Back 39

1. MS within cells

2. MS rows

3. MS columns

4. MS interaction

Front 40

The other estimates are sensitive to the effects of the ________

Back 40

Independent Variables

Front 41

The estimate MS rows is called the _________

Back 41

row variance estimate

Front 42

The estimate MS columns is called the __________.

Back 42

column variance estimate

Front 43

The estimate MS interaction is the ________.

Back 43

interaction variance estimate

Front 44

If variable A has no effect, MS rows is an independent estimate of the _______.

Back 44

σ-squared

Front 45

In a 2-way ANOVA, we partition the SS (total), into four components:

Back 45

1.SS (within-cells) ,

2. SS (row)

3. SS (column)

4. SS (interaction)

Front 46

what tests do we use to answer the question of where the differences come from.

Back 46

Post Hoc tests

Front 47

“Post hoc” is Latin, and means.

Back 47

“after this”

Front 48

What problem might occur if we just use t-test when comparing groups?

Back 48

alpha inflation

Front 49

What type of error does alpha inflation is

Back 49

Type 1

Front 50

A Type I error is where we reject a null hypothesis that is _______

Back 50

true

Front 51

What assumption do we make for Bonferroni Correction

Back 51

homogeneity of variance

Front 52

Calculate df in Bonferroni correction

Back 52

df= N-2

Front 53

The problem with Bonferroni is that it's not that precise. And the p-values required for statistical significance rapidly become _____.

Back 53

very small

Front 54

Why is homogeneity of variance is important?

Back 54

Homogeneity of variance means that the variances (standard deviations) of each of the groups that we are investigating are approximately equal. This is an assumption made by ANOVA and t-tests, and it is important for two reasons.

First, we use the variance to estimate the standard error (and hence the confidence intervals and p-values). If we can assume homogeneity of variance, we can pool the variances from each group. If we estimate the variance with a larger number of people, the standard error will be smaller (because se = σ/√Ν ). And a smaller standard error means reduced confidence intervals, and lower p-values.

Second, and perhaps more importantly, if we can assume homogeneity of variance, the calculations are much, much easier.

Front 55

Post hoc tests are based on ____?

Back 55

t-test