Research Methods Level 5

Front 1

Cohen's effect - small and how do you find it ?

Back 1

0.20 when comparing 2 groups we take the differences of their means and divide it by the SD.

Front 2

Cohen's effect medium and how do you find it

Back 2

0.50 and you find the differences of both groups means and divide it by the SD

Front 3

Large Cohen's effect and what does it measure?

Back 3

The magnitude of an effect and large is 0.80

Front 4

Partial eta squared - what does it measure

Back 4

Effect size

Front 5

Partial eta squared - what does it measure

Back 5

Effect size

Front 6

Small partial eta squared is...

Back 6

.02

Front 7

Partial eta squared - what does it measure

Back 7

Effect size

Front 8

Small partial eta squared is...

Back 8

.02

Front 9

Medium partial eta squared is

Back 9

.13

Front 10

Partial eta squared - what does it measure

Back 10

Effect size

Front 11

Small partial eta squared is...

Back 11

.02

Front 12

Medium partial eta squared is

Back 12

.13

Front 13

Large partial eta squared is classed as

Back 13

.26

Front 14

Correlations can also measure

Back 14

Effect size

Front 15

Correlations can also measure

Back 15

Effect size

Front 16

If r=.10 is what size is the effect size?

Back 16

Small

Front 17

Correlations can also measure

Back 17

Effect size

Front 18

If r=.10 is what size is the effect size?

Back 18

Small

Front 19

If r=.30

What size is the effect size?

Back 19

Medium

Front 20

Correlations can also measure

Back 20

Effect size

Front 21

If r=.10 is what size is the effect size?

Back 21

Small

Front 22

If r=.30

What size is the effect size?

Back 22

Medium

Front 23

what is the large effect sizes in correlation?

Back 23

.50

Front 24

Front (Term)

Back 24

NONE as both means are similar on both sides (tasks). So there was an effect on the DV on both iv's.

Front 25

Back 25

No interaction as the patterning hasn't changed from task to task

Front 26

Front (Term)

Back 26

There is an interaction as there is a difference in both conditions (the dv increases on one iv and decreases on the other)

Front 27

What are the three assumptions/worries in factorial ANOVA?

Back 27

Normality

Levene's test

Mauchly's test

Front 28

Front (Term)

Back 28

Yes there is an interaction as means we're close together initially but big difference in end.

Front 29

Where does the f ratio come from?

Back 29

Mean square divided by error mean square

Front 30

Mixed factorial ANOVA, what table do we find it in?

Back 30

BOTH TABLES

BETWEEN TESTS AND WITHIN SUBJECTS EFFECTS.

Front 31

Mixed factorial ANOVA, what table do we find it in?

Back 31

BOTH TABLES

BETWEEN TESTS AND WITHIN SUBJECTS EFFECTS.

Front 32

What is an interaction?

Back 32

Happens in ANOVA. Where the effect on one of the iv's on the dv's differs across the levels on the other dv.

Front 33

Levenes test, what does it test?

Back 33

ANOVA

Tests whether we have homogeneity of variances

If levene's data is significant we need to transform data.

Front 34

Mauchly's test of sphericity when do we use it and when is it significant?

Back 34

We use it for repeated measures and mixed design as need three or more levels.

If significant we use the greenhouse grosser and hunh delft corrections.

Front 35

What is the main ANOVA table for repeated measures factorial ANOVA

Back 35

Tests of within subjects effects

Front 36

What is the main ANOVA table for repeated measures factorial ANOVA

Back 36

Tests of within subjects effects

Front 37

Total variance, how do we find it?

Back 37

Difference between the mean scores and actual scores squared and added together

Front 38

Simple regression - what is it?

Back 38

A straightforward extension of correlation.

Front 39

If you have a correlation of R=.40 what % is the variance in variable A that can be explained by variable B?

Back 39

16%

As 40 X 0.40 = 0.16

Front 40

If you have a correlation of R=.40 what % is the variance in variable A that can be explained by variable B?

Back 40

16%

As 40 X 0.40 = 0.16

Front 41

Bartlett's test of sphericity

What is it and what is the assumption?

Back 41

figures out correlations of variables we want 0 correlations and we want it to be dig so p<.05

Front 42

Working out variance with the mean:

Back 42

Find the difference between actual score and mean

Eg mean = 10 actual score = 8

(10-8) squared and then add them all up to find out how many units

Front 43

Working out variance with the mean:

Back 43

Find the difference between actual score and mean

Eg mean = 10 actual score = 8

(10-8) squared and then add them all up to find out how many units

Front 44

What are the cronbachs alpha thresholds?

Back 44

0.60 too low ( unless v short questionnaire)

0.70- acceptable

0.80-good

0.90- excellent (maybe too high)

Front 45

Working out variance with the mean:

Back 45

Find the difference between actual score and mean

Eg mean = 10 actual score = 8

(10-8) squared and then add them all up to find out how many units

Front 46

What are the cronbachs alpha thresholds?

Back 46

0.60 too low ( unless v short questionnaire)

0.70- acceptable

0.80-good

0.90- excellent (maybe too high)

Front 47

Total variance explained table - what do you use it for?

Back 47

We use it to work out how many factors to extract.

Front 48

Kolmogorov - smirnov test

What do you use it for? When? What table is it found in? What is the assumption?

Back 48

Tests the distribution of data used on a large sample size of 50 participants or more

P>.05

V sensitive so b careful

Front 49

Homoscedasticity what is it for and where do we look for it?

Back 49

Tells us whether the residuals differ at different predicted scores.

We look at the ZPRED &&ZRESID plots

Front 50

What does 'how each item loads onto each factor' mean? What table is it on?

Back 50

How much each item belongs to a factor, it is found on the rotated component matrix

Front 51

Good or bad homoscedasticity?

Back 51

Good as all residuals are random

Front 52

On the graph....r=.67

R(squared) =.0.45

Y=16.18 + 1.08(X)

What does this all mean?

Back 52

R= correlation

R(squared) = the line explains 45% of the variance in data.

Y is the line of best fit.

Front 53

What is the determinant value?

Back 53

Tests correlations between variables we want the determinant value to be LARGER than .00001

Front 54

What to say if the regression scores are different to actual scores

Back 54

"Some error/residual variance left unexplained by the regression model"

Front 55

How do we know if a regression model is significant?

Back 55

P<.05 change predictor variables until both are sig.

Front 56

When looking at any influential cases you need to look at

Back 56

OUTLIERS

be alarmed if outliers average are 3x bigger than the leverage values

Cook's distance is more than 1.

Mahalanorbis distance

if n=100 worry if MD >15

If n =500 worry if MD >25

Front 57

Depression score = 16.18+1.08(rumination score)

So if participant 1 has the rumination score =30 what is the depression score?

Back 57

16.18+1.08(30)

So 16.18 +32.4=48.58

Only multiply the second number

Front 58

How do you find the r squared value?

Back 58

The regression value divided by total this can be found in the SPSS ANOVA table

Front 59

Cronbach's alpha if deleted - where to find it and what does it show?

Back 59

Found in the item total statistics it shows whether cronbach's alpha will be IMPROVED or REDUCED if you remove that particular item

Front 60

What should the Eigen value be? + what table is it in?

Back 60

Eigenvalue> 1

Front 61

Kaiser-Meyer-olkin measure what is it and what are the thresholds?

Back 61

Calculates the sample size varies from 0-1

Below 0.50 - need to collect more data

0.50-0.70 -mediocre

0.70-0.80- good

0.80-0.90- great

0.90+ superb

(Rule of thumb is 300 pp's should be ok 10-15 pp's per variable)

Front 62

Cronbach's alpha - what is it?

Back 62

It tests the internal reliability

ARE THE ITEMS MEASURING THE SAME THING?

Front 63

F(*1,*2) =*3, p=.*4 n(squared)=.*5

What are they all?

Back 63

*1= df

*2 error

*3 = f ratio

*4 =sig

*5= partial eta squared

Front 64

Main table for independant factorial anova

Back 64

Tests of between subjects effects

Front 65

Mixed factorial ANOVA is

Back 65

All participants do all conditions for one iv

For the other iv participants only complete one condition

Front 66

Independant factorial ANOVA is

Back 66

Participants only complete one condition

Front 67

Independant factorial ANOVA is

Back 67

Participants only complete one condition

Front 68

Repeated measures factorial ANOVA is

Back 68

All participants complete all conditions

Front 69

3 different types of factorial ANOVA are

Back 69

Independant

Mixed

Repeated measures

Front 70

Independence of residuals- what do we look at?

Back 70

Look at the dustbin-Watson it should be approximately 2

(The errors should be independant and not correlated)

Front 71

Normality of residuals

Back 71

The errors should be normally distributed.

Only 5% of cases should have residuals more than 2 SD's away from the mean.

Front 72

Multiple regression

Back 72

Simple regression can sometimes have the predictor variables correlating with eachother so we enter them together into the analysis and that's multiple regression

Front 73

In ANCOVA a table called estimates is produced. What does it show ?

Back 73

It produces the adjusted means whilst controlling any covariates.