08 Oct Repeated- Measures Designs

Repeated- Measures Designs

 

 

 

 

 

When the assumption of sphericity is violated what action is needed:

Correct the model degrees of freedom.

Correct the error degrees of freedom.

Do both a and b.

Correct the F-ratio.

 

 

Which of the following statements about the assumption of sphericity is not true?

It is tested using Mauchley’s test in SPSS.

Does not apply when a variable has only two levels.

Does not apply when multivariate tests are used.

It is the assumption that the variances for levels of a repeated measures variable are equal.

 

 

A nutritionist conducted an experiment on memory for dreams. She wanted to test whether it really was true that eating cheese before going to bed made you have bad dreams. Over three nights, the nutritionist fed people different foods before bed. On one night they had nothing to eat, a second night they had a big plate of cheese, and the third night they had another dairy product, Milk, before bed. All people were given all foods at some point over the three nights. The nutritionist measured heart rate during dreams as an index of distress. How should these data be analysed?

One-way independent ANOVA.

One-way repeated measures ANOVA.

Three-way repeated measures ANOVA.

Three-way independent ANOVA.

 

 

 

 

What is NOT an advantage of repeated measures designs in comparison to independent measures designs?

 

 

Each participant acts as their own control

Researchers can study cross-cultural effects more easily

Researchers can study trends more easily

They require fewer participants overall

Sphericity is

 

 

An assumption that means the data distribution must be round.

The critical value area of the graph is round

A way of rounding up the decimal points

An assumption that means the data in each level should uncorrelated

 

 

If there is sphericity in a repeated measures design the outcome could be that

 

 

The p value will be too high

The p-value will be too low

A p-value cannot be computed

The p-value will not be related to the model

 

 

The Greenhouse-Geisser correction refers to:

 

 

Temperature control

A way of dealing with sphericity

Raising the sample size

Lowering humidity

 

 

An experiment was carried out in which participants learned words in several conditions: no learning strategy, a verbal learning strategy, visual one and a verbal-visual one.

 

 

 

What considerations would the researchers NOT need to take into account?

 

 

Fatigue

Learning effects

Asymmetric transfer

Parametric assumptions

 

 

What safeguard could the researchers put in place to overcome the difficulties with repeated measures?

 

 

Counterbalancing

Something to keep the participants awake

Measure an appropriate covariate

Before and after measurements

Consider the table below, the results from the above experiment.                                      What do they indicate?

 

 

There is too much sphericity in the model

There are too many errors in the model

The null hypothesis can be rejected

The null hypothesis must be rejected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Chapter 12 – Mixed Design ANOVA

 

 

 

 

 

Field & Lawson (2003) reported the effects of giving 7-9 year old children positive, negative or no information about novel animals (Australian marsupials). This variable was called ‘Infotype’. Gender of the child was also examined. The outcome was the  time taken for the children to put their hand in a box in which they believed either the positive, negative, or no information animal was housed (Positive values = longer than average approach times, negative values = shorter than average approach times). Based on the output below, what could you conclude? [see Field, A. P., & Lawson, J. (2003). Fear information and the development of fears during childhood: effects on implicit fear responses and behavioural avoidance. Behaviour Research and Therapy, 41, 1277–1293.]

 

 

 

 

 

 

Male Female

 

0.8

 

 

 

 

 

 

 

Latency to Approach (Z-Score)

 

0.6

 

 

 

 

 

0.4

 

 

 

0.2

 

 

 

0.0

 

 

 

-0.2

 

 

 

-0.4

 

 

 

 

 

-0.6

 

 

 

 

 

Negative

 

 

 

 

 

Positive

 

 

 

 

 

None

 

 

 

 

 

Type of Information

 

 

 

Tests of Within-Subjects Effects

 

Measure: MEASURE_1

 

 

 

Source

 

Type III Sum of Squares

 

 

 

df

 

 

 

Mean Square

 

 

 

F

 

 

 

Sig.

 

INFOTYPE                               Sphericity Assumed

 

9.177

 

2

 

4.588

 

7.283

 

.001

 

Greenhouse-Geisser

 

9.177

 

1.940

 

4.730

 

7.283

 

.001

 

Huynh-Feldt

 

9.177

 

2.000

 

4.588

 

7.283

 

.001

 

Lower-bound

 

9.177

 

1.000

 

9.177

 

7.283

 

.010

 

INFOTYPE  GENDER Sphericity Assumed

 

.599

 

2

 

.299

 

.475

 

.623

 

Greenhouse-Geisser

 

.599

 

1.940

 

.309

 

.475

 

.618

 

Huynh-Feldt

 

.599

 

2.000

 

.299

 

.475

 

.623

 

Lower-bound

 

.599

 

1.000

 

.599

 

.475

 

.495

 

Error(INFOTYPE)                  Sphericity Assumed

 

51.664

 

82

 

.630

 

 

 

 

 

Greenhouse-Geisser

 

51.664

 

79.544

 

.650

 

Huynh-Feldt

 

51.664

 

82.000

 

.630

 

Lower-bound

 

51.664

 

41.000

 

1.260

 

 

 

 

 

Tests of Between-Subjects Effects

 

 

 

 

 

Source

 

Type III Sum of Squares

 

 

 

df

 

 

 

Mean Square

 

 

 

F

 

 

 

Sig.

 

Intercept

 

2.034E-02

 

1

 

2.034E-02

 

.049

 

.826

 

GENDER

 

1.822E-03

 

1

 

1.822E-03

 

.004

 

.948

 

Error

 

17.109

 

41

 

.417

 

 

 

 

 

 

 

Measure: MEASURE_1 Transformed Variable: Average

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Approach times were significantly different for the boxes containing the different animals, but the pattern of results was affected by gender.

Approach times were significantly different for the boxes containing the different animals, but the pattern of results was unaffected by gender.

Approach times were not significantly different for the boxes containing the different animals, but the pattern of results was affected by gender.

Approach times were not significantly different for the boxes containing the different animals, but the pattern of results was unaffected by gender.

 

 

Based on the information in the previous question, what analysis has been done?

A two-way Mixed ANOVA.

A Three-way Mixed ANOVA.

A two-way repeated measured ANOVA.

A Two-way Independent ANOVA.

 

 

For the same data as in the previous question, contrasts were performed. Based on the SPSS output given, which of the following statements is true? (Levels of Infotype were entered in the following order: negative information, positive information, no information)

 

 

Tests of Within-Subjects Contrasts

 

Measure: MEASURE_1

 

 

 

Source                                         INFOTYPE

 

Type III Sum of Squares

 

 

 

df

 

 

 

Mean Square

 

 

 

F

 

 

 

Sig.

 

INFOTYPE                               Level 1 vs. Level 3

 

Level 2 vs. Level 3

 

11.090

 

.447

 

1

 

1

 

11.090

 

.447

 

8.762

 

.420

 

.005

 

.521

 

INFOTYPE  GENDER  Level 1 vs. Level 3

 

Level 2 vs. Level 3

 

1.177

 

.174

 

1

 

1

 

1.177

 

.174

 

.930

 

.163

 

.341

 

.688

 

Error(INFOTYPE)                  Level 1 vs. Level 3

 

Level 2 vs. Level 3

 

51.896

 

43.689

 

41

 

41

 

1.266

 

1.066

 

 

 

 

 

 

 

Approach times for the box containing the negative animal were not significantly different to those for the box containing the positive information animal.

Approach times for the box containing the positive animal were significantly shorter to the box containing the control (no information) animal.

The profile of results were different for boys and girls.

Approach times for the box containing the negative animal were significantly longer than for the box containing the control (no information) animal.

 

 

What would be the appropriate SPSS commands for a mixed design?

 

 

 

 

Analyze – GLM – multivariate

Analyze – Mixed model – linear

Analyze-GLM-repeated measures-define and add a between subjects factor

Analyze-GLM-univariate and define fixed and random factors

 

 

Which of the following is a mixed design?

 

 

An investigation of the effect of sex of participant on age of attaining a degree

An investigation of the effect of the sex of the participant on driving simulator errors before and after drinking alcohol

An investigation of the effect sex of participant on driving simulator errors with and without training

An investigation of the effect of sex of participant on choice of degree topic.

 

 

A mixed factorial design

 

 

Is one in which both men and women take part

Has at least one between subjects variable and one within subjects variable.

Utilises both categorical and continuous

Needs a non-parametric tests

 

 

 

 

Chapter 13 – Non-parametric Tests

 

 

 

 

 

A researcher was interested in stress levels of lecturers during lecturers. She took the same group of 8 lecturers and measured their anxiety (out of 15) during a normal lecture and again in a lecture in which she had paid students to be disruptive and misbehave. The data were not normally-distributed. Which test should she use to compare her experimental conditions?

Paired t-test.

Mann-Whitney test.

Wilcoxon signed ranks test.

Wilcoxon rank sum test.

 

 

A psychologist was interested in whether there was a gender difference in the use of email. She hypothesised that because women are generally better communicators than men, they would spend longer using email than their male counterparts. To test this hypothesis, the researcher sat by the email computers in her research methods laboratory and when someone started using email, she noted whether they were male or female and then timed how long they spent using email (in minutes). How should she analyze the differences in males and females (use the output below to help you decide)?

 

 

 

 

Tests of Normality

 

 

 

 

 

Kolmogorov-Smirnova

 

Shapiro-Wilk

 

Statistic

 

df

 

Sig.

 

Statistic

 

df

 

Sig.

 

Average Time spent using

 

Email (minutes)

 

.261

 

16

 

.005

 

.735

 

16

 

.010

 

. This is an upper bound of the true significance.

 

Lilliefors Significance Correction

 

 

Paired t-test.

Mann-Whitney test.

Wilcoxon signed ranks test.

Independent t-test.

 

 

Another term for non-parametric tests is

 

 

Non-normal tests

Data-free tests

Non-continuous tests

Distribution-free tests

 

 

The non-parametric equivalent of the paired t-test is the

 

 

Mann-Whitney U test

Wilcoxon sign test

Friedman test

Kruskall-Wallis test

 

 

The non-parametric equivalent of the two-sample independent t-test is the

 

 

Mann-Whitney U test

Wilcoxon sign test

Friedman test

Kruskall-Wallis test

 

 

 

 

 

 

 

 

 

 

Chapter 14 – Multivariate Analysis of Variance

 

 

 

 

 

A psychologist was interested in gauging the success of a mood manipulation during one of her experiments. She had three groups of participants who underwent different types of mood induction: disgust mood induction, negative mood induction and positive mood induction. After the mood induction, participants were asked to endorse nine statements relating to their mood (on a 5 point Likert scale from 1—disagree to 5—agree): (1) When you’re smiling the whole world smiles with you, (2) I love the pretty flowers, (3) I could never touch a dead body, (4) I would never eat cat food,

If someone served me monkey brain soup I would vomit, (6) I feel fed up, (7) Bodily fluids are nasty, (8) I could not drink from a glass that I’d used to catch a spider, (9) I am a worthless piece of scum. What analysis should be done to see if  the mood inductions had an effect on responses to these 9 items.

Factor analysis.

MANOVA.

Repeated Measures ANOVA.

Mixed ANOVA.

 

 

 

 

A psychologist was interested in gauging the success of a mood manipulation during one of her experiments. She had three groups of participants who underwent different types of mood induction: disgust mood induction, negative mood induction and positive mood induction. After the mood induction, participants were asked to endorse nine statements relating to their mood (on a 5 point Likert scale from 1—disagree to 5—agree): (1) When you’re smiling the whole world smiles with you, (2) I love the pretty flowers, (3) I could never touch a dead body, (4) I would never eat cat food,

If someone served me monkey brain soup I would vomit, (6) I feel fed up, (7) Bodily fluids are nasty, (8) I could not drink from a glass that I’d used to catch a spider, (9) I am a worthless piece of scum. What analysis should be done to see if  the mood inductions had an effect on responses to these 9 items. Part of the SPSS output is below. Which of the following statements best summarizes the output.

 

 

 

 

Multivariate Testsc

 

Effect

 

Value

 

F

 

Hypothesis df

 

Error df

 

Sig.

 

Intercept Pillai’s Trace

 

.983

 

366.116a

 

9.000

 

58.000

 

.000

 

Wilks’ Lambda

 

.017

 

366.116a

 

9.000

 

58.000

 

.000

 

Hotelling’s Trace

 

56.811

 

366.116a

 

9.000

 

58.000

 

.000

 

Roy’s Largest Root

 

56.811

 

366.116a

 

9.000

 

58.000

 

.000

 

GROUP         Pillai’s Trace

 

.416

 

1.723

 

18.000

 

118.000

 

.044

 

Wilks’ Lambda

 

.615

 

1.772a

 

18.000

 

116.000

 

.037

 

Hotelling’s Trace

 

.575

 

1.820

 

18.000

 

114.000

 

.031

 

Roy’s Largest Root

 

.465

 

3.048b

 

9.000

 

59.000

 

.005

 

Exact statistic

The statistic is an upper bound on F that yields a lower bound on the significance level.

Design: Intercept+GROUP

 

 

 

 

The type of mood induction had a significant effect on responses to all of the 9 items.

The type of mood induction had a significant effect on responses to at least one of the 9 items.

The type of mood induction that a person had could be determined from a linear combination of responses to the 9 items.

The type of mood induction had a significant effect on responses to more than half at least of the 9 items.

 

 

The next part of the output is shown below. Which statement best sums up this part of the output?

 

 

Levene’s Test of Equality of Error Variancesa

 

 

 

 

 

F

 

df1

 

df2

 

Sig.

 

When you’re smiling the whole world smiles with you

 

2.235

 

2

 

66

 

.147

 

I love the pretty flowers

 

1.004

 

2

 

66

 

.378

 

I could never touch a dead body

 

.291

 

2

 

66

 

.771

 

I would never eat catfood

 

3.239

 

2

 

66

 

.073

 

If someone served me monkey brain soup I would vomit

 

1.591

 

2

 

66

 

.236

 

I feel fed up

 

.216

 

2

 

66

 

.847

 

Bodily fluids are nasty

 

11.176

 

2

 

66

 

.000

 

I could not drink from a glass that I’d used to catch a spider

 

1.438

 

2

 

66

 

.266

 

I am a worthless piece of scum

 

3.978

 

2

 

66

 

.044

 

Tests the null hypothesis that the error variance of the dependent variable is equal across groups.

 

a. Design: Intercept+GROUP

 

 

 

There were significant differences between the mood induction conditions on two items: ‘I am a worthless piece of scum’, and ‘Bodily fluids are nasty’.

There were significant differences between the mood induction conditions on three items: ‘I am a worthless piece of scum’, ‘I would never eat catfood’ and ‘Bodily fluids are nasty’.

I need to transform some of the items.

I need to transform all of the items.

The next part of the output is shown below. Which statement best sums up this part of the output?

 

 

There were significant differences between the mood induction conditions on all items.

There were significant differences between the mood induction conditions on two items: ‘I would never eat catfood’, and ‘Bodily fluids are nasty’.

There were significant differences between the mood induction conditions on four items: ‘I would never eat catfood’, ‘I could never touch a dead body’, ‘I feel fed up’ and ‘Bodily fluids are nasty’.

The mood induction had no effect on responses to the 9 items.

 

 

Multivariate Analysis of Variance (MANOVA) is

 

 

An extension of analysis of variance with more than one interaction

 

 

An extension of analysis of variance used to accommodate more than one dependent variable.

 

 

An extension of analysis of variance with more than two independent variables

 

 

An extension of multiple regression that allows the variance to calculated from means

 

 

Chapter 15 – Exploratory Factor Analysis

 

 

 

 

 

Based on this scree plot, how many factors should be extracted?

 

 

8

 

 

 

6

 

 

 

4

 

 

 

2

 

 

 

0

 

1      2      3      4      5      6      7      8      9

 

Factor

 

 

 

2.

b. 3.

 

4.

5.

 

 

Varimax rotation should be used when,

Factors are expected to correlate.

Factors are non-orthogonal.

Factors are independent.

Kaiser’s criterion is met.

 

 

Oblique rotation should be used when,

Factors are expected to correlate.

Factors are orthogonal.

Factors are independent.

Kaiser’s criterion is met.

 

 

A scree plot in factor analysis is a plot of:

Each factor against its eigenvalue.

The factor loadings of each variable onto each factor.

The correlations between variables.

The regression coefficient of each variable with each factor.

Kaiser’s criterion for retaining factors is:

Retain any factor with an eigenvalue greater than 0.7.

Retain any factor with an eigenvalue greater than 1.

Retain factors before the point of inflexion on a scree plot.

Retain factors with communalities greater than 0.7.

 

 

 

 

Chapter 16 – Categorical Data

 

 

 

 

 

933 people were asked which type of programme they prefer to watch on television. Results are below. What is the expected frequency for men who liked to watch sport?

 

 

 

 

 

 

News

 

Documentaries

 

Soaps

 

Sport

 

Total

 

Women

 

108

 

123

 

187

 

62

 

480

 

Men

 

130

 

123

 

68

 

132

 

453

 

Total

 

238

 

246

 

255

 

194

 

933

 

 

 

a. 132

 

b. 94.19

 

c. 64.09

 

d. 27.45

 

 

 

Based on the data above, what are the odds of being a man if you watch sport? a. 0.47

b. 0.14

 

c. 0.41

 

d. 2.13

 

 

 

For the same data, a chi square test produced the SPSS output below. What can we conclude from this output?

 

 

Chi-Square Tests

 

 

 

 

 

 

 

Value

 

 

 

df

 

Asymp. Sig. (2-sided)

 

Pearson Chi-Square

 

82.112a

 

3

 

.000

 

Likelihood Ratio

 

84.840

 

3

 

.000

 

Linear-by-Linear Association

 

.105

 

1

 

.746

 

N of Valid Cases

 

933

 

 

 

 

 

0 cells (.0%) have expected count less than 5. The minimum expected count is 94.19.

 

 

Women watched significantly more programs than men.

Significantly more soap operas were watched.

The profile of programs watched was significantly different between men and women.

Men and women watch similar types of programs.

Chi-square is a test of

 

 

Difference,

Relationship

Association

Factors

 

 

In which sub-dialog box can the Chi Square test be found?

 

 

Frequencies: Percentages

Crosstabs: Statistics

Bivariate: Pearson

sex: Female