You are on page 1of 6

Research paper

Alcohol and driving ability

One-way Analysis of Variance (ANOVA)

Thirty-six people took part in an experiment to discover the effects of alcohol on drinking ability. They were randomly assigned to three conditions: placebo (no alcohol), low alcohol and high alcohol. The non-alcoholic drink looked and tasted exactly the same as the other drink. Participants were weighted and given the appropriate amount of drink. After half an hour of drinking, participants drove in a simulator for ten minutes, and the number of errors made was automatically registered by the computer.

20 Januar 2003

20 Januar 2003

Research paper
Results: A one-way ANOVA revealed a significant effect of the amount of alcohol on the error rate (F(2,33) = 9.91; p = 0.001).

SPSS output

Degrees of Freedom
ANOVA SCORE

P-value

One-way ANOVA??????
Between Groups

Sum of Squares 97.556 225.667 323.222 df 2 33 35 Mean Square 48.778 6.838 F 7.133 Sig. 0.003

F(2,33) = 9.91; p = 0.001 ????

Within Groups Total

20 Januar 2003

20 Januar 2003

Learning objectives
Principles of ANOVA

ANOVA
Null Hypothesis: T h e m e a n s o f all g r o u p s a r e e q u a l . Experiment: Alcohol has no influence on driving ability

Read research papers

R e a d S P S S s printouts Between-groups variance vs. Within-groups variance

20 Januar 2003

20 Januar 2003

Within-groups variance

Within-groups variance

Group 1 Group 2 Group 3

Sources:
Individual differences Experimental error

Alternative Names: within, error

20 Januar 2003

20 Januar 2003

Between-groups variance
Grand mean

Between-groups variance

Group 1 Group 2 Group 3

Sources:
Treatment effects Individual differences Experimental error

Alternative Names: treatment; between; systematic

20 Januar 2003

10

20 Januar 2003

F-statistic
Small F:

F-statistic

Broadly: F is the ratio between Between-groups variance and Within-groups variance

Grand mean

The larger F the more likely to reject the null hypothesis

11

20 Januar 2003

12

20 Januar 2003

F-statistic
Large F: Grand mean

F-statistic

Grand mean

13

20 Januar 2003

14

20 Januar 2003

F-statistic
Large F: Grand mean

F-statistic
P-value given by F-distribution: F ( d fb e t w e e n , d fw i t h i n )

Degrees of Freedoms:
dfb e t w e e n = groups 1

dfw i t h i n = participants groups

15

20 Januar 2003

16

20 Januar 2003

F-statistic

Research Paper
F-statistic p-value

Important: Rejection of null hypothesis means at least one of the means is different from at least one other mean, but we do not know exactly which means a r e d i f f e r e n t f r o m w h i c h o t h e r m e a n s!

F(2,33) = 9.91; p = 0.001


Degrees of freedom

One-way
= one factor = one independent variable

17

20 Januar 2003

18

20 Januar 2003

SPSS output

ANOVA SCORE Sum of Squares Between Groups Within Groups Total 97.556 225.667 323.222 df 2 33 35 Mean Square 48.778 6.838

~ Variance

One-way Analysis of Variance


F 7.133 Sig. 0.003

Part 2

p-value

19

20 Januar 2003

20

20 Januar 2003

Research paper

Research paper
Results: A repeated-measures ANOVA revealed a significant effect of the amount of alcohol on the error rate (F(2,22) = 10.83; p = 0.001). Repeated-measure????

Alcohol and driving ability


Twelve people took part in an experiment to discover the effects of alcohol on drinking ability. The experiment included three conditions: placebo (no alcohol), low alcohol and high alcohol. The non-alcoholic drink looked and tasted exactly the same as the other drink. Participants were weighted and given the appropriate amount of drink. After half an hour of drinking, participants drove in a simulator for ten minutes, and the number of errors made was automatically registered by the computer. Each participant performed the task in all three conditions.

12 participants F(2,22 ) = 10.83; p = 0.001

21

20 Januar 2003

22

20 Januar 2003

Repeated-measures ANOVA
Between groups Degrees of freedom

Learning objectives

~ Variance

p-value

Principles of repeated-measures ANOVA

Tests of Within-Subjects Effects Measure: MEASURE_1 Source ALCOHOL Type III Sum of Squares 145.167 145.167 145.167 145.167 147.500 147.500 147.500 147.500 df 2 1.833 2.000 1.000 22 20.164 22.000 11.000 Mean Square 72.583 79.194 72.583 145.167 6.705 7.315 6.705 13.409 F 10.826 10.826 10.826 10.826 Sig. .001 .001 .001 .007

Sphericity Assumed Greenhouse-Geisser

Read research papers

Huynh-Feldt Lower-bound Error(ALCOHOL) Sphericity Assumed Greenhouse-Geisser Huynh-Feldt Lower-bound

R e a d S P S S s printouts

23

20 Januar 2003

24

20 Januar 2003

Two experimental designs

Total variance (between participants)

Between-participants design: Experimental design in which different participants serve under the different treatment levels

Within-participants design: E x p e r i m e n t a l d e s i g n s in which each participants receives all levels if at least one independent variable
Total Variance
25 20 Januar 2003 26 20 Januar 2003

Partitioning variance (between participants)


Total variance

Partitioning variance
Total variance Individual diff. variance

Withingroups variance Betweengroups variance

Error variance

Betweengroups variance

27

20 Januar 2003

28

20 Januar 2003

F-statistic

Degrees of freedom

Broadly: F is the ratio between Between-groups variance and E r r o r v a r i a n c e ( W i t h i n v a r i a n c e without individual d i f f e r e n c e s)

d fT o t a l = participants * groups 1 d fi n d i v i d u a l _ d i f f e r e n c e s = participants 1 d fw i t h i n = d fT o t a l d fi n d i v i d u a l _ d i f f e r e n c e s

P-value given by F-distribution: F ( d fb e t w e e n , d fe r r o r)

d fb e t w e e n = g r o u p s 1 d fe r r o r = df within - d fb e t w e e n

29

20 Januar 2003

30

20 Januar 2003

Repeated-measures ANOVA

Advantages

Tests of Within-Subjects Effects Measure: MEASURE_1 Type III Sum Source of Squares ALCOHOL Sphericity Assumed 145.167 Greenhouse-Geisser 145.167 Huynh-Feldt 145.167 Lower-bound 145.167 Error(ALCOHOL) Sphericity Assumed 147.500 Greenhouse-Geisser 147.500 Huynh-Feldt 147.500 Lower-bound 147.500 df 2 1.833 2.000 1.000 22 20.164 22.000 11.000 Mean Square 72.583 79.194 72.583 145.167 6.705 7.315 6.705 13.409 F 10.826 10.826 10.826 10.826 Sig. .001 .001 .001 .007

Avoidance of variability from participant to participant

Require fewer participants than betweenparticipant designs

Error variance (without individual differences)


31

Report Greenhouse- Geisser


20 Januar 2003 32 20 Januar 2003

Disadvantage

Counterbalancing

Order effects

or

carry-over (practice) effects

An arrangement of treatment conditions design to balance practice effects

Example:
Participant: Placebo; Low Alcohol; High Alcohol Participant: Placebo; High Alcohol; Low Alcohol Participant: Low Alcohol; Placebo; High Alcohol Participant: High Alcohol; Placebo; Low Alcohol Participant: Low Alcohol; High Alcohol; Placebo Participant: High Alcohol; Low Alcohol; Placebo Participant: like first participant
20 Januar 2003

C o u n t e r b a l a n c i n g for avoidance

33

20 Januar 2003

34

You might also like