One-way between-subjects ANOVA

Question 1

assumptions

Answer 1

DV is interval or ratio level data
Data is normally distributed (histogram)
Homogeneity of variance (same variance)
Independent-random samples taken from each population

Question 2

SPSS outputs

Answer 2

between subjects factors
descriptives table
levenes test for equality of error variances
tests of between-subjects effects
profile plots

Question 3

between subjects factors

Answer 3

identify levels of factors and number of ppts

Question 4

descriptives table

Answer 4

mean and SD to report results
shows where differences occur if significant result

Question 5

levenes test for equality of error variances

Answer 5

use based on mean
want a non-significant value (p>0.05)
suggests variances aren’t significantly different

Question 6

tests of between-subjects effects

Answer 6

shows DF, mean square, F at significant level

Question 7

reporting the ANOVA (main effect) results
(what do we use?)

Answer 7

use tests of between subjects effects table
F( , )= . ,p<.001

Question 8

effect size calculation

Answer 8

sum of squares for BG condition (hypothesis)/total sum of squares (hypothesis+error)

partial eta square
eta square

Question 9

Cohen 1988 guidelines

Answer 9

effect sizes

small = 0.01
medium = 0.059
large = 0.138

Question 10

steps for formally reporting ANOVA

Answer 10

1. State ANOVA type, effect of DV
2. Present means and SD in table or text
3. Mention assumptions
4. Report ANOVA results (F ratio)
5. Report effect sizes and meaning
6. Report comparisons

Question 11

Next steps after the anova

Answer 11

planned/unplanned comparisons

Question 12

when to use planned comparisons for one way between subjects anova

Answer 12

when no significant main effect

used when hypothesised difference made in advance is explored

Question 13

planned comparisons for one way between subjects anova

Answer 13

Contrasts tests table (assumes equal variances line)
Contrasts effects sizes table (point estimate is effect size)
- Find both the above for each comparison

Report analysis resultsInterpret results

Question 14

reporting results of planned comparisons for one way between subjects anova

Answer 14

• Planned comparisons were performed to test hypotheses.
• See table 1 to see means and SD for each groups DV (draw table)
• Comment on means for DV and whether they were significantly higher or lower in each condition.
• t( )= . ,p<.001
• state effect and size

interpret in words

Question 15

when to use unplanned comparisons for one way between subjects anova

Answer 15

used if no hypothesised difference is explained

main effect needs to be significant to be run

Question 16

another name for unplanned comparisons

Answer 16

post hoc comparisons

Question 17

unplanned comparisons for one way between subjects anova

Answer 17

• Bonferroni post hoc test to reduce type 1 error
• multiple comparisons table shows significance between groups (p<.001 is significant, p>0.05 is non-significant)
• effect sizes (mean condition 1 - mean condition 2)/mean SD (see cohens effect sizes)

Question 18

reporting results of unplanned comparisons for one way between subjects anova

Answer 18

bonferroni post hoc tests revealed the DV of factor was significantly higher/lower than factor with a s/m/l effect size (p<.001), d= . (state for all comparisons)
interpret results in words