Basic One-Way ANOVA

One-way independent measures ANOVA

“One-way” analysis of variance

Utilisation• Independent groups– One independent variable with two or more levels • Will most likely state randomly allocated.

– The independent variable (i.e. Colour) should become one variable name in variable view• Each level of the independent (i.e Red, Blue, Green,

Yellow) should become a value (i.e Red = 1, Green = 2, etc…)

– The dependent variable should be a separate variable name in variable view.

Variable view

Data inputReaction Time (hundredths of a second)

Warning Light

Yellow Red Green Blue

20 23 21 2620 20 21 2321 21 20 2222 21 23 2321 23 22 2320 22 20 2319 22 21 2121 21 22 2419 22 20 2520 22 19 22

Measured(dependent) variable data is here. Values should be put under each other, so column four under column three and column two under column on to create one long column of results. They should then be assigned values based on their original columns

These are the levels which will become values

This is the condition

This is the what the dependent variable is measured as

Data input

• Data should then look like this when put in SPSS as one long column of data and assigned values.

Running the analysis

• Analyse -> General Linear Model -> Univariate• Move the dependent variable to the

“dependent variable” box.• Move the independent variable to the

“fixed factors” box

Running the analysis

Running the analysis• Options -> Select Descriptive statistics -> OK.• If the independent variable has more than 2

levels (3+) then post-hoc comparisons will need to be used.• Post hoc -> Move independent variable with

more than two levels across to “Post hoc tests for:” -> Select Tukey -> Continue.

• As this is a one-way ANOVA a plot is not really needed.

Reading the result output

• 3 main tables (2 if no post-hoc ran)• Descriptive statistics• To show means and SD’s to judge direction of

significance.• Tests of between-subjects effects• To gain ANOVA equation and significance.

• Multiple comparisons• To judge significance between levels.

Reading the result output• Descriptive statistics• This will need to be reported but in a different way

to that produced by SPSS.• Values from this table will also be used in

explaining significant differences between levels.

Reading the result output• Test of between-subjects effects

F(3,36)=11.649, p<.001

Reading the result output• Test of between-subjects effects• For the exam, you could put a table in,

basically reshaping this one, and I think it may get you some marks.

Mean square F-value P-valuedfType II Sum of Squares

Reading the result output• Multiple Comparisons (Tukey post-hoc)

• Significant results will be marked with a star (*) next to the mean difference.

• It may be worth putting this in table form.• See below!

Reporting the result output

• Tables• Descriptive statistics• Test of between-subjects effects• Post-hoc significance

• Written description• See next slide

Reporting the result output• The use of a one-way analysis of variance (ANOVA)

showed a significant effect of warning light colour on reaction time in milliseconds [F(3,36)=11.649, p<.001]. Furthermore, a Tukey post-hoc pairwise comparison test revealed that the participants who viewed blue warning light (M=23.2) reacted significantly slower in comparison to those that viewed the red (M=21.7)[p=.032], yellow (M=20.3)[p<.001] and green (M=20.9)[p<.001] warning lights.