Unrelated ANOVA Flashcards
What is the unrelated ANOVA?
It is an extension of the independent t-test. It requires data from two or more different conditions.
What does the unrelated ANOVA give us?
It estimates the population variance by finding the cell means variance (between variance/truth) and the random population variance (within variance/error)
It calculates the overall variance across conditions.
What formula is ANOVA based on?
Variance but with new components.
It is applied to calculate two new forms of variance.
True and error variance
What does the ANOVA assume about our obtained score?
That it is made of two components. True component and error.
What are errors?
Comes in many forms, can be tiredness, experimenter error, hangover etc.
How do we find the ‘true scores’ when doing an ANOVA?
We find the means of each group. We then replace the scores with the groups mean. This is the true score.
How do we find the error score?
We just subtract the mean score from the data.
What mean do error scores always give us?
0
How do we estimate population variance in unrelated ANOVA?
We use these scores separately to estimate the population variance.
The formula states that to find the Estimated Population Variance (F score) you divide the variance estimate for the true scores by the variance estimate for error scores.
Then check F against significance tables
What are we looking for in the unrelated ANOVA?
If null hypothesis is true there will be no difference between these scores, but if alternative hypothesis is true then the true scores should be more than the error score.
We are not simply comparing the 2 estimates. It is one-tailed. We need to know if true variance estimate is bigger than errors’
Is the degrees of freedom worked out differently in the unrealised ANOVA?
Yes, it’s different for true and error.
For true you calculate: no. Of conditions - 1
For error you calculate: N - no. Of conditions
Does the variance formula change when conducting the unrelated ANOVA?
Yes, only the bottom half, the degrees of freedom. This is worked out differently for both the true estimate and error estimate.
Does the unrelated ANOVA show us statistical difference across or between the groups?
It only shows us a difference across the groups, not between. To find out the difference between the groups, we can run post-hoc tests on spss. We can only do this if our ANOVA comes out significant.
How do we conduct the unrelated ANOVA in spss?
Analyse -> compare means -> one-way ANOVA then I out variables and if you want to select options and choose descriptive and homogeneity of variance.
What is an ANOVA?
ANOVA is used when looking to compare three or more groups. It stands for Analysis of Variance
