Lecture 11 - Analysing 3+ Conditions Flashcards
How could the problem of increased alpha from running multiple t-tests be solved, and what is the main problem with this?
Reducing the alpha for each test to 1.66666% making the combined error still 5%.
This reduces power.
What is the main issue with conducting multiple t-tests?
If each test has an alpha of 5%, the overall likelihood of error becomes just under 15% (1-(0.95sq))
What is an ANOVA?
An extension of the t-test which allows us to test whether 3 or more population means are significantly different, without reducing power.
What are the different types of ANOVA?
ANOVA - one-way/one factor, multi-factor
MANOVA - Multivariate Analysis of Variance (extension for multiple DVs)
ANCOVA - Analysis of Covariance (extension for continuous variables, e.g. correlations)
Who invented the ANOVA, what is its test statistic and what does it actually do?
Fisher = F
Compares the amount of variance explained with the amount of variance unexplained
What are the assumptions of ANOVA?
- the scores were sampled randomly and are independent
- roughly normal distribution
- roughly the same number of pts per group (improved power and special version if exactly the same)
- roughly equal variance for each condition
In between-groups ANOVAs, what is the unexplained and explained variance?
Explained = variance between groups Unexplained = variance within a group
What is another word for the variance calculation?
Mean squared (MS) error
How is the F-ratio calculated?
F= MS between (the effect) / MS within (the noise)
= explained/unexplained
What degrees of freedom values do ANOVA critical values require?
Degrees of freedom between conditions and residual degrees of freedom
If an ANOVA confirms that groups differ, what should be done?
Run the multiple comparisons (some are planned comparisons some are post-hoc tests)
Correct for multiple comparisons (Bonferroni)
How do you conduct a single IV between-groups ANOVA in SPSS?
Analyse|CompareMeans|One-way-ANOVA
How do you conduct a between-groups ANOVA in SPSS (generally)?
Analyse|GeneralLinearModel|Univariate
When conducting an ANOVA in SPSS, what goes in the ‘fixed factor’ box?
The IV(s)
In an SPSS ANOVA, how do you select post-hoc tests?
From the univariate menu, select ‘postHoc’ and insert the correct factors, selecting ‘Bonferroni’ underneath.
