Lecture 3 Flashcards
ANOVA I
Video 1
What is the difference between a true and a quasi experiment?
In a true experiment the IV is manipulated, the people are randomly assigned and there is a control group. Whilst in a quasi experiment at least one of these is not true.
Video 1
What is a causal inference?
When other things also affect the experiment, but they are not taken into account.
Video 1
What type of experiment takes the causal inference into account?
A true experiment
Video 1
What is the problem when there is a higher variance?
Then the distribution overlaps more and this then obscures the difference in the group means.
Video 1
What causes systemic variance?
A known or manipulated factor (or factors)
Video 1
What causes unsystemic variation?
Factors that are not explained by the manipulated factors
Video 1
What does the t/F test show?
The ratio of systemic vs unsystemic variance
Video 1
How do you calculate the SD?
The difference between the observation and the mean.
Video 1
What does the SE actually state?
The unsystemic variation, or the variation in a sample.
Video 2
How can you make sure to get the max. t ?
By maximizing the mean difference, lower the variance in each group (unsystemic variation) or get a higher N
Video 2
When do you use a one-way ANOVA?
When there are multiple means
Video 2
When do you use a two-way ANOVA?
When there are 2 factors manipulated
Video 2
When do you use a three-way ANOVA?
When there are 3 factors manipulated.
Video 2
What is a familywise error?
This type of error is caused by the amount of tests that are used: 1*0.95, if it is more than 1 test it becomes bigger than 0.05 (too high alpha), can correct for it, but then it becomes less likely to get a significant result
Video 2
How can the pairwise comparison be calculated?
By using the formula [k (k-1)]/2
Video 2
Which are the two steps needed to take when doing an ANOVA?
Step 1: test the difference between k groups and test whether the null hypothesis of all means are equal
Step 2: do a posthoc comparison, whether the groups actually differ.
Video 2
What are the dfs used in the F statistic?
df1 = k-1, df2 = N-k
Video 3
How can you get a smaller deviation from the mean?
When you have different means per group instead of one big mean
Video 3
What is the formula to calculate the F statistic?
F = (SSm/df SSm) / (SSr/df SSr) which equals to MSm/MSr
Video 3
What formula is used to calculate the SSt?
(Xi - x(grand mean))^2 (of all the different group means added up) and the df is calculated by df = N-1
Video 3
What formula can be used to calculate the SSr?
(Xik - Xk (mean))^2 (of all the means combined), df = (N-1) - (k-1)
Video 3
Which formula can be used to calculate the SSm?
nk(Xk(mean) - (x grand mean)^2 (with all the means added up) and df= k-1
Video 4
What is a contrast?
A selection of hypothesis-driven pairwise comparisons
Video 4
What is a simple contrast?
When you compare a control group to each experimental group
