Week 10 Flashcards
When are t-tests used?
When you want to compare up to two sample means
What does a one-sample t-test tell you?
Does a sample mean differ from some known mean?
What does an independent samples t-test tell you?
Do means of two independent sample differ?
What does a paired-sample t-test tell us?
Do means of two related samples differ?
As a general assumption of t-tests, what kind of data is assumed?
interval or ratio data
As a general assumption of t-tests, what is assumed about the distribution?
Population of scores is normally distributed. T-tests are reasonably robust to violations.
As a general assumption of t-tests, what is assumed about the scores?
Scores for independent samples t-test are independent (e.g knowing one score provides no information about any other score).
As a general assumption of t-tests, what is assumed about the variance?
That the variance is equal (for independent samples t test).
- homogeneity of variance (assume SDs are equal)
- Levene’s test
Give some steps for using a one sample t-test with the example: do newborns show preference for more “attractive” faces?
- newborns shown pictures of various female faces to measure average time looking at them
- Locate critical region
- Calculate the t-value
- make a decision (compare obtained t to critical t)
Therefore, infants stares significantly longer at attractive pictures
When is an independent samples t-test used?
When we want to compare means from two groups
What does the denominator look like in an independent samples t-test?
The denominator is standard error of difference between means
In an independent samples t-test, what is known about the t-distribution?
The t-distribution varies with sample size.
With the example of “how much anorexic individuals eat is a result of how much they feel in control”, demonstrate how an independent samples t-test would be used.
- Twelve anorexic females were randomly assigned to a high control and low control group
- Participants were expected to taste test. High control was hypothesis to eat a higher mean of cereal.
- Two means of low and high control group were compared
- Test of equality of variance. If p
How can we summarise an independent samples t-test?
t value, p value, 95% CI and cohens d.
Does Jamovi provide numerical 95% confidence intervals?
No.
What will it mean if 95% CIs on means for independent t-test overlap no more than ~25% of their total length?
It will mean that the difference is statistically significant.
What data is needed for a paired-samples t-test?
Interval or ratio
What kind of distribution is needed for a paired-samples t-test?
population of scores is normally distributed (t-tests are reasonably robust to violations)
With a paired samples t-test, what do the scores usually require?
For them to be independent (knowing one pair of scores provides no information about any other pair of scores)
What is the expectation of scores in a paired-samples t-test?
Expect that pairs of scores will be correlated (not a requirement)
What is a paired samples t-test actually testing?
Is the difference between means different from zero?
What is an example of when a paired-sample t-test should be used?
Investigating whether reaction time for bimanual movements depends on which arm has to move accurately to precise location.
What does it mean if in a paired-samples t-test (within-subject design), the 95% CIs overlap?
It means that visual inspection is not able to answer questions of statistical significance.
If in the 95% CIs for two conditions overlap at any amount for a within subjects design, what must you do?
Consult the p-value in the t-test results for statistical significance.
For a paired-samples t-test, what does Cohen’s d not take into account?
The correlation of the two conditions. This effect size is not directly comparable with d from independent samples t-test
What is effect size in an t-test?
significance test is sensitive to sample size. Even small effect can be significance with a large enough n.
What do we use to measure the effect size when testing significance (t-test)?
Cohen’s d. Measure of effect size which states the size of effect in SD units.
What does a cohens d of 0.2 mean?
A small effect
What does a cohens d of 0.5 mean?
A moderate effect
What does a cohens d of 0.8 mean?
A large effect
