Flashcards in Chapter 14: Correlated Groups t-Test Deck (11):
When do we use single sample t-test?
- mean of null hypothesis population is unknown
- usually only have the mean
What are the limitations of single sample designs?(2)
- Usually mu is not known
- if u were known, cannot be certain if conditions under which mu was calculated are applicable.
Characteristics of correlated groups designs?(2)
- Each subject used before conditions ( before and after) aka control and experimental .
- Pairs of subjects matched on one or more characteristic serve in both conditions.
What is the information tested by correlated group design?(2)
- Difference scores:
* magnitude of difference scores
* direction of difference scores
What are difference scores?
- random sample from a population of difference scores having a mean of zero!
What is the formula for tobt for correlated groups?
* Dobt- uD/ Sd/>/N
or Dobt - uD/ >/ SSd/ N(N-1)
What is the formula for SSd for correlated groups?
- SSd= ED^2 - (ED)^2/N
List the steps to solving a correlated group question!
1. calculate differences from the before and after values, calculate D^2
2. Calculate the summation of D, summation of D^2
3. Calculate the Dobt= ED/N
4.Calculate SSd than Tobt
5. find df ( N-1)
6. evaluate the statistic based on its sampling distribution. ( look at alpha and use df, look at table D for tcrit )
Whats the conclusion for evaluation of the statistic?
- If | tobt | > | tcrit |, reject ho
What is the calculation for Cohen's D? (effect size)
- d= | Dobt |/ Sd
- Sd= standard deviation sample difference scores
- Dobt= value of mean of the sample difference scores
- d ? estimated d