Last minute research methods Flashcards
Sign test
Non parametric
- test of difference
- Repeated measures
- Data is nominal
Wilcoxon
Non-parametric
- Test of difference
- Repeated measures
- Data is at least ordinal
Unrelated t test
Parametric test
- Test of difference
- Independent groups design
- Data is at least interval
Chi- squared
Non-parametric
- Test of difference
- Independent groups design
- Data is nominal
Mann-Whitney U
Non-parametric
- Test of difference
- Independent groups
- Data is at least ordinal
Related t test
Parametric
- test of difference
- repeated measures design
- Data is at least. interval
Pearson’s r
Parametric
- test of relationship
- Data is at least interval
Spearman’s rank
Non-parametric
- test of relationship
- Data is at least ordinal
Assessing reliability (observations)
Inter observer reliability
- correlate the observations of the observers
- use spearman’s rank
- If there is enough agreement it would be +0.80 (high inter observer reliability)
Assessing validity (observations)
Concurrent validity
- correlate the scores from the observation from another known observation with high validity.
- A stronger correlation means it has high validity
Improving reliability (observations)
Do a pilot study
Make sure the behavioural checklists are fully operationalised.
Improving reliability (observations)
Train observers
So that there is no misunderstanding between the behavioural categories
Assessing reliability (experiment)
Test retest
- Repeat the test at a different time with the same ppts
- If the correlation of the data is positive then the reliability is high.
Assessing validity (experiment)
Face validity
- Look at whether the study is measuring what it was meant to measure.
- See if the measures used for the study are relevant to its aims.
Improving reliability (experiment)
- Fully standardising the procedure.
- Strict control over different aspects of the research, can increase the reliability of the study.
- A replication of the precise method will lead to high test-retest reliability.
Improving validity (experiment)
Improving ecological validity:
- make stimulus material as similar as possible to what would happen in real life- memory- avoid word lists and use meaningful info of the type that people would have to remember.
Assessing reliability (Self-report)
Test retest reliability
- Ppts should complete self report multiple times and have short intervals in between tests.
- Then correlate the results to see any strong correlation. - Similar results means there is a high test-retest reliability
Assessing validity (Self-report)
Look at whether the self report measure is measuring what it was meant to measure.
For example if the question was obviously related to the study and its aims.
Improving reliability (Self- report)
Pilot study
identify any ambiguous, unclear questions and then remove these questions so that they are not interpreted incorrectly.
Improving validity (Self-report)
Single blind
Make the ppts unaware of the aims of the questionnaire/interview and therefore less likely to answer untruthfully and social desirability bias is not a factor.
