Research Methods 3 Flashcards
Probability
-Significance levels
- Significance levels are assigned to establish the probability of the results being due to chance. If this is acceptably low, then we can reject our null hypothesis and we accept our experimental/alternative hypothesis
- If the results of a statistical test are significant we accept our experimental hypothesis and reject the null hypothesis
- If the results are not significant we must accept the null hypothesis and reject the experimental hypothesis
Analysing data
-Nominal data (+strengths and weaknesses)
Nominal data = A frequency count for distinct categories where something can only belong to one category e.g. number of people who pass or fail a driving test
- Strengths = Reliable as it is taken directly from a survey or another collection method
- Weaknesses = As it does not give a numerical score for each ppt, this type of data does not permit sensitive analysis
Analysing data
-Ordinal data (+strengths and weaknesses)
Ordinal data = This is where numbers can be placed in ascending or descending rank order e.g. on a rating scale where 1= unattractive and 10= highly attractive
- Strengths = There is ease of comparison between the variables, convenient to group variables and efficiently used in surveys
- Weaknesses = It uses subjective data as scaled data is open to interpretation
Analysing data
-Interval data (+strengths and weaknesses)
Interval data = Measurements are taken from a scale where each unit is the same size and the gap between each unit is fixed and equal e.g. the difference between 100 and 90 is still the same as the difference between 90 and 80
- Strengths = Easy to generate from closed questions, points are linear with equal gaps and there are more points of equal value
- Weaknesses = The lack of absolute zero, there is no true zero point or fixed beginning
Analysing data
-Statistical tests
Chi square = test of difference, independent measures, nominal data
Man Whitney = test of difference, independent measures, ordinal data
Unrelated T test= test of difference, independent measures, interval data
Sign test= test of difference, repeated measures, nominal data
Wilcoxon test=test of difference, repeated measures, ordinal data
Related T test= test of difference, repeated measures, interval data
Spearman’s Rho= test of relationship, correlational design, ordinal data
Pearson’s R= test of relationship, correlational design, interval data
Probability
-Probability stats
- p< (or equal) 0.01: the probability that the results are due to chance factors is less than or equal to 1%
- p< (or equal) 0.005: the probability that the results are due to chance factors is less than or equal to 0.5%
- p< (or equal) 0.001: the probability that the results are due to chance factors is less than or equal to 0.1%
Type 1 and type 2 errors
Type 1= Known as a false positive and occurs when a researcher incorrectly rejects a true null hypothesis
- A type 1 error is more likely to occur with a p<0.01
Type 2= Known as a false negative and occurs when a researcher fails to reject a null hypothesis which is really false
- A type 2 error is more likely to occur with a p<0.10
Reliability
-External reliability
External reliability = Producing the same results each time the test/ study is carried out
-The test retest method assesses external reliability
Reliability
-Internal reliability
Internal reliability = This is concerned with the consistency within a test. It is usually associated with questionnaires and tests e.g. personality
-The split half method assesses internal reliability
Reliability
-Observer reliability
Observer reliability = When 2 or more observers produce the same record of their observations
-The inter observer (or rater) assesses observer reliability
Reliability
-The split half method
- This compares a person’s performance on 2 halves of a test or questionnaire
- You would expect a person’s performance on the first half of the test to be consistent with their performance on the second half of the test
Reliability
-The test retest method
- This is where a test is repeated several times using the same participants
- if it is reliable you would expect the same or very similar results each time from the same individuals
- You would then correlate the results and hopefully obtain a strong positive coorelation
Reliability
-Inter observer (or rater) reliability
- This is where the observations of two or more observers are compared for similarity
- The observers would first be trained in the use of a coding system so they can identify and are clear on exactly what they are looking for
- Once you have completed the observation, you would correlate the results on each coding category and hopefully obtain a strong positive correlation
Improving reliability
- Take multiple measurements - taking more than one measurement from each participant reduces impact of any anomalous scores
- Alteration of the experimental method - fix control variables through choice of equipment
- Improve the reliability of single measurements - Use averaging and be more precise and accurate with recordings. Use the line of best fit to improve measurement recordings for the results of whole studies
Validity
-Internal validity
-In an experiment, the extent to which our findings are due to the manipulation of the IV and not any other uncontrolled variables
Validity
-Temporal validity
-A type of external validity that concerns the extent to which research findings hold true over time
Validity
-Ecological validity
-The extent to which an experimental effect can be generalised from the other study to other settings and situations
Validity
-Population validity
-The extent to which findings can be generalised from the sample tested to other populations
Validity
-Face validity
-The extent to which research looks as though it is doing what it claims to, on the surface levels
Validity
-Concurrent validity
-The extent to which a new measure (or instrument) compares (or concurs) to a previously validated measure
Issues with validity
- Confounding variables = Those ‘uncontrolled variables’ above can get confused with the IV. If these are not removed then our study will have low internal validity
- Demand characteristics = The social desirability effect or ‘screw you’ effect. The cues or clues that give away the study’s aim
- Investigator bias = The experimenter manipulates what you say for the study
4 ways to assess validity
- Face validity = The eye ball test
- Concurrent validity
- Temporal validity
- Ecological validity
7 ways to improve validity
- Improving concurrent validity through concurrent validity checks
- Single blind technique - improves internal validity
- Double blind technique - improves internal validity
- Eye-ball test by ‘experts’ - face validity
- Large samples and careful sampling - improves population validity
- Improving temporal validity through replication of tests with periodic testing
- Improving ecological validity through replicating research in different settings and even diverse methods
Conventions of reporting psychological investigations
- The Title
- The Abstract
- The Introduction - literature review
- The Method
- The Results
- The Discussion
- This includes the 4 pillars; identify flaws is your research, Discuss whether results are in line with previous results, discuss why your findings were found, identify future directions for research - Appendices
- References