Research Methods Flashcards
Statistical testing
Used to determine the likelihood that a relationship has been found due to chance
It determines whether a hypothesis should be rejected or accepted
Spearman’s Rho
Use for testing correlation
When data is ordinal
Pairs of scores
Pearson’s R
Use when testing correlation
Data is interval
Pairs of scores
Mann-Whitney
Use when independent measures design
Data is ordinal
Testing for difference
Unrelated T-test
Use when independent measures design
Data is interval
When testing for difference
Wilcoxon
Use when repeated measures/matched pairs
When data is ordinal
Testing for difference
Related T-test
Used when repeated measures/matched pairs
When data is interval
To test for difference
Chi-Squared
Use when independent groups design
Data is nominal and data within each category is independent
Used to test for association/difference
Nominal Data
Represented in the form of CATEGORIES e.g small or tall
Participants cannot be in more than one category which makes the data discrete
It is the crudest and least informative type of data as it does not have a numerical result for each participant
Ordinal Data
Represents data in PLACE ORDER e.g 30 students would be numbered with the tallest being 30 and the smallest being 1
A strength is that the intervals are equal however the intervals are not equal between each unit
It is criticised as being based on subjective opinion rather than on objective measures
Interval Data
Based on STANDARDISED measurements e.g weight or time
The most informative and accurate data
Significance level
“p<0.05” means that the probability of the results occurring by chance is less than 5 times in 100
Null Hypothesis
States there is no difference/association between variables
Type One Error
When the researcher rejects the null hypothesis (accepts the alternative hypothesis) when in fact the effect is due to chance
Known as error of optimists
Type Two Error
When the researcher rejects the alternative hypothesis (or accepts the null hypothesis) when in fact the results occurred due to chance
Referred to as error of pessimists
One tailed test
Used when the hypothesis is DIRECTIONAL
If previous research has been conducted to suggest the direction of the results it is used
Two tailed test
Used when the hypothesis is NON-DIRECTIONAL
If no previous research has been done to suggest the direction of the research, a two tailed test is used
Case study
A detailed study about an individual or a group of individuals that is either LONGITUDINAL or RETROSPECTIVE
Strength and Weakness of a Case Study
It gives a lot of depth and understanding of the individual and allows us to investigate a situation that could not be set up for ethical reasons
If is difficult to replicate therefore it is difficult to say whether or not it is reliable. The findings can also not be generalised
Content Analysis
A technique of analysing qualitative data
It is an indirect way is studying behaviour by analysing the communications the individual has used such as texts or emails
It uses coding to get quantities data from qualities data
By identifying categories e.g. words or phrases, the researcher then goes through the communications and uses a tally chart to go through and look for examples
Evaluation of Content Analysis
It is easy to perform and doesn’t require contact with the participant
HOWEVER
It is descriptive and just not show the underlying reasons for the behaviours. Reasons may be attributed to the author which were not intended
Thematic Analysis
Involves assessing data to identify patterns within it. For example, a thematic analysis of the portrayal of the mentally ill in the press may involves looking at themes such as "they drain NHS resources". It has 6 stages: Familiarisation Coding Searching for themes Reviewing themes by collecting new data Defining and naming themes Writing up
Assessing Reliability
Test-Retest
Inter-observer reliability
Split-half method
Questionnaire
Test-retest
Tests for EXTERNAL reliability
It involves getting the same participants to perform he same tests after a time delay.
If the results are similar then the test would be deemed reliable
Inter-Observer Reliability
Used to assess the reliability of OBSERVATIONS
Could be measured by getting two observers the observe the phenomena
A strong correlation of scores (0.8) would suggest that the observation was reliable
Split-Half method
Tests for INTERNAL reliability
Involves splitting the test in half and if the results are similar on each side it is deemed reliable
Improving reliability
In observational research, the observer could gain more training
Or create a clear criteria for behaviours categories
Internal Validity
Whether or not it measures what it is suppose to measure i.e the effect of the IV on the DV
Reduced by demand characteristics and investigator effects
External Validity
Comprised of ecological and temporal validity:
ECOLOGICAL- the extent to which findings can be generalised to other situations and settings
TEMPORAL- whether findings can be generalised to other historical times and eras