Research methods A2 Flashcards
Define content analysis
The systematic research technique in which qualitative data is analyzed to produce quantitative results, through the use of coding systems or themes to categorize behavior.
State two advantages of content analysis.
- Converts qualitative data to quantitative data
2. Reliable way to analyse qualitative data - coding units are objective
State two disadvantages of content analysis.
- Causality can’t be established - just describes
2. Can’t show deeper meaning or explanation for data patterns
How does thematic analysis differ from content analysis?
Thematic analysis involves the re-analyzing of data to identify broader patterns of meaning (themes)
Define thematic analysis.
A qualitative approach to content analysis, involves identifying prominent, recurring themes which are usually more descriptive and general than coding categories
Define reliability
How consistent the findings from an investigation or measuring device are
Define internal validity
The extent to which a test is investigating the true effect of the IV on the DV
Define concurrent validity
Measurement of whether the results of the new test match the results of a similar, but older test
Define face validity
Measurement of whether the test or measuring instrument is measuring what it should
Define external validity
The extent to which results can be generalised to other settings/people/time eras (factors outside the research setting)
Define coding
The stage of quantitative content analysis in which the communication is analyzed by identifying each instance of the chosen categories
Define test-retest
When the observation is repeated and the results are compared with the first observation - they should produce a correlation coefficient of +0.8 or more
What is a pilot study? What is it’s purpose?
When a small trial run of the observation is run before the main study is carried out.
They ensure the procedures and instruments used improve precision of results e.g.standardised instructions, debriefing, and operationalisation. It minimises human error and variation.
Define inter-observer reliability.
More than one psychologist/observer observes the behavior separately. The results are compared to obtain a correlation coefficient of +0.8 or more
Define inter-researcher reliability
More than one psychologist/researcher conducts an interview with the pps - act in similar and consistent ways to make it reliable. Results are compared to gain correlation coefficient of +0.8 or more
How can reliability be improved?
Observation:
- Operationalise categories
- Give more training and practice to observers
Self-report:
1. Make sure the question isn’t ambiguous - rewrite
Experiments:
1. Standardised instructions
What factors could reduce internal validity?
- Investigator effects
- Demand characteristics
- Confounding variables
- Social desirability bias
- Lack of operationalisation
How could internal validity be assessed?
- Concurrent validity
2. Face validity
How could internal validity (concurrent and face) be improved?
Concurrent:
Remove irrelevant/ambiguous questions
Face:
Getting an expert to examine the questions on the questionnaire - rewrite any ambiguous questions
How can you reduce investigator effects to improve internal validity?
- Get rid of psychologist completely
2. Use multiple researchers
How can you reduce demand characteristics to improve internal validity?
Single/Double blind study
How can you reduce confounding variables to improve internal validity?
Carry the experiment out in a controlled environment (lab study)
Define temporal validity
Whether the findings are true over a period of time and can be generalised to other time eras
How can external validity be assessed?
1) Meta analysis - comparison of findings of different research studies studying the same hypothesis
2) Check environment - should be naturalistic
3) Assess how the dependent variable is measured - high mundane realism
4) Assess natural behaviour of pps
How can external validity be improved?
- Reduce demand characteristics through double/single blind study
- Carry out in naturalistic setting e.g. Field experiment, covert experiment
Define a paradigm
A shared set of assumptions and agreed methods that are found within scientific disciplines
Why is Psychology seen as a ‘pre-science’?
There are too many internal disagreements and conflicting approaches, but it does have some scientific qualities
What are key scientific features?
1) Systematic, controlled approach
2) Use of experiments and observational studies
3) Falsifiability
4) Objectivity
5) Replicability
Define empirical methods
A method of gaining knowledge that relies on direct observations or testing. This can help separate unfounded beliefs from the real truth. Theory cannot be scientific unless tested empirically
Define a paradigm shift
The result of a scientific revolution. A significant change in the dominant unifying theory of a scientific discipline occurs, causing a paradigm shift.
Describe the stages of a paradigm shift.
1) One theory remains dominant. Researchers begin to question the theory and have contradictory research.
2) Evidence accumulates - can’t ignore
3) Present paradigm overthrown - new one established
How do paradigm shifts help science?
Allows for rapid progress to be made - scientific revolution
Define objectivity.
Dealing with facts in a way that is unaffected by beliefs, opinions, feelings or expectations
Define replicability.
The extent to which the findings of research can be repeated in different contexts and circumstances
Why is replicability important?
- Guards against scientific fraud
- Check to see if results were a “one off”
- Check for reliability
- Check for validity
Define falsifiability.
The notion that scientific theories can potentially be disproved by evidence. Refers to proving a hypothesis wrong.
Why is falsifiability important?
It’s the only way to be certain of a theory - have to be able to conduct tests to try and prove a theory wrong, just one contradictory piece of evidence can refute said theory
What is a limitation of psychology in regards to objectivity?
- It’s humans studying humans - difficult to study without being biased
- Psychologists from different approaches may examine in different ways e.g behaviourist wouldn’t instinctively look for cognitive elements - biased
What is a limitation in regards to falsifiability?
Many of the theories can’t be disproved, so can’t be certainly true e.g. Psychodynamic approach - Id, ego, superego
What are the stages in deductive reasoning?
- Propose theory
- Develop hypothesis
- Test theory
- Conclude
What are the stages in inductive reasoning?
- Observe instances in environment
- Develop hypothesis
- Test hypothesis
- Conclude
- Devise theory
Define ‘level of statistical significance’.
The level at which the decision is made to reject the null hypothesis and accept the experimental
Define significance
How sure we can be that the IV is having an effect on the DV and isn’t due to chance. Shows that there’s a real difference between the sets of data
Define probability
A numerical measure that determines the extent to which our results are likely due to chance or whether there’s a real difference between the experimental and control conditions
What is the conventional significance level?
5% (p<0.05)
Why is 5% significance used?
- Not too lenient or strict
2. Minimises chance of type 1 or type 2 error
When would a 1% significance level be used?
When research findings are critical and important e.g drug testing
What is a type 1 error?
False positive.
Null hypothesis is rejected even though it should’ve been accepted. There is no real difference between the two variables.
What is a type 2 error?
False negative.
Null hypothesis should’ve been rejected instead of accepted. There is a real difference between the two variables.
State three advantages of using correlations.
- Can establish strength of relationship between two variables
- Can investigate variables that couldn’t be manipulated experimentally
- Predictions can be made of one of the variables based on results
State three disadvantages of using correlations.
- Cannot determine cause and effect
- May be a confounding that is influencing results
- Only measure linear relationships - doesn’t detect curvilinear relationships (bell-curve)
Describe nominal data
Data consisting of numbers of pps falling into different categories - can’t be placed in both categories
Describe ordinal data
Data that can be ranked from highest to lowest (or vice verse) - can be unequal intervals
Describe interval data
Data that has fixed and even intervals
How do parametric tests differ to non-parametric tests?
Parametric tests rely on data collected rather than the order of the data. Parametric tests are more likely to detect if data is significant or not.
What three factors allow conduction of a parametric test?
- Level of measurement is INTERVAL
- Data from normal distribution curve
- Data has similar variance/spread of scores
What factors help to choose the correct statistical test?
- Test of difference, correlation or association?
- What design is used?
- What level of measurement is used?
- Is the test non parametric or parametric?
If testing for a correlation what statistical tests should be used?
- Spearman’s Rho
2. Pearson’s R
If testing for a difference which statistical tests should be used?
- Mann Whitney
- Chi-Squared
- Wilcoxon
- Unrelated T-test
- Related T-test
If testing for an association which statistical test is used?
Chi-squared
In which two instances is chi-squared used?
- Non Parametric - Nominal data - Test of Difference - Independent measures
- Non Parametric - Nominal data - Test of Association
In which instance is the sign test used?
Non Parametric - Nominal - Test of Difference - Repeated measures/Matched pair
In which instance is Mann Whitney used?
Non Parametric - Ordinal - Test of difference - Independent measures
In which instance is Wilcoxon used?
Non Parametric - Ordinal - Test Of Difference - Repeated Measures/Matched pair
In which instance is Spearman’s rho used?
Non-Parametric - Ordinal - Test of correlation
In which instance is Unrelated T-test used?
Parametric - Interval - Test of difference - Independent measures
In which instance is Related T-test used?
Parametric - Interval - Test of Difference - Repeated measures/Matched pair
In which instance is Pearson’s R used?
Parametric - Interval - Test of Correlation
An investigation into whether psychology students or maths students revise differently occurs. They must study only one of the subjects. What’s the level of measurement, and statistical test used?
Nominal data
Chi-squared
Define the critical value.
A numerical value found in the statistical tables that helps determine the significance of the results
What does Na stand for in the Mann Whitney test?
Number of pps in smaller sample
What does Nb stand for in the Mann Whitney test?
Number of pps in larger sample
What does T stand for in the Mann Whitney test?
Sum of the ranks in the smaller sample
Define ‘degrees of freedom’
The number of values that are free to vary
What are the steps to working out the results in a Wilcoxon test?
- Calculate sum of positive ranks
2. Calculate sum of negative ranks
What does T stand for in the Wilcoxon statistical test?
The smaller value
In which statistical test is ‘N-1’ used?
Related T-test
In which statistical test is ‘N-2’ used?
Unrelated T-test
