Research Methods (A2) Flashcards
Define validity
When it measures what it claims to measure
What’s external validity about
External validity is about applying/generalising the results of a study
What are the 3 types of external validity
- Ecological validity
- Population validity
- Temporal validity
What’s low ecological validity
When results cannot be generalised to behaviour in the real world, as the study environment isn’t reflective of real life
- high ecological validity is the opposite
What’s low population validity
When results cannot be generalised to the target population as the sample isn’t representative enough of the target population
- high population validity is the opposite
What’s low temporal validity
When results cannot be generalised to modern day behaviour
- high temporal validity is the opposite
What’s internal validity about
Whether or not the study has been conducted accurately, whether data has been collected accurately, and whether or not the study has measured what it thought it measured
There are 5 threats to internal validity
- Situational variables
- Participant variables
- Investigator bias
- Demand characteristics
- Social desirability bias
How and why do Situational variables threaten internal validity
How could this be overcome
- E.g. time of day, noise, temperature
- Factors in the environment that affect a condition of the IV. This change caused results rather than the IV. = Low Internal validity
Overcome by all participants having same environment conditions (standardised procedures)
How and why do Participant variables threaten internal validity
How could this be overcome
- E.g. mood, intelligence, gender, age, personality
- Features of Personality In participants account for results rather than the IV. = Low Internal validity
Overcome by using Same participants in both conditions of IV (repeated measures), or match upon certain criteria (matched pairs)
How and why does investigator bias threaten internal validity
How could this be overcome
- Where the researched could have influenced participants in some way so behaviour isn’t accurate
Overcome by using a double blind strategy
How and why do demand characteristics threaten internal validity
How could this be overcome
- Where participants have guessed the aim of the study and changed behaviour to spoil study, therefore accurate behaviour hasn’t been measured
Overcome by using deception- hide aim from participants
How and why does Social desirability bias threaten internal validity
How could this be overcome
- Where participants change answers or behaviour to make themselves look better so accurate behaviour hasn’t been measured
Overcome by making questionnaires anonymous, or study people without their knowledge (covert observations)
What’s the 2 ways of checking validity
- Concurrent validity
- Face validity
What’s concurrent validity
How’s it done
Concurrent validity is a check that the measuring tool you are using is equal to an existing validated measuring tool
- Its checked by comparing a participants scores on your measure with their scores on an existing measure
What’s face validity
How’s it done
Face validity is whether a test, scale or measure appears “on the face of it” to measure the thing it’s supposed to measure
- Checked by examining two measuring tools closely or having tool examined by an expert
Two ways of checking reliability of a study
1) test-retest = doing a study/experiment and then repeating, 2 weeks later for example
2) inter-rater reliability (inter-observer)= having two researchers/observers so they can check each others scores afterwards
What’s the order of a psychology report
8 things
1) Abstract
2) Introduction
3) Method
4) Results
5) Discussion
6) Conclusion
7) References
8) Appendices
What does the abstract include
- IV, DV, hypothesis, aim, method and brief summary of background research and results
- Also, includes the experimental design and type of sample
What does the introduction include
- What was previously done, and what you expect to happen (background research). Aim and hypothesis at the end
What does the method include
- What design you’re using and why, IV and DV, attempts to control confounding variables
- Procedures detailed so a person who’s never done psychology could understand what needs to be done
What does the results include
- Descriptive statistics involves measures of central tendency (mean, median, mode) also measures of dispersion
- Types of data (nominal, ordinal, interval, ratio)
- Inferential statistics
- When deciding inferential statistics you take into account design, data and difference
What does the discussion include
- Explaining findings, was the hypothesis supported
- Reference type 1 and 2 errors
- Background research and results for them
- Limitations of your research, what would you do differently, what’s required in further research
What’s a critical value
The values in the statistical tables, which determines significance
- because you compare to observed value
What’s the calculated or observed value
Results from a calculation
What’s inferential statistics
Used to make inferences to establish whether results are significant.
- Used to draw conclusions and predictions
What’s degrees of freedom
Number of participants there are
What’s significance level
Importance level of study
What’s probability
The likelihood results are due to chance
What’s a null hypothesis
Hypothesis that predicts there will be no effect
No change in DV when IV is manipulated
What’s an alternative or experimental hypothesis
Hypothesis that predicts there will be an effect
Will be a change in DV when IV is manipulated
What are descriptive statistics
Describes the data and what it shows (mean, median, mode)
The 4 things needed to understand statistical tables
- Number of participants
- Directional (one-tailed) or non-directional (two-tailed) hypothesis
- What type of design used
- How confident are you that IV caused DV. E.g. 5% level means 95% sure IV causes the DV
What does 5% level mean
What does 2% level mean
- 5% level means researcher is 95% sure the IV caused the DV
- 2% level means the researcher is 98% sure the IV causes the DV
What level is most common and why is the 2% level used
- 5% level is the most common, 95% sure the IV caused the DV, as it’s easier to prove than 98%
- 2% level is used when it’s more serious for example drug testing
What happens when the result is significant
We can reject the null hypothesis and accept the alternative hypothesis
What happens when the result is not significant
We reject the alternative hypothesis and accept the null hypothesis
What’s a type 1 error
- Rejecting the null hypothesis when we shouldn’t.
- Happens when probability level too high (probability level at 98% instead of 95%)
What’s a type 2 error
- Accepting the null hypothesis when we shouldn’t
- Happens When probability level is too low (probability level at 95% instead of 98%)
Non-parametric data is what
When we have nominal or ordinal data
- So chi-squared, sign Test, wilcoxon, spearman’s rho and Mann-Whitney U test is used
Parametric data is what
When there’s interval data, normally variability of scores, and distributed for each condition should be the same
- So independent t test, related t test and Pearson’s test is used
How do you do the sign test
1) Add up the (+) and (-). In this case n=5
2) To calculate observed value, add up number of times the less frequent sign occurs (s). In this case the sign value is S=2
3) Work out whether it’s one or two tailed hypothesis, we will assume it’s two tailed. 0.05 (95%) significance level. Check sign value against table of critical values.
4) For study to be significant, observed value has to be smaller or equal to critical signs test value
What are the antonyms used for statistical tests
Chi-squared
Independent measures
Nominal
Mann-Whitney
Ordinal
Independent measures
Interval
Spearmans rho
Ordinal
Correlation
Wilcoxon
Ordinal
Repeated measures
What’s an experimental design
The way in which participants are organised and allocated to conditions
What’s an independent groups design
Two groups for different conditions. Results are compared across groups
- They can be in more than one condition, but results are compared across groups and not across conditions
What’s a matched pairs design
Participants are matched due to a common characteristic
- most effective way to resolve nature vs nurture debate (separating mono and dizygotic twins, and see if they develop same characteristics or not)
What’s a repeated measures design
When the same participants are used for all conditions. Participants take part in more than one condition. Results are compared across conditions
What’s counterbalancing
When a group is split to experience events in a different order
- resolves order effects
What’s attrition
When participants drop out of a study, data is then lost
Adv and disadv of independent groups
Adv= - Don’t have to control for Participant variables, as participants are randomly allocated to conditions
Disadv= - Need a larger sample of people
- More time-consuming, therefore costly
Adv and disadv of repeated measures
Adv= -Smaller Number of people needed
-Less time-consuming, not as costly
Disadv= -Order effects, to resolve this, counterbalancing is used
Adv and disadv of matched pairs
Adv= -Solve nature vs nurture debate
Disadv= -Trying to find participants who have same characteristics
-Attrition, because matched pairs are done over a long period of time
What’s a paradigm
A shared set of assumptions, methods and terminology about what should be studied and how
Example of paradigms
The assumptions of the approaches
When is there a paradigm shift
When new studies emerge, or new ideas and evidence are put forward
AO3 points for paradigms
- Appear at different times, therefore there are various paradigm shifts
- Psychology hasn’t fully developed into a scientific approach yet, due to the fact it has many paradigms
Explain what is meant by a paradigm shift (4)
PPQ
- A paradigm is a set of shared assumptions/beliefs about what should be studied and how
- A shift occurs when members of science change from one established way of explaining behaviour to a new way
- This shift leads to a ‘scientific revolution’
What is a science
Something that can be used to make predictions, due to it being tested and re-tested
- Predictions are then validated or falsified on the basis of objective evidence
What does falsifiable mean
Test the study to try and prove it wrong
What’s the reason for scientific studies evolving and changing
Because psychologists/scientists are trying to falsify / prove it wrong.
E.g. new medications
Criteria for a science
- constructing theories and testing hypotheses
- falsifiability
- replicability
- objectivity
- having a paradigm
AO3 point on people assessing people
If people are assessing people there’s an implicit subjective bias, because the researcher chooses the aim, hypothesis, method, sample, location, results, conclusion
What’s an inductive scientific method
Theory is constructed before the hypothesis
What’s a deductive scientific method
Theory is constructed after the hypothesis
What approach challenges that psychology is a science
The psychodynamic approach, as Freuds studies are unfalsifiable
What study supports the idea that psychology is a science
Skinner (operant conditioning)
Methods to increase objectivity
- double blind collection of data
- standardised procedures
- operationalised variables
- Peter review
- controlled variables
What research methods are replicable and not replicable
Replicable= Lab, controlled observations
Non-replicable= Case studies, field, natural, covert/overt
Why do scientists use standardised procedures
So others can retest the same study again
What approaches are falsifiable and which ones aren’t
Biological and behaviourist are falsifiable
Humanistic and psychodynamic are unfalsifiable (they use case studies)
What does popper argue about falsification
That no amount of positive validations of a theory prove it to be true. However, one piece of falsification evidence is enough to render a theory untrue