Research methods Flashcards
An aim is
A general statement of the purpose
Dependent variable
Measured
Independent variable
Manipulated/changing
Hypotheses
A prediction
Directional hypotheses
Uses previous research, clear difference
Non directional hypothesis
No previous research
Example of non directional hypothesis
There will be a difference in (dv) between (group A) and (group B)
Example of directional hypothesis
(Group A) will have a higher/lower (dv) in comparison to (group B)
Operationalisation
Ensuring variables are in a form where they can be easily tested
Extraneous variables
Unwanted variables. Any other variables that isnt the IV that could effect the DV.
Common extraneous variables - (participant and situational)
Participant - age, IQ, personality
Situational - time, weather
Demand characteristics
Change in behaviour to fit the experiment
Investigator effect
Change in investigators behaviour (unconscious or consciously) towards participants making them give a desired answer/result
How do you overcome demand characteristics/extraneous variables and investigator effect
Randomisation and standardisation
Randomisation
Equal chance, eliminates error, controls bias
Methods of randomisation
Dice roll, random name/number generator, names out of a hat, coin flip
Standardisation
Every participant receives the same instructions = same experience
Repeated measures
One group, take part in all conditions
Advantages of repeated measures
No individual difference, no personal errors
Disadvantage of repeated measures
Participants may figure out the experiment and display demand characteristics = unreliable results
Match pairs design
Pairs are matched on a variable
Disadvantages of matched pairs
More time, effort and money
Independent group design
One group for each condition
Advantages of matched pairs
More variation, limits order effect/demand characteristics
Advantages of independent group design
Less likely to work out the aims, order effects/demand characteristics aren’t a problem
Disadvantage of independent group design
More people = more money, individual differences
How to control order effects
Counterbalancing
Counterbalancing
Half participants participate in condition A then B, half participate in B then A to attempt to control order effects
Order effects
The outcome may differ due to the order of conditions
Laboratory experiment
Carried out in a controlled setting, high internal validity, good control over all variables
Advantages of laboratory experiments
It’s well controlled over extraneous variables as a safe and secure environment
Disadvantages of laboratory experiments
Demand characteristics = give less accurate results and lack generalisation, as people can tell
Field experiment
Natural environment, people don’t know its an experiment, high external validity, low internal validity, covert
Advantages of a field experiment
Behaviour is more natural - more mundane realism, less likely to show demand characteristics = accurate, reliable, genuine
Disadvantage of a field experiment
Can’t control all IV
Covert
People don’t know they are being observed
Overt
You know your being observed
Natural experiment
IV is natural occurring, would be unethical/impractical to change it
Advantages of natural experiment
Provides opportunity for experiments that wouldn’t have normally taken place and study real-life issues and problems
Disadvantages of natural experiment
Unethical to change the IV so research is limited. Research is targeted at a specific group.
Quasi experiment
You can’t manipulate the IV, it simply exists (sex, age, personality)
Advantages of a quasi experiment
Carried out in well controlled conditions, share the same strengths as lab experiments
Disadvantage of a quasi experiment
Targets specific people, unable to control the IV.
Internal validity
We can be sure change in DV is due to IV
External validity
Results are genuine and can be applied to others
Random sampling
Members of the population have equal chance of being selected
Disadvantage of random sampling
Difficult and time-consuming, selected participants may be unwilling to participate
4 steps of random sampling
- Get target population
- Number participants
- Random number generator
- All have equal chance
Advantage of random sampling
No bias = valid, reliable results
Systematic sampling
Every nth member is selected, can use a random generator to select the person to start with
Advantage of systematic sampling
No bias, representative, easy, simple, accurate
Disadvantage of systematic sampling
Work out the experiment due to grouping
Stratified sampling
Sub-groups according to frequency in population. People are then selected randomly from the sub-groups
Advantages of stratified sampling
No bias, representative, reflects different groups, results can be generalised
Disadvantages of stratified sampling
Doesn’t reflect all people differences; so full representation of population isn’t possible
Opportunity sampling
Researchers ask random people who are available
Advantages of opportunity sampling
Convenient - saves time, money and effort
Disadvantages of opportunity sampling
May be bias in who the researcher picks, unrepresentative of specific people/areas (not generalised)
Volunteer sampling
Participants select themselves and volunteer
Advantages of volunteer sampling
Easy, no effort, less time consuming
Disadvantages of volunteer sampling
Study may attract a specific group/type of people so bias/not generalised
How could you promote volunteer sampling
Adverts, posters
Informed consent
Information given, concerning the nature and purpose of the experiment and their role
How do people get consent for experiments
A consent form
Deception
Deliberately withholding/misleading information. Not told true aims.
Overcome deception
Debrief after (experiment could cause distress)
Protection from harm
Should not experience physical or psychological harm (injury, lowered self-esteem or embarrassment)
Overcome protection from harm
Right to withdraw, offer therapy/reassurance/counselling
Confidentiality
Data is not shared
Privacy
Person has control over the flow of information about themselves
Achieve privacy and confidentiality
Remain anonymous, password protected/burnt/destroyed after test
Naturalistic observation
Setting where behaviour would naturally occur
Observations
Not manipulate any variable just watching behaviour
Controlled observation
Lab/controlled setting, variables can be managed
Covert observations
Public setting without participant consent
Overt observation
Recorded with participant knowledge
Participant observation
Researcher becomes a participant
Non - participant observation
Researcher remains separate from experiment
Advantage of naturalistic observations
High external validity and behaviour can be generalised
Disadvantages of naturalistic observations
Hard to replicate as lack of control, may be extraneous variables
Advantage of controlled observations
Less extraneous variables and less easy to replicate
Advantage of covert
No demand characteristics as natural and increase validity
Advantage of overt
Ethically acceptable
Advantage of participant observations
Increase validity and reliability as you get an insight into real lives
Advantage of non - participant observations
Maintains objective from a distance, less danger of going native
Disadvantage of controlled observations
Findings cannot be generalised, low external validity
Disadvantage of covert observations
Ethical issues - people don’t know they are being observed
Disadvantage of overt observations
Demand characteristics = unreliable data
Disadvantage of participant observations
Researchers loses the objective - line between researcher and participant is blurred
Disadvantage of non - participant observations
May miss valuable information as at a far distance
Unstructured observations
Continuously recording everything, good for small scale
Structured observations
Use a behavioural checklist (behavioural categories)
Behavioural categories
Target behaviour that is easily observed and measurable
Event sampling
Record every time a specific behaviour/event is done
Time sampling
Records at a fixed time (e.g every 15 seconds)
Inter - observer reliability
2 observers carry out observations and agree to avoid bias and objective data
4 steps to carry out inter - reliability
- Familiarise with categories
- Observe behaviour
- Compare data
- Analyse data
What percentage is high inter - observer reliability
80%
Advantage of structured observations
Systematic easy to analyse and easy to record
Advantage of unstructured observations
Qualitative data - rich in detail
Advantage of behavioural categories
Structured and objective data
Advantage of event sampling
Target behaviour that could occur infrequently/could be missed
Advantage of time sampling
Reduces number of observations, simpler
Disadvantage of structured observations
Lack detail
Disadvantage of unstructured observations
Difficult to analyse and record
Disadvantage of behavioural categories
Categories should not overlap and should not be dustbin categories
Disadvantage of event sampling
May overlook important details if behaviour is too complex
Disadvantage of time sampling
Unrepresentative of whole observation
Dustbin categories in behavioural categories
Categories that are too vague so lots fall under
Questionare
List of written questions
Open questions
Any answers
Closed questions
Fixed choice
Interview
Face to face or on the phone
Structured interview
Pre-determined set of questions in a set order
Unstructured interview
Conversation like/discuss
Semi-structured interview
Set questions but follow up questions as well
Qualitative
Language/words
Quantitative
Numbers
Primary data
First hand by researcher
Secondary data
Collected by someone else
Advantage of qualitative data
Details, greater external validity, wide views
Advantage of quantitative data
Simple, draw conclusions, less bias, objective
Advantage of primary data
Authentic, reliable
Advantage of secondary data
Inexpensive, easy to get = save time and effort
Advantage of structured interviews
Easy to replicate
Advantage of unstructured interviews
Gain a better insight
Advantage of questionnaires
Cost effective (large amounts of data) quickly, easy to analyse
Disadvantage of primary data
Time consuming, effort
Disadvantage of secondary data
Out of date, Variation in quality
Disadvantage of qualitative data
Difficult to analyse, subjective interpretation = bias
Disadvantage of quantitative data
Fail to represent real life, narrow scope
Disadvantage of structured interview
Lack of details, can’t go off topic
Disadvantage of unstructured interview
Sift through data - difficult to analyse and time consuming
Disadvantage of questionnaires
Design characteristics, response bias (all people who answer may be similar to each other)
Correlation
An association between two variables, method used to analyse data
Pilot study
Small scale trial run
In positive correlation, variables go
Both go up/down
In negative correlation, variables go
1 goes up other goes down
What does correlation doesn’t cause causation mean
There could be a third variable which causes one of the variables to rise/fall
Advantages of correlation (3)
- Useful for looking at trends
- Easy to replicate (confirm findings)
- Quick economical to find
Disadvantage of correlation (2)
- Can’t find a conclusion as cannot demonstrate cause and effect between variables
- May be another variable (intervening variable)
Mean
Add all, divide by number of values
Median
Middle
Mode
Common
Range
Difference between smallest and largest
Standard deviation
How much data deviates from the mean
Measure of dispersion
Range, standard deviation
Measure of central tendency
Mean, median, mode
Advantages of mode
Easy to calculate, good for all data (nominal data) not just numbers
Advantages of mean
Includes all values, representative
Advantages of median
Not effected by extreme values, easy to calculate
Advantages of range
Easy to calculate
Advantages of standard deviation
More representative, include all values
Disadvantage of mode
Not representative of all data
Disadvantage of mean
Easily distorted by extreme values
Disadvantage of median
Not all scores are included
Disadvantage of range
Use extreme values, unrepresentative of whole data
Disadvantage of standard deviation
Can be distorted by extreme values
Dispersion
How spreed out the data is
A table represents
Mean and standard deviation
Bar charts represent
Mean, discrete/categorical data
Scatter-grams represent
Correlation/relationships. Can’t come to a cause and effect (variables aren’t linked)
Histograms represnt
Continuous data, bars touch
Line graphs represent
Continuous data
Normal distribution on a graph looks like
Same/similar mean mode and median in middle, bell shape, symmetrical
Positive skewed distribution on a graph
Peak towards left (most people on left) , mode —> median —> mean, hard test
Negative skewed distribution on a graph
Peak towards the right (most people on right), mean –> median, —> mode, easy test
Statistical testing
Determines if hypothesis should be accepted or rejected
Sign test defornition
A statistical test to analyse the difference in scores between related items
What is the sign test
Calculated value must be < (less than or equal to) the critical value for it to be seen as significant
Where do you get the critical value from
Critical value table
How to get the critical value from table
- Determine hypothesis (directional - 1 tailed, non - 2 tailed)
- Significance level (always 0.05/5% unless told otherwise)
- Number of participants (N value)
How to calculate the calculated value
SMALLEST number of +’s or -‘s
What to do if there are = in the calculated value
Remove them from the number of participants
Peer review
The assessment of scientific work by other specialists to make sure work is high quality
3 aims of peer review
- Quality and accuracy
- Suggest amendments and improvements
- Allocate research funding
Positives of peer review
Validity and accuracy of research
Negatives of peer review
Anonymity - may allow rivals to criticizes research who find research a threat/competition as ‘peer’ is anonymous
Publication bias - only publishing things that will create hype (may be controversial)
Research done on parents was
Both parents form an equal bond with the child which promotes flexible working arrangements (mothers don’t need to take traditional roles)
How does both parents forming bonds benefit the economy
Both parents can earn income and contribute effectively to society and the economy
Why does the treatment for mental illnesses benefit the economy
Treatments mean less people are off work
How much does absence from work cost the economy a year
£15 billion
What fraction of absences from work is to do with mental illness
1/3
