Research Methods Flashcards
What is reliability?
How consistent findings from a controlled investigation are within itself and overtime
What is validity?
Extent to which a study is consistent in measuring what it claims to be, and if the findings can be generalised beyond the setting of study
What are the 2 types of internal validity?
Face validity and concurrent validity
What is face validity?
Extent to which a study appears to measure what it claims to
What is concurrent validity?
Extent to which test produces same results as an established measure
What are the 3 types of external validity?
Temporal validity, ecological validity, population validity
What is temporal validity?
Study generalised across time
What is ecological validity?
Study generalised to other settings or conditions
What is population validity?
Study generalised to other groups of people
What is the word associated with reliability?
Consistancy
What is internal reliability?
How consistent different parts of the test are within itself
What is external reliability?
How consistent a study is when replicated
How is external reliability assessed?
Test-retest reliability or inter-observer reliability
How does test-retest reliability work?
Test is given to participants and after a short interval same/similar test is given again
What should happen for something to have inter-observer reliability
Consistent results between observers/raters
How does inter-observer reliability work?
Observers agree on categories and independently observe/rate people at the same time
To have high reliability/validity, what must comparing date have?
A correlation coefficient of +0.8 or higher when tested for significance using Pearson’s R or Spearman’s Rho at the 5% level
What are ethical issues, and name the 4?
Conflict between participants’ rights and researchers’ need to gain valuable, meaningful and valid findings
Informed consent, deception, protection from harm, privacy/confidentiality
What is deception?
Deliberately misleading/withholding information
How are privacy and confidentiality dealt with?
Privacy- the participant has a right to control information they release
Confidentiality- data protection act, anonymity
What is the British Psychological Society (BPS) code of ethics and what do they do?
Quasi-legal document psychologists have a professional duty to observe. A cost-benefit analysis is done, with the cost being Ψ reputation ruined or participant damage and the benefit being valuable research.
If under 16, can consent be given?
Yes, by a parent
What is done to deal with informed consent?
A consent form
How is a consent form written?
In Verbaitim…
Dear Participant…
-Aim + procedure
-Can withdraw
-Can withdraw data
-Data anonymous
-Ask any questions
-Debrief at the end
Circle YES/NO
Name____
Sign____
Date____
What are 3 alternate ways of consent?
Presumptive consent- similar group consent, so consent of this group is presumed
Prior general consent- participants consent for a number of studies in one
Retrospective consent- consent in debreif
How is deception and protection from physical/psychological harm dealt with?
-full debrief with true aims, all info and right to withhold data
-reassure normal behaviour
-offer provided counselling
What is an aim, and how can one be written?
A general statement stating what the researcher intends to investigate- the purpose of the study
To investigate…
What is a hypothesis?
Stated at the start of a study, it’s a clear, precise and testable statement stating the relationship between the variables to be investigated
What are the 2 types of hypotheses?
Alternative
Null
What are the 2 types of alternative hypotheses?
Directional/one-tailed
Non-directional/two-tailed
What is a directional/one-tailed hypothesis, and when is one used?
Predicts specific outcome
Previous research done or a replication of research
What is a non-directional/two-tailed hypothesis, and when is one used?
There will be a difference… indicates no direction
No research done before or contradictory research
What is a null hypothesis?
Hypothesis that the research hypothesis is tested against, states that the researcher won’t find what they expect
What does operationalisation mean?
Clearly defining variables in terms of how they can be measured
Give 3 ways in which you can operationalise variables
Counting
Rating scales
Time period
What are extraneous variables?
All other variables/features of the research other than the IV/DV/variables studied that have the potential to affect the DV if not controlled. Doesn’t vary systematically with the IV.
What are confounding variables?
An EV that varies systematically with the IV, and affects the DV to the extent where the researcher is unsure if the IV is the true source of change to the DV.
What are the 5 types of confounding variables?
Participant variables
participant effects
Situational variables
Demand characteristics
Investigator effects
What are participant variables?
Individual differences in participants eg: age memory
Give 2 ways participant variables be controlled
Repeated measures design
Random allocation
What are participant effects?
Participant reactivity
Social desirability bias
Hawthorne effect (change due to attention)
Give 2 ways participant effects can be controlled
Single-blind trial
Experimental realism
What are situational variables?
Features of the research environment eg: time
Give 2 ways situational variables can be controlled
Pilot study
Standardised procedure
What are demand characteristics?
Cues that tell the participant what is expected of them/what the investigator hopes to find
Give 4 ways demand characteristics can be controlled
Pilot study
Standardisation
Experimental realism
Matched pairs design
What are investigator effects?
(Un)conscious cues from the investigator that impact the outcome of the study
What are direct investigator effects, and how can these be controlled (2)?
Unconscious cues given during interaction
Double-blind trial
Standardisation
What are indirect investigator effects, and how can these be controlled (3)?
Bias in the research design
Pilot study
Counter-balancing if repeated measure design
Random allocation
What is a pilot study?
A small-scale trial run before an investigation with smaller + similar participants
What is the purpose of a pilot study?
Improves validity + reliability and saves time/money
What do pilot studies do?
Check the research is clear and controlled, any potential issues are identified
Change thinks that don’t work, modify designs/procedures
What is standardisation?
Keeping everything the same for all participants in a condition so the investigation is fair and validity is increased
What are standardised procedures?
Exercising control by ensuring all participants have the same experience
What are standardised instructions?
Exercising control by ensuring all participants have the same instructions in the same way
What should stay the same when using standardised procedures?
Place
Equipment
Materials
Conditions
Time
Instructions
How are standardised instructions written?
Procedures explained
Can ask any questions
Written verbatim
What is the schema for the experimental method?
Difference
Comparison groups
Causal
Bar graph
What is the experimental method?
The only method to make causal conclusions as the IV is directly manipulated to measure the effect on the DV
What is the independent variable (IV)?
The variable manipulated/changes naturally and the direct effect on the DV is measured
What is the dependent variable (DV)?
The variable affected by the IV and is measured, the change should only be caused by the IV
Why do levels of the IV/experimental comparison groups exist?
A comparison needs to be made to determine the effects of the IV on the DV
What are the 2 different types of experimental comparison groups?
Comparison of 2 experimental conditions
Comparison to a control condition
How do you write a directional experimental hypothesis?
IV → DV → IV
How do you write a non-directional experimental hypothesis?
There will be a difference in DV between IV + IV
What are the 4 types of experiments?
Lab
Field
Natural
Quasi
What are the 2 types of true experiments
Lab
Field
What is a lab experiment?
In a highly controlled environment
The IV is directly manipulated to produce a change in the DV
Participants are randomly allocated
What are 3 strengths of a lab experiment?
High internal validity (variable control)
Replication possible increasing validity + reliability
Strong causal relationship
What are 3 weaknesses of a lab experiment?
Low ecological validity + mundane realism
Demand characteristics
Investigator effects
What is a field experiment?
A controlled experiment in a natural/ordinary environment
The IV is directly manipulated to produce a change in the DV
Participants are usually unaware
What are 3 strengths of a field experiment?
High external validity + mundane realism
Reduced demand characteristics as unaware
Casual relationship
What are 3 weaknesses of a field experiment?
Low internal validity + reliability (hard to control EVs + CVs)
Ethical issues (consent + privacy)
Participant variables
What is a natural experiment?
The IV is naturally occurring and the researcher takes advantage of this, will vary
Effect on the DV observed
Tentative causal relationship
What are 3 strengths of a natural experiment?
Allows for research which can’t be done ethically
High external validity
Low demand characteristics
What are 2 weaknesses of a natural experiment?
Low internal validity (no RA)
Low reliability (replication)
What is a quasi-experiment?
The IV conditions already exist in participants
Often controlled in a laboratory
What are 2 strengths of a quasi-experiment?
High internal validity (if controlled)
Can compare when impossible to manipulate variables
What are 2 weaknesses of a quasi-experiment
Participant variables lower internal validity (no RA)
Demand characteristics (if unnatural)
What are the 3 designs for experiments?
Independent groups
Repeated measures
Matched pairs
What is an independent groups design?
Different participants used in each condition
Results are compared
Random allocation used
What are 3 strengths of using an independent groups design?
Order effects avoided
Demand characteristics lowered
Same test used
Although also less economical, what is the main disadvantage of using an independent groups design? How can this be controlled?
Participant variables act as EVs
-random allocation of participants
-matched pairs design
What is a repeated measures design?
All participants experience all conditions with a suitable time interval inbetween
What are 2 strengths of using a repeated measures design?
Participant variables controlled
Fewer participants needed so less time recruiting, more economical
What are 3 disadvantages of using a repeated measures design? How can these be controlled?
Order effects (practice, boredom, fatigue)
-counter-balancing distributes effect
-half do condition 1 then 2, half do condition 2 then 1
Demand characteristics (more opportunity to guess aim)
-single-blind technique
-experimental realism
Different tests used
-random allocation of questions
What is a matched pairs design?
Different but similar participants are used paired on a key variable
How would you match pairs in a matched pairs design?
- Identify suitable variable
- Variable is measured in participants eg: IQ test
- Paired on similarity
- Each participant in pair randomly allocated to condition
What are the strengths of using a matched pairs design?
Order effects avoided
Demand characteristics reduced
Same test can be used
Participant variables controlled
What are 2 disadvantages of using a matched pairs design?
Participant variables (matching not perfect)
-control with pilot study to make sure all PV are considered
Time consuming + expensive as more participants required
How should validity be improved in an experiment?
Control group
Single/double-blind to reduce participant reactivity/investigator effects
Standardisation (also for reliability)
What do correlations do?
Illustrate the strength and direction of the relationship between 2 or more co-variables
What does strength mean in terms of correlations?
Between -1 and 1
Perfect, high, low, no correlation
What does direction mean in terms of correlations?
Positive/negative
What is a correlation coefficient?
A number indicating the strength and direction of the relationship between the co-variables
What are the 4 types of correlations?
Positive
Negative
No correlation
Curvi-linear
What are the 2 reasons as to why a correlation may be used?
Variables cannot be manipulated (eg: for ethical reasons)
Want to investigate relationship, not difference, between variables (maybe before designing an experiment)
What are the 4 reasons why a cause-and-effect relationship cannot be established with a correlation?
- Not an experiment, no IV is manipulated
- No causal conclusions
- Direction of the influence between the co-variables unknown
- May be a third factor affecting the co-variables
How do you write a directional hypothesis for a correlation?
There will be a +/- correlation between … and …
How do you write a non-directional hypothesis for a correlation?
There will be a relationship between … and …
What are 2 strengths of a correlational analysis?
Useful starting point
-assessing strength and direction provides precise quantifiable measure of relation
If it’s strong, may suggest hypothesis for future
Quick + economical
-no control/manipulation
-secondary data
Less time consuming
What are 2 weaknesses of a correlational analysis?
May have flawed methodology
-method used to measure the variables may be flawed
Lowers validity
Cannot demonstrate cause and effect
-none manipulated
-not sure which variable caused change
-may be third untested variable
Base a research design on false, misinterpreted information
What is an observation?
Recording observable behaviour to study complex interactions, often used as a technique within the experimental method
What are the 6 types of observations?
Naturalistic or controlled
Covert or overt
Participant or non-participant
What is a naturalistic observation?
Takes place in the setting the target behaviour would usually occur
All aspects of the environment are free to vary and not interfered
What is a strength and a weakness of a naturalistic observation?
High external validity
Lack of control (low internal validity + reliability)
What is a controlled observation?
Controlled, specific and measureable
ALl aspects of the environment not free to vary
Control over EVs via manipulation
What are 2 strengths and 2 weaknesses of a controlled observation?
Can replicate to improve reliability
High control improves validity
Low ecological validity
Demand characteristics
What is a covert observation?
Participants are unaware they’re being observed
What is a strength and weakness of a covert observation?
Reduced demand characteristics
Ethical issues as may not want to be observed
What is an overt observation?
Participants know they’re being observed and have given consent
What is a strength and weakness of an overt observation?
More ethically acceptable
Increased demand characteristics (decreased accuracy)
What is a participant observation?
The observer becomes part of the group, overtly or covertly
What is a strength and 2 weaknesses of a participant observation?
Increased insight and understanding improves validity
Lack of objectivity as may identify with the group
May influence and decrease natural behaviour
What is a non-participant observation?
The researcher is seperate and records behaviour in a more objective manner
What is a strength and a weakness of a non-participant observation?
More objectivity due to psychological distance, increasing accuracy
Lack of insight
What are behavioural categories?
When a target behaviour is broken into operationalised components via coding
What 5 things should behavioural categories be?
Visible
Objective
Self-evident
Mutually exclusive
Cover all behaviours
What do behavioural categories/behavioural checklists increase?
Reliability
What are the 2 ways to record observational data?
Structured observation
Unstructured observation
What are structured observations?
Target behaviour tallied using pre-determined checklist from a pilot study
What are unstructured observations?
Everything is recorded with no pre-set criteria, usually small scale
What is a strength and weakness of using a structured observation?
Recording is easy + systematic, quantifies information to analyse and compare
If not all behaviour is covered in the checklist, important behaviours ignored reducing accuracy
What is a strength and 2 weaknesses of using an unstructured observation?
Data is rich in depth + detail improving accuracy
Observer bias
Difficult to record + analyse qualitative data
What are the 3 sampling methods for observations?
Continuous recording
Event sampling
Time sampling
What is continuous recording?
Key feature of unstructured observation, all behaviour is recorded although not practical for complex behaviours
What is event sampling?
Recording each time a target behaviour occurs
Useful for complex behaviours, but if the behaviour is too complex, details may be overlooked
What is time sampling?
Recording in a pre-established time frame
Reduces the number of observations made but may be unreprasentitive
How can ethics be ensured during an observation?
Informed consent- usually resumptive or retrospective
Privacy- confidentiality + anonymity ensured
How can validity be improved for observations?
Behavioural categories (via pilot study)
Naturalistic/covert to reduce demand characteristics
Various people for population validity
How can reliability be improved for observations?
Train observers and allow to practise
Operationalised behavioural categories that cover everything and don’t overlap
What is a meta-analysis?
The process of combining results from a number of studies (secondary data) involving the same aims/hypotheses and methods of research. Results are then reviewed together
How does a meta-analysis work?
Data is statistically tested to assess effect size
Provides a joint, overall conclusion
What is effect size?
A measure of the strength of relationship/difference between 2 variables
The bigger the number, stronger effect
What are 2 strengths of a meta-analysis?
Reviews lots of data to be more confident in conclusions
Larger, varied and more representative sample so more generalisable and increased population validity
What is a weakness of a meta-analysis?
Publication bias
-not all relevant studies selected
-negative/insignificant left out
Only represents some data in the analysis, incorrect + biased conclusions
What is a content analysis? (speaking generally)
AN observational research technique where behaviour is studied indirectly be analysing artefacts/communications
What is the aim of a content analysis?
Summarising + describing artefacts/communications systematically to identify patterns/themes and draw conclusions
What is a content analysis?
When qualitative data is categorised and the frequency is counter by producing quantitative data
How is a content analysis carried out?
Categorise- view the artefact multiple times and code (identify relevant categories) give an eg of a category
Count- re-view artefact and tally the number of examples in each category to produce quantitative data
Compare- summarise in descriptive statistics to draw conclusions
What is a thematic analysis?
Used to keep participant perspectives, qualitative data is analysed, summarised and described in themes
How is a thematic analysis carried out?
Code via familiarisation- carefully view to look for behaviour/words/images repeatedly occurring
Reduce this to 3/4 theme, which are ideas repeatedly occurring
Re-view, and provide egs of behaviour using quotes and descriptions to draw conclusions
What are 3 strengths of a content analysis?
Non-invasive, reducing ethical issues, as data is in the public domain
High ecological validity _ internal validity as genuine communications
Replicable, reliability can be tested
What are 3 weaknesses of a content analysis?
Lack of causality
Outside of the context, observer bias may occur and reduce objectivity
Culture bias, interpretation is the language/culture of the observer
In a content analysis, how can objectivity be improved to increase validity?
Double-blind technique
In a content analysis, how can culture bias be reduced to increase validity?
Triangulation, compare results from a variety of different studies about the same artefact
How can reliability be ensured in a content analysis?
Coding system must be clear, concise and easy to use
What is a case study?
A specific, detailed, in-depth investigation of an individual, group, institution or event carried out in the real world
How is a case study done in 3 steps?
Info is gathered from a variety of sources using many different research techniques, mostly qualitative
A case history is constructed by combing and organising findings
Usually longitudinal, may also involve data from others eg: family
What are 2 strengths of a case study?
Only possible method for rare events (eg: HM) or sensitive events (eg: Romanian orphans)
Rich detailed data, info isn’t overlooked, greater understanding of human behaviour (this can generate hypotheses and lead to revising contradictory theories), high ecological validity
What are 4 weaknesses of a case study?
Difficult to generalise- unique, low sample size, reduces external validity
Researcher bias- subjective selection of info on report, may lose objectivity with connection
Lack of causality- only after a key event, can’t be sure of changes
Reliability- hard to recall past events, can’t be replicated
How can the validity of case studies be improved?
Interpretive validity- use direct quotes
Triangulation- variety of sources for coherence
What are self-report techniques?
Participants giving information to the researcher about their own feelings, opinions and behaviour by responding to questions
What are questionnaires?
Pre-set list of written questions/items participants that people respond to
What are open questions, and what type of data do they produce?
Have no fixed range of answers so participants can answer how they wish
Qualitative
What are closed questions, and what type of data do they produce?
Have a fixed number of responses
Quantitative
What is a pro and con of qualitative data for open questions?
Rich in depth, best for attitudinal research
Difficult to collate, summarise, analyse
What is a pro and con of quantitative data for closed questions?
Easy to analyse, compare and test statistically
Lacks depth + detail, can’t explain why
What are 2 strengths of questionnaires?
Large amounts of data to a large number of people, cost-effective + representative so increases external validity
Easy to administer + replicate, can establish reliability
What are 2 weaknesses of questionnaires?
May not be truthful due to social desirability bias reducing validity
Response bias/acquiescence bias as answer too quickly or don’t read, inaccurate
What are interviews?
Face-to-face/over-the-phone interaction between an interviewer and interviewee
What are the 3 types of interviews?
Structured
Semi-structured
Unstructured
What are structured interviews?
Pre-determined set of questions asked in a fixed order in real time
What are semi-structured interviews?
In between other types, most used in everyday life, for example at a job interview
There is a pre-set list of questions but the interviewer can ask follow-up questions
What are unstructured interviews?
Like a conversation, with no set questions, just a general aim and topic discussed
Free-flowing and the interviewee is encouraged to elaborate
What is a strength and weakness of structured interviews?
Standardised, easy to replicate with decreased differences between interviewers
Can’t deviate from the questions, limiting the richness of the data
What is a strength and weakness of unstructured interviews?
Flexible so can follow-up and gain greater insight
Lots of information so difficult to analyse the relevant information and draw firm conclusions
How may social desirability bias be increased or decreased in interviews? (compared to questionnaires)
Not anonymous, so risk of untruthfulness to present themselves in the best light
Can establish a rapport with the interviewer so sensitive/personal information is answered trruthfully
How can validity be improved in interviews and questionnaires in 6 ways?
Use relevant questions
Pilot study
Lie scales
Anonymity
Representative samples
Distracting questions
How can reliability be improved in interviews in 4 ways?
Same interviewer
Train interviewers to prevent asking leading questions
Standardised
Use structured interview to decrease bias and make the interview replicable
How can reliability be improved in interviews in 3 ways?
Deselect unclear questions
Rewrite ambiguous questions
Use closed questions
How do you design an open question?
Explain…
Tell…
Describe…
How do you design a closed question?
Likert scale
Rating scale
Fixed choice option
What does a likert scale question look like?
Rate strength of (dis)agreement
Strongly agree
Agree
Neutral
Disagree
Strongly disagree
What does a rating scale question look like?
Value representing the strength of feelings
Very fun
1
2
3
4
5
Not fun at all
What does a fixed-choice option question look like?
Choose from a list of what applies
▢
▢
▢
▢ N/A
▢ Other
How do you design an interview well?
Standardised interview schedule (question list)
Take notes/record
Group/1to1 in calm, relaxed place
Start with neutral questions
Ensure confidentiality
What should NOT be used when designing self-report questions?
- Jargon
- Emotive language
- Leading questions
- Double-barrelled questions
- Double-negative questions
No overlapping boxes in fixed choice questions
What are the 4 types of data?
Primary or secondary
Quantitative or qualitative
What is primary data?
Original data collected directly, first-hand by the researcher from the participants specifically for the purpose of the study
What are 2 strengths and a weakness of using primary data?
High control over quality/quantity of data, improving validity + reliability
Designed specifically to fit the aims and hypotheses of the study
Time consuming + expensive to recruit participants, conduct the study and analyse data
What is secondary data?
Data collected for a different purpose by another person, exists before the study but used to test the aims and hypothesis of the study?
What are 2 strengths and 2 weaknesses of using secondary data?
Simpler + cheaper as less time + equipment needed
May have had statistical testing so the significance is shown
May not fit the needs of the study
May be inaccurate (eg: outdated)
What is quantitative data?
Measures numberically
More objective/scientific measure
Easier to analyse
Superficial data
Easily replicated
What is qualitative data?
Non-numberical + descriptive
Open to subjective interpretation
Harder to analyse (done via content)
Rich, detailed data
Not easily replicated/generalised
What are the 3 levels of measurement that quantitative data can be categorised into?
Nominal
Ordinal
Interval
What is nominal data?
Placed into discreet categories and the frequency is counted using a contingency table used
What is the measure of central tendency, measure of dispersion, graph, and example for nominal data?
Mode
N/A
Bar/pie
Favourite food
What is ordinal data?
Data ordered/ranked from high to low
Non-scientific/non-standardised as the difference is unknown and not fixed
What is the measure of central tendency, measure of dispersion, graph, and example for ordinal data?
Median
Range
Bar/scatter
Satisfaction rating
What is interval data?
Based on a standardised numerical scale with fixed intervals
Scientific, technical public scales of measurement
What is the measure of central tendency, measure of dispersion, graph, and example for interval data?
Mean
Standard deviation
Histogram
Number of calories consumed
How do you convert interval data to ordinal data?
Give the same number of ranks as participants, rank from highest to lowest
Tied scores are given the average rank of the spaces occupied
How do you convert interval data to nominal data?
Draw a contingency table with a title
Tally data into appropriate categories
What is a measure of central tendency, and what are the 3?
An average best describing the data set
Mean
Median
Mode
What is the mean, and what type of data is it used for?
Average value when all values in the data set are added and divided by the total number of the values
Interval/ratio data, no extreme values
What is the median, and what type of data is it used for?
The middle value when values are placed in rank order
Ordinal/extreme values
What is the mode, and what type of data is it used for?
Value occuring most frequently, may be bi-modal or have no mode
Nominal
What is a strength and weakness of using the mean as a measure of central tendency?
Most sensitive measure as includes all values so more representative
As highly sensitive, extreme values distort the mean making it unrepresentative
What is a strength and weakness of using the median as a measure of central tendency?
Not distorted by extreme values so if there is there are EVs, the value is representative
Less sensitive measure, not all scores included, not representative
What is a strength and weakness of using the mode as a measure of central tendency?
Easy, not influenced by extreme values so representative, also only one that can be used for nominal data
Crude, communicates little about the data, not as useful if more than 1 mode
What is a measure of dispersion, and what are the 2?
How spread out/similar the data is
Range
Standard deviation
What is the range?
The difference between the highest and lowest value
What is standard deviation?
Average distance of data from the mean
What is a strength and weakness of using the range as a measure of dispersion?
Quick and easy to calculate so can get feel for dispersion
Distorted by extreme values and no indication of dispersion between the maximum and minimum value, unreprasentitive
What is a strength and weakness of using standard deviation as a measure of dispersion?
Most sensitive taking all scores into account so more accurate
Time-consuming, also distorted by extreme values so not a fair respresentation
How can conclusions be drawn from descriptive statistics?
The mean/median is the same/slightly/higher/lower
This suggests… group A is/no difference
HOWEVER/ALSO
The range/SD is smaller for…
This suggests… less variation/more consistent, more had similar results
What are the 3 types of data distributions?
Normal distribution
Positively skewed distribution
Negatively skewed distribution
What does a normal distribution look like?
Bell-shaped, no bias to left/right so symmetrical
Mean=median=mode at the highest peak
50% of data less than the mean and 50% more than the mean
What is an example of a human characteristic with a normal distribution?
IQ
What % of the population is within 1 standard deviation of the mean with a normal distribution?
68%
What % of the population is within 2 standard deviations of the mean with a normal distribution?
95%
What % of the population is within 3 standard deviations of the mean with a normal distribution?
99.7%
What are skewed distributions?
Scores not distributed symmetrically so the data clusters at one end
What does a positive distribution look like?
Concentrated to left, long tail on right
Mean highest value, mode lowest value, median inbetween
What does a negative distribution look like?
Concentrated to right, long tail on left
Mean lowest value, mode highest value, median inbetween
How do you fix a positively skewed distribution?
- recognise the test is too hard as there are too many low scores (floor effect)
- Adjust the difficulty by making easier more accessible questions
- There will be fewer low scores, achieving normal distribution
How do you fix a negatively skewed distribution?
- Recognise the test is too easy as too many high scores (ceiling effect)
- Adjust the difficulty by adding harder questions
- Fewer achieve a high score, achieving normal distribution
Why does the mean shift for skewed distributions?
The mean is affected by extreme values
What must graphs have? (AHA)
Appearance- must reflect graph accurately
Headings- relationship vs difference
Axes- labelled + operationalised, start at 0 (can us broken line), use same scale
What type of data are contingency tables used for, and what are they?
Discrete/nominal
Examines relationships, can compare
Data connected vertically horizontally, with MCT or MD, not raw stats
What type of data are scattergrams used for, and what are they?
Pairs of scores
Examines relationship (strength + direction) between co-variables, participant scores are plotted (not group)
What type of data are bar charts used for, and what are they?
Discrete/nominal
Shows difference, compares levels of the IV
Spaces, same-size bars
If comparative, key needed
What type of data are histograms used for, and what are they?
Continuous/interval (grouped)
Shows how data is distributed
Bars are couching as it is continuous
yaxis is frequency xaxis is groups
What are 3 other types of graphs not specifically listed in the specification?
Pie charts
Line graphs
Frequency polygons
What does ≈ mean?
Approximately equal to
What is the equation for %increase or decrease?
Difference
————— x100
Original
In the %change equation, how do you know which number the ‘original’ is?
If % increase- original is smaller number
If % decrease- original is bigger number
What is a population?
A group of people with specific characteristics who are the focus of the researchers interest
Often called the target population as they’re a subset of the general population
What is a sample?
A group of people selected from a target population in a research investigation, and are presumed to be representative of that population
What does a representative sample look like?
Group of people forming the sample accurately reflect the target population as they have a similar distribution of relevant characteristics
What is generalisation?
Extent to which findings/conclusions from an investigation can be broadly applied to the population
When is generalisation possible in sampling?
If the sample is representative of the target population
What is bias in sampling?
When certain groups are over/under-represented in a target population
Why is bias bad in sampling?
Limits the extent generalisations can be made to the target population
What are the 5 sampling techniques?
Random
Systematic
Stratified
Opportunity
Volunteer
What is a random sample?
Where every member of the target population has an equal change of being selected
How is a random sample obtained in 4 steps?
- Obtain a complete list of all in a target population
- Assign every name to a number
- Use hat/computer randomiser
- The people assigned to the first n numbers drawn form the sample
What are 2 strengths of using a random sample?
Unbias as EV and CV distributed evenly, increased internal validity
Free from researcher bias
What are 3 weaknesses of using a random sample?
Still possible for the randomised list to be unrepresentative
Difficult + time consuming to obtain list
Selected participants may refuse to participate, affecting how representative the sample is
What is a systematic sample?
Every nth member of the target population is selected
How is a systematic sample obtained in 3 steps?
- Sampling frame produced (organised list of all in a target population, randomly or alphabetically)
- Sampling system is chosen, random or deliberately (population/sample size)
- Researcher selectors every nth person from the sampling frame until the sample is complete
What are 2 strengths of using a systematic sample?
Unbiased if randomised
Free from researcher bias (if start randomly selected)
What are 3 weaknesses of using a systematic sample?
If a periodic trait exists in the population, or not randomised, not representative
Time-consuming
Selected participants may refuse to participate (affects how representative)
What is a stratified sample?
The composition of the sample reflects the proportions of people in certain subgroups (strata)
How is a stratified sample obtained in 3 steps?
- Identify the strata that make up the population
- Calculate the proportion for each stratum based on the proportion of the population
- Participants that make up each stratum selected randomly
What is a strength of using a stratified sample?
Most representative as proportionally reflected, so can be generalised
What are 3 weaknesses of using a stratified sample?
Cannot reflect all so not a complete representation
Time-consuming
Selected participants may refuse (affects how representative)
What is an opportunity sample?
People who are available, willing to participate and fits the researcher’s criteria
How is an opportunity sample obtained?
The researcher approaches people around the time of the study and asks them to participate
What are 2 strengths of using an opportunity sample?
Convinient, saving time + money
Must be used for natural experiments
What are 2 weaknesses of using an opportunity sample?
Unrepresentative as people drawn from a specific area
Researcher bias as they have complete control over participant selection
What is a volunteer sample?
Participants select themselves to form the sample (self-selection)
How is a volunteer sample obtained?
An advert is placed in a place the target population would see it (eg: newspaper, common room noticeboard) and volunteers respond to the advert
What are 2 strengths of a volunteer sample?
Convenient + time-effective
Engaging + willing participants who won’t sabotage the study
What are 2 weaknesses of a volunteer sample?
Volunteer bias attracts a certain profile of those who are keen, helpful, curious, cooperative- demand characteristics
Can be unrepresentative, reducing validity
How are reports written?
Conventionally + academically
Standardised with all relevant info
Why are reports written?
Refine understanding + theories of human behaviour
Allows replication
What is the order of a scientific report?
Title
Abstract
Introduction
Method
results
Discussion and conclusion
References
What is the purpose of a title in reports? What must it include?
Gives focus of study
Must include key variables
What is an abstract?
Short summary of the report with key details (aim + hypotheses, method, results, conclusion, implications)
Approx 150-200 words
What is the purpose of an abstract?
Allows reader to have a quick overview of the report and decide if they want to read it
What is the purpose of an introduction?
Provide rationale/context for research, reviews previous research + background on theories/concepts
Logical- starts broad and narrows down to aims + hypotheses with key variables
What is the purpose of the method?
Provides detailed description of what was done so other researchers can replicate
All needs to be explained + justified
What are the 5 subsections of the method?
Design with key variables
Sample, sampling method + target population, give sample demographics whilst not compromising anonymity
Apparatus/materials
Procedure, exactly how carried out beginning-end, verbatim record of all said to participants, control of EV + CV
Ethics- how were these addressed
What is the purpose of the result?
Give summary of key findings
What is reported in the results?
Descriptive statistics (tables/graphs/MCT/MD)
Inferential stats (accept/reject null)
Qualitative analysis
What is the purpose of the discussion/conclusion?
Interprets study + considers implications
What is done in the discussion/conclusion?
- Summarise/explain findings in verbal form in context of intro evidence/other research
- Limitations (validity) + how to be addressed in future study
- Wider implications (RW, contribution to knowledge base, future research)
What is the purpose of the references?
Gives full standerdised details of any course cited to…
avoid plagiarism,
give credit
allow access of sources
What is the format for a journal article?
Author, date, article title, journal title, volume (issue), pages
What is the format for a book?
Author, date, book title, place of publication, publisher
What is the format for a web page?
Source, title, date, weblink, date accessed
What is a peer review?
Process where research papers are independently scrutinised by experts/peers in same/similar field before publication
What do peers assess for in peer review?
Validity/accuracy
Significance
Originality/contribution to body of psychological knowledge
High quality
What do peers do in peer review?
Ensure objectivity + anonymity
Decides whether research is published/reviewed/rejected
What are 4 purposes of peer review?
Improves + validates quality/integrity of research
- ensures high standard, can suggest amendments/improvements
- increases probability of weakness/errors identifies and addressed
- research taken seriously and adds to body of scientific knowledge to frequently become mainstream thinking
Prevents dissemination (spread) of irrelevent/flawed/fraudulent findings
- as if entered public domain, integrity of field and psychology damaged
- loss of trust in scientific claims
Considers wider context/relevance/significance/originality
- ensures research has clear + accurate + relevant practical applications with no negative consequences
Allocates research funding
- evaluates aims/quality/value
- peer review panel helps government bodies/charities decide which research is likely worthwhile
What is a strength of peer review?
Establishes validity of research
- may lead to social policy change
- if not valid, may have negative consequences
What is a strength/weakness of peer review?
Anonymity leads to honesty and objectivity
BUT
Minority may criticise rival researcher as in competition for funding
Thus, some journals favour open reviewing
What are 2 weaknesses of peer review?
Publication bias
- tendency to publish +/significant findings as increases credibility and circulation of publications
- avoids publishing less significant/replications (file drawer phenomenon)
Creates misconception of facts –> biases reviews and meta-analyses
Status quo, findings may be suppressed if contradictory
- PR may be assessing acceptability not validity
SLows rate of change of the paradigm within the discipline
What are the 6 features of science?
Objectivity
Empirical method
Falsifiability
Replicability
Theory construction
Hypothesis testing
What is ‘objectivity’ as a feature of science?
All bias minimised and personal opinions eliminated
Keeping critical distance to not distort/influence the research process
Operationalisation, standardisation and control
What is ‘the empirical method’ as a feature of science?
Scientific process objectively gathering evidence via direct experimentation/observation
Prevents abstract theories leading to bad treatment
Is objectivity via the empirical method achievable in psychology?
Behaviourist/biological/cognitive approaches use direct experimentation + observation to develop + refine theories
Psychodynamic/humanistic approaches use subjective methods as believe behaviour is unique + different
Some things can’t be directly observed, have to use inference which is subject to bias
What is ‘falsifiability’ as a feature of science?
Theory is only scientific is possible to establish as false
Popper said not possible to confirm theory, can only disconfirm
Even if theory repeatedly tested, may not be true
So focus on disproving hypothesis (confirming null)
Theories gradually fine-tuned via continual hypothesis testing
Is falsifiability achievable in psychology?
Behaviourist/biological/cognitive approaches produce falsifiable hypotheses eg: Milgram
Psychodynamic/humanistic/evolutionary unfalsifiable as cannot undergo testing
What is ‘replicability’ as a feature of science?
Extent to which scientific procedures + findings can be repeated in different contexts + time
Important as increases reliability, which has a role in determining validity
Strengthens theory as less likely findings due to chance
Done via operationalisation/standardisation/controlling/peer review
Is replicability achievable in psychology?
Behaviourist/biological/cognitive approaches generate testable hypotheses with empirical method allowing for replication
But true replication almost impossible as humans are dynamic + influenced by many factors
What is ‘theory construction’ as a feature of science?
Process of developing explanations for the causes of behaviour by systematically + empirically collecting evidence and organising into coherent account
Popper’s hypothetico-deductive model states theories come first which generate falsifiable hypotheses, which are tested + fine-tuned via the empirical method
What is a theory?
Set of general laws to explain events/behaviour
What is ‘hypothesis testing’ as a feature of science?
Process of gathering evidence
A good theory should suggest multiple testable hypotheses
If there is no empirical evidence, theories modified
Has psychology developed theories allowing for hypothesis testing?
Many theories explain + can predict behaviour and propose hypotheses for empirical evidence, eg: Bickman
Some lack predictive power eg: psychodynamic
Humanistic has no theories as doesn’t want to predict + control behaviour as individuals do that
What is a paradigm?
Shared set of assumptions and methods agreed on within a discipline
Why is psychology a pre-science according to Kuhn?
Too much internal disagreement and conflicting approaches
What is a paradigm shift?
Result of a scientific revolution when there is a change in the dominant unifying theory of a discipline due to new/contradictory evidence
What are the paradigm shifts in psychology?
1870s- introspection
1900s- psychoanalysis
Beginning of 20th century- observable behaviour
1950s- humanistic
1960s- cognitive
1980s- biological
21st century- cognitive neuroscience
Is psychology a science?
According to Kuhn, no…
Lacks universal paradigm, many conflicting perspectives
Many methods of investigating human behaviour (eg: experiment, case study)
Maybe yes…
Universal agreement that psychology is the study of the mind and behaviour
Has undergone many paradigm shifts
What is the purpose of inferential statistics?
Allows researchers to make inferences about whether differences/relationships are significant (not due to chance) and thus can be applied to a target population
What do statistical tests decide?
If null is accepted/rejected
As hypotheses must be falsifiable, seek to falsify null
What is probability?
Degree of certainty the difference/relationship is real, or if it occurred by chance
What is the usual level of significance in psychology?
5%
p ≤ 0.05
Why is the 5% level of significance used in psychology?
95% certain results not by chance- can never be 100% sure as haven’t tested all members of target population under all circumstances: so this is an acceptable level of error
Strikes balance between risk of making Type I / Type II errors
When would the 1% level of significance be used?
Replication
Challenging a theory
Controversial
Socially sensitive
Life/death
What type of error is more likely to occur with the 1% significance level?
Type II
When would the 10% level of significance be used?
Unresearched area (with a non-directional hypothesis)
What type of error is more likely to occur with the 10% significance level?
Type I
What is a type I error, and why does it occur?
False positive
As level of significance too lenient, too optimistic so the null is rejected when should be accepted
Stating there is a real difference/relationship when there isn’t
What is a type II error, and why does it occur?
False negative
As level of significance too stringent, too pessimistic so null is accepted when should be rejected
Stating there is no real difference/relationship when there is
What are 3 ways to reduce Type I/II errors?
Use less stringent (reduces Type II) or less lenient (reduces Type I) level of significance
Use 5% significance level as strikes balance between risk of Type I/II errors
Increase sample
How do you choose what statistical test to use?
Note: must be interval AND normally distributed
What is the difference between a relationship and an association?
Relationship is linear
Association not linear
What are the 3 parametric tests? What does this mean?
Unrelated T-test
Related T-test
Pearson’s R
More sensitive in measuring significance as use mean and standard deviation
What are 3 things to note when using the unrelated T-test?
If the calculated value if negative, ignore sign when comparing to critical value
Uses df= (sum of ppts in both groups) - 2
Difference must be in direction predicted
What are 2 things to note when using the related T-test?
Uses df= number of participants - 1
Difference must be in direction predicted
What do parametric tests assume?
Participants drawn from normally distributed population
Homogeneity of variance (standard deviation)
What are 4 things to note when using the Mann-Whitney test?
Group with smallest ranked sum is used
2 N values
Data from both groups ranked as 1 set, lowest value = rank 1
Difference must be in direction predicted
What are 3 things to note when using the Wilcoxon test?
T = sum of ranks of less frequent sign
N, excluding participants with no difference
Rank/order the difference between 2 conditions
Difference must be in direction predicted
What are 4 things to note when using Spearman’s rho test?
If calcuated value negative, ignore sign when comparing to critical value
N
Data ranked separately, square of difference between ranks is summed
Sign of calculated value must be in same direction as hypothesis
What are 3 things to note when using the Peaarson’s r test?
If calcuated value negative, ignore sign when comparing to critical value
Df = number of participants - 2
Sign of calculated value must be in same direction as hypothesis
What are 2 things to note when using the Chi-Squared test?
Df = (number of rows -1) x (number of columns -1)
Results recorded in contingency table
What are the 5 steps for how to calculate the sign test?
- Convert data to nominal and record the sign of difference, keep consistent
- If no difference remove from sample (N)
- Count total number of +/-, total of less frequent sign = calculated/observed/S
- Compare to critical value table know: significance, one/two-tailed, N
- Calculated value must be ≤ critical value
How do you state a conclusion after an inferential test?
Calculated value of ___ is more/less/equal to critical value of ___
at a significance level of ___% for a one/two tailed test.
Therefore, the null hypothesis is accepted/rejected.
There is/is not a significant difference… *link to stem.