research methods Flashcards
Experimental method
manipulating the IV to have an effect on the DV
what is levels of the IV?
-testing the effect of the IV using different experimental levels i.e a control condition and an experimental condition
What is operationalising?
-ensuring that both the IV and DV are measurable and clear
what is an aim?
area of psychology that the experimenter is looking into
-developed from theories and similar research
-i.e to investigate whether…
what is a hypothesis?
-testable statement which clearly states the relationship between variables and is developed from past research
-must be operationalised
Directional hypothesis
-predicts a specific direction of the effect between the variables
-i.e there will be an increase/decrease in…
Non-directional hypothesis
-non specific relationship between the variables
-i.e there would be a difference between…
Null hypothesis
-no effect on the variables
Control of variables
-any variable interfering with the IV and DV should be removed or well controlled.
Extraneous variables
- a ‘nuisance’ variable which is not the IV but affects the DV
-do not vary systematically with the IV
-make results harder to detect and should be removed before study.
situational variables
-factors in the environment which impact the study
-i.e weather, time
-extraneous
participant variables
-individual differences between participants which affect the study
-i.e gender, age
-extraneous
Confounding variables
-does change systematically with the IV
-becomes difficult for the researcher to be sure of what impacted the DV
-turns into a second unintended IV
Demand characteristics
-ppt may guess the aim due to cues from the researcher or situation.
-extraneous variable
Social desirability
participants automatically try to respond in ways that make them seem likeable in a study, even if it means misrepresenting how they truly feel.
‘please-U’ effect
-ppt may act in a over-way they think is expected or wanted from the researcher to fulfil hypothesis
‘screw-U’ effect
-ppt may deliberately under-perform to sabotage the results of the study
Investigator effects
-when a researcher influences the outcome of any research they are conducting
conscious or unconscious
-i.e leading question, selection of ppts.
Order effects
when participants’ responses in the conditions are affected by the order of conditions to which they were exposed
Randomisation
-use of chance methods wherever possible to reduce the effect of bias from investigator effects
- minimises impact of confounding/extraneous variables.
Standardisation
all ppts should be subject to the same environment, info and experience
-i.e standardised instructions read to each ppt
-means that non-standardised changes in procedure do not act as extraneous variables
Experimental designs
Independent groups
-different ppts complete different levels of the IV/conditions.
-groups are randomly allocated to prevent bias
Advantages +Disadvantages of Independent groups
-demand characteristics are minimised.
-no order effects
HOWEVER
-extraneous variables i.e differences between ppts due to random allocation may impact results
-requires more ppts so less economical(time/money)
Repeated measures
-same ppts take part in all conditions/levels of the IV.
- produces related data, can compare how ppts did in both conditions
Advantages+ Disadvantages of Repeated measures
-needs less ppts so more economical
- higher validity as ppt variables will not affect results as much
HOWEVER
order effects:
-may create boredom or fatigue(important in skill-based tasks)
-practice effects
-first condition may have an effect on the other one
How can you control the order effects in a repeated measures design
Counterbalancing
- this uses an ABBA format
- Half the ppts complete condition A then B
- other half complete condition B then A
Matched pairs
-pairs of ppts are matched in terms of key variables i.e age
-one member of each pair is placed in two different conditions
-minimises ppt variables but does not eliminate them
-avoids order effects
-time consuming, expensive as may require pre-test
What is reliability?
-how consistent results are, whether research can be repeated with those results
What is validity?
-accurately measuring what you are claiming to measure
Internal validity
-whether the IV changed the DV without influence of other factors.
what affects internal validity?
particapant variables (demand charctertics, personality, age)
- Lack of control (order effects, investigator effects)
- situational variables (temp, room size)
- Researcher Bias (lack objectivity)
-face validity(claim to measure what its measuring)
What is external validity?
can you apply the findings to the public or day to day life (generalisability)
Ecological validity
can generalise to a different place or setting
mundane realism
is the task similar to what we would do in real life
population validity
ability to generalise the findings to the wider population
-affected by androcentric / gynocentric
Historical validity
Can you generalise to a different century, or decade
How would we improve our external validity?
Using field study, natural observations etc
Types of experiments
Lab experiments
-conducted under controlled environment for conditions/lab
strengths+limitations of lab experiments
-high internal validity= control of EVs and CVs mean cause and effect is certain
-high replicability because there is a standardised procedure
HOWEVER
-low external validity=cannot be generalised to un artificial contexts
-give rise to demand characteristics
-lack of mundane realism
-experimenter bias
field experiments
-IV is manipulated in the ppt’s more natural,everyday setting
strengths+limitations of field experiments
-High external validity which means you can generalise it to real life situations
-higher mundane realism as people show more naturalistic behaviours
HOWEVER
-loss of control of CVs and EVs —> difficult to establish cause and effect
-possible ethical issues about consent and privacy
natural experiments
-Change in IV has occurred naturally so not be manipulated by researchers
strengths+limitations of natural experiments
- high external validity (changes happened in real life)
- allow research in areas that could not happen due to ethical or cost reasons
HOWEVER
-may occur rarely and reduce research opportunities
-may not be randomly allocated to experimental conditions so other factors may affect the DV
Quasi-experiments
-IV based on an existing difference between ppts and un manipulated
-DV can either be naturally occurring or devised by experimenter
Strengths and limitations of quasi-experiments
-same as lab due to controlled conditions
HOWEVER
-cannot randomly allocate ppts to conditions so may be confounding variables
-other ppt variables may come into play
Population&target population
-large group of individuals that a researcher is interested in studying
-subset of general populations that research is specifically about
sample
-ppts representative of the target population so results can be generalised
Random sample
each person in target population has an equal chance of being chosen
1. have a list of the target population
2. enter names into hat, computer system
3. Place names into either conditions until there is an equal number of ppts in each condition(reference the number according to target population in the question)
Advantages+disadvantages of random sample
-avoids researcher bias, as researcher cannot choose ppts from sample
HOWEVER
-could produce an unrepresentative sample
- can be time consuming to get full list of target population
systematic sample
-ppts chosen from list of target population and every nth ppts chosen to from sample
-process is repeated until sample required is complete
-method is subjective
advantages+disadvantages of systemic sample
-avoids researcher bias because researcher can’t choose who they want in sample(objective)
HOWEVER
-could result in unrepresentative sample
-time consuming and ppts may refuse to take part
- if target population is too large full list difficult to obtain
opportunity sample
-researcher directly asks members of the target population if they are avaliable
-likley to be someone researcher has easy access to or familiar with
advantages+disadvantages of opportunity sample
-convenient as there is no need to divide the population into different strata
-less costly
HOWEVER
-researcher bias as researcher asks who they want to take part in the study, may choose those who they prefer
- unlikley to be representative (for example research conducted in Unis is likley to have young undergrad students so cannot be generalised
volunteer sample
-ppts offer to take part study after seeing an ad in a newspaper or online
advantages+disadvantages of volunteer sample
-easy, minimal input from researcher
-less time consuming
-ppt ends up more engaged and willing to participate
HOWEVER
-may not be generalisable to target population becuase of volunteer bias (more likely attract a certain profile i.e more curious and helpful)
BPS CODE OF CONDUCT
Informed consent: ppts must be told about research before taking part and their right to withdraw
Deception: deliberately misleading or withholding info from ppts at any stage of research.
protection from harm: ppts should not be exposed to psychological/physical harm more than expected in day to day life
Confidentiality: right to have all personal info protected under the data protection act away from publications
privacy: right to control info about themselves.
how we deal with informed consent issue
Prior consent: ppts can sign a consent form before the study
presumptive consent: details explained to a similar representative group and ask if they would agree to conditions
retrospective consent: ppts asked to give consent after the study
how is protection from harm and deception dealt with?
a debrief where:
1) you thank them for taking part
2)a full explanation of the study is offered(aim, hypothesis) and further psychological support
3) right to confidentiality(withhold data) and right to withdraw info is given
how do we deal with confidentiality?
-try and maintain anonymity through giving ppts numbers or using initials in a case study.
-remind ppts before and after study that data will be protected
psychological realism
thought processes that participants use in the study may be quite common in the real world
costs and benefits of ethics issues
:( harmful for ppts, time consuming, invasion of privacy, can be upsetting, lead to harmful conclusions
:) ppts get paid for their time, researchers can develop theories from it
Explain the role of ethics committees
-researcher must write a proposal outlining how the researcher will meet ethical guidelines+ justification for any breaches
-all research must be approved by the committee
-made up of lay people+independent staff members
-if a researcher behaves unethically, they are barred from practicing as a psychologist
What is the ‘measures of central tendency’?
‘averages’ which give us information about the most typical values in a set of data
What are descriptive statistics?
the use of graphs, tables and summary statistics to identify trends and analyse sets of data
The mean(the average)
-most sensitive of the measures of central tendency as it includes all of the scores/values in the data set within the calculation= most representative
-harder to calculate than other measures
-cannot be used for categorised data
-can be affected by extreme values
What is categorised data?
can be divided into different categories but cannot be ordered/measured
can be nominal i.e hair colour, binary i.e has a pet or doesn’t and ordinal i.e ranking things 1st,2nd and 3rd
The median(middle value when scores are ordered highest to lowest)
-easily identified (just find the middle of the 2 medians in an even number of values)
-easier to calculate than the mean
-unaffected by extremes
-less representatives as it doesn’t not use all scores and ignores the actual highest/lowest values
The mode(most frequent score)
-very easy to calculate
-data is often multi-modal so doesn’t show much
-unrepresentative as it doesn’t use all scores
-sometimes the only method that can be used
What are measures of dispersion?
based on the spread of scores: how far they vary from one another
What is range?
-taking the lowest score from the highest and (usually) +1
-adding 1 that allows for the fact that raw scores are usually rounded when recorded
Evaluation of range
-shows the overall spread of data in a set
-easy to calculate
-may not be representative of the data if there are extreme values at the top/bottom of the dataset
Standard deviation
-calculate mean, subtract it from each score to get the deviation, square the deviation, find the total of the squared deviations, divide that by the number of scores in the set minus 1, find the square root of this
-answer is how many values away scores lie from the mean
evaluating SD
-provides an accurate measure of data spread as it takes all into account
-provides useful information about how individual scores relate to each other and to the mean
-gives a measure of the spread of the data as lower SD mean there was little variations between the scores
-it is harder to calculate than the range
Bar charts
-used when data is divided into categories(discrete data)
-difference in mean value can be easily seen
-bars are separate to show different conditions
-frequency is plotted on y axis and condition is plotted on x axis
Histograms
-data is continuous not discrete(bars touch each other)
-x axis is equal sized intervals of a single categories
-y axis represents the frequency within each interval
Scattergrams
-depict associations between co variables
-either co variable occupies the x axis and y axis and each point on the graph corresponds to the x and y position of the covariables
What is qualitative data?
-meaningful data produced in a natural setting
-expressed in words and may be a written description of thoughts, feelings and opinions
-relies on interpretation of language
Evaluating qualitative data
-HIGH EXTERNAL VALIDITY= more rich in detail as it broader in scope and allows ppts to go in depth about thoughts, feelings and opinions
-difficult to analyse as it tends not to be summarised statistically
-may rely on subjective interpretations, vulnerable to bias
What is quantitative data?
-collected and expressed numerically(normally via individual scores)
-aim to produce results that can be compared and analysed i.e graphs/charts
-i.e questionnaires, experiments
Evaluating quantitative data
-easy to analyse so comparisons can be drawn
-can be replicated easily
-more objective and less vulnerable to bias
-produces more narrow, oversimplified data = may not represent real life
Define primary data and evaluate it
-researcher generates their own data to answer a specific research question i.e experiment, questionnaire, interview, observation
D:its time consuming, and its expensive
A: Has authentic data that answers that question
Define secondary data and evaluate it
-researcher uses data collected by others that were created to answer another research question (reports, government statics, content analysis), and significance is known due to prior statistical testing
D: less valid as data wasn’t collected to answer that research question i.e outdated and incomplete
A: data is collected easier as its cheaper and less time consuming
What is meta analysis?
-a statistical technique used to gather data from lots of studies with the same aim/hypothesis and combine them to see the overall effect in the conclusion(secondary data)
A: large number of ppts, so results can be generalised across populations(validity)
D: studies vary, with different methods, aims etc
-publication bias = researcher may choose to ignore studies with negative/ non significant results
Name 2 self report techniques
-Questionnaires
-Interviews
What are questionnaires and how are they designed?
-pre set list of written questions where ppts are assessed on how they think/feel or the DV is assessed
Likert scale= indicates agreement with usually 5 points
Rating scales= gets ppt to identify a value to represent strength of feeling towards something
Fixed choice option= list of possible options and ppt has to ‘tick all those that apply’
Evaluating questionnaires
A: -cost effective, easy to collect representative data, easy to analyse, less social desirability bias than interviews, fast + easily distributed
D: low response rate, unreliable responses as ppts are unmotivated, confusing questions leading to acquiesce bias, (closed) answers are restricted, bias sample(location/helpful people), cannot be used on young children or people who are illiterate,
What is an open question
-advantages and disadvantages
-questions phrased in way that allows ppts to answer in a way they choose
Advantage
-freedom to choose what they say, can lead to more valid responses
disadvantage
-makes data analysis more difficult(qualitative data)
What is a closed question
-advantages and disadvantages
question phrased in a way that limits ppts responses to a few options
Advantages
- allows easy data analysis of large number of responses(quantitative)
- easier to spot patterns
Disadvantages
-ppts responses are fixed so it lacks depth, which is less valid
What does the researcher need to do when conducting a self-report
- Avoid complex terminology: ppts may not understand or by too embarrassed to say they don’t, this can then result in inaccurate responses
- Rewording questions: when a ppts doesn’t understand a skilled interviewer could reword it in way that doesn’t change the meaning of the question
- Double barreled questions: ppts may agree with one part of the question but not the other, makes the question confusing (CLARITY)
- Leading questions: this can bias responses in a certain direction, to avoid this questions should be written in a way that doesn’t suggest a correct way (BIAS)
-Must be written in a way that they are easy to analyse
How can a researcher test whether their self-report study has issues
run a pilot study
small scale study, which helps the interviewer or researcher identify problems like confusing questions, make sure questions don’t give away the aim, or to check if they don’t produce a meaningful and detailed response
- this allows for things to be changed
Why are filler questions used in self-report studies
Questionnaire: to stop ppts from finding out the aim
Interviews: to ease the ppts in, to make them feel more comfortable
Designing an interview
-most include interview schedule, with questions they intent to cover(reduces interviewer bias)
-group interviews common in clinical settings
-one to ones should be done in a private, quiet room to illicit vulnerability
-neutral questions to establish rapport
-interviewees should be reminded that answers will be treated in the strictest confidence
What is a structured interview and what are the advantages and disadvantages
the interview asks a list of prepared questions in a set order
Advantages: interviews do not have to be highly trained
responses are easier to compare because same questions were used
easy to replicate
Disadvantages: responses can’t be followed up with additional questions to get ppts to elaborate
What is an unstructured interview and what are the advantages and disadvantages
When there is not a setlist of questions, so its an open conversation about the topic
Advantages: interviewer can build a rapport, so ppts feels more comfortable answering questions
-responses can be followed up with questions to gain insight + receive unexpected info
Disadvantages: need a highly trained interviewer to think about questions
-every interview will be different so it’s hard to come up with comparisons
-interviewer bias
What is a semi-structured interview and what are the advantages and disadvantages
combination of prepared questions and use of additional questions for elaboration
Advantages: easy to compare because same questions used
- interviewer can ask follow up questions for ppts to elaborate on ideas
- rapport can be built
Disadvantages
-highly trained interviewer
Why are observations important?
-allow researchers to study what people do without having to ask them
-allows behaviours to be observed in natural or observed settings
-study complex interactions between variables
What is naturalistic observation and what are the advantages and disadvantages
-In a real life setting i.e studying workers in a factory
Advantages
- high external validity likely to show more natural behaviour, as its easier to generalise
- fewer demand characteristics
Disadvantages
-low levels of control makes replication difficult confounding/extraneous variables contributing to behaviour
What is a controlled observation and what are the advantages and disadvantages
Aspects of environment are controlled and some variables manipulated, to give ppts same experience, Often conducted in a lab (Ainsworth and Bandura)
Advantages
- high control reduces likelihood of confounding/extraneous variables which makes replication easy
- Results reliable as they used same standardised procedures
Disadvantages
-low external validity because environment is artificial, Behaviour may not be repeated in actual environment
What is an overt observation and what are the advantages and disadvantages
Ppts know they are being observed
Advantages
-ethically correct as ppts gave informed consent to being observed
Disadvantages
-risk of demand characteristics and social desirability bias
What is covert observation and what are the advantages and disadvantages
Ppts don’t know they are being observed and are unaware of the focus of the study
Advantages
-no demand characteristics and research has high validity
Disadvantages
-highly unethical
What is participant observation and what are the advantages and disadvantages
Observer joins the group and takes part
Advantages
-builds a rapport, insight (you understand more about the situation through experience)
Disadvantages
-researcher bias (may start to take on opinions), may lose objectivity as they identify too strongly with their role
What is non participant observation and what are the advantages and disadvantages
Observer is outside of the group and takes a separate role to the group
Advantages
-increases objectivity due to psychological distance from ppts
Disadvantages
-may miss details or insight because unable to build a rapport so behaviour less naturalistic
What is an unstructured observation?
-when the behaviors of interest are not clearly specified in advance of the study; researcher might write down everything they see
-produce accounts of behavior that are rich in detail
Evaluating unstructured observations
:) rich, detailed data
good for small scale level with ppts
:( greater risk of observer bias , due to lack of objective behavioral categories –> researcher may only record what ‘catches their eye’(may not be important/useful)
What is a structured observation?
-the psychologist target certain behaviors which allows the researcher to quantify their observations using a predetermined list of target behaviors and sampling methods
Evaluating structured observation
:) make recording of data easier and more systemic
quantitative data= easy to analyse + compare behavior observed
:( less rich data
Behavioural categories
-breaking up target behaviours into categories that are precisely defined, observable and measurable
-made up of what the researcher sees+hears
A: -can make data more structured and objective(if follows 3 criteria)
D: -must ensure all possible forms of behaviour are included in checklist, no ‘dustbin’ category to dump behaviours
-categories should be exclusive and not overlap
Event sampling
-tally the number of times a behaviour occurs in a target individual/group
-A: -if behaviour is on list, always will be recorded
-helps catch behaviours that might be missed
D: -miss detailed behaviour if event is too complex
-need lots of observers
Time sampling
-record behaviours within a pre-established time frame
A: more flexible are able to record more unexpected behaviours
-reduces no. of observations that have to be made
D: can miss behaviour that is not in time frame–} unrepresentative of behaviour as a whole
Content analysis
-technique for analysing data of various kinds.
-Data can be placed into categories and counted (quantitative) or can be analysed in
themes (qualitative)
-uses coding systems before analysing material(flagging certain info and counting how many times ‘code’ occurs)
How do you do content analysis?
-Sampling: decide what kind of info will be used to analyse
-Pilot: familiarise with material(helps us decide what common themes occur)
-Highlight key common themes and create coding categories (specific, measurable things within theme) based on them
-Tally number of occurrences of coding categories(nominal data/quantitative figures)
-Researchers will compare different results to specify data + pick best data to back up hypothesis
Cumberbatch and Gauntlett (2005)
-commissioned by OFCOM
-looked at how smoking, alcohol + drug abuse was shown in TV shows watched by 10-15 year olds
-looked at 246 programmes(76% soap operas) between Aug-Oct 2004 before 9pm watershed
found that there were:
-1.2 drug related incidences per hour(bad light)
-3.4 smoking incidences per hour(neutral/negative light)
-12 drinking incidences per hour(neutral/negative light)
-only 4% of shows tended not to show any of these(mostly game shows)
Evaluating content analysis
A: -avoid ethical issues because it uses public data so no need to obtain permission
-high external validity because of real life data
-can produce qualitative AND quant data
D: - subjective as researcher needs to interpret data without context, may lead to researcher bias
data was not created for purpose of content analysis, so may lack internal validity(lack of objective categories)
What is thematic analysis?
-qualitative method of identifying, analysing and reporting themes in a set of data
-i.e interview transcripts
Steps of thematic analysis
-Familiarisation with the data (reading it)
-Coding (labels to identify important parts)
-Search for themes (examine labels/codes for patterns)
- Review themes (checking if themes explain the data + refining them)
- Defining and naming themes (analysis)
- Writing up (info is combined from the analysis to create a report)
Evaluating thematic analysis
:) -allows analysis of qualitative data so conclusions can be drawn
-flexible + can be modified for many types of studies
:( -time consuming and complex
-relatively new so it is difficult to find details of other relevant studies that have used it
Correlations
-involves comparing data from same ppts
-test relationship between 2 variables
-alternative hypothesis = will be about relationship not difference between conditions
-have covariables(2 factors that compare to each other)
Difference between correlations and experiments
-Correlations do not manipulate or have IV + DV whereas exps looks at effect of IV on DV
-difference between variables vs relationship/association between variables
-correlational designs are studies i.e observations, self report and exp use experimental design i.e matched pairs, ind groups, repeated measures
Types of correlations
Positive: co-variables increases with the other co-variable
Negative: as one co variables increases the other co variable decreases
Zero: no relationships between co variables
Correlation coefficient
shows us the numerical strength and direction of the relationship between co-variables as a number between -1(perfect negative correlation) and +1(perfect positive correlation)
If its higher than 0.8 it has a strong positive correlation
lower than -0.8 has a strong negative correlation
Evaluate the use of correlations
A: -highlight potential causes and relationships between variables, which can be tested experimentally later –} may be unexpected, highlight pattern
-little manipulation of data so it is easier and more economical to carry out(secondary/existing data)
D: -correlation does not mean causation, even if a relationship exists(could be coincidence, unknown direction of causality)
-unknown co variable/3rd variable may cause relationship
-data is not operationalised so it may lack validity
Case study
-in-depth, detailed investigations of one individual/group/event
-usually conducted since individuals are rare/unique
-used a range of RMs
-could happen over extended time period.
-RM originated in clinical medicine (i.e patients personal history)
What is triangulation?
Use of different approaches to gather data to improve the credibility of the conclusions.
Increased validity of the case study
Strengths of case studies
-rich and detailed qualitative data increase validity(higher mundane realism)
-allows for investigation of otherwise impractical/unethical situations
-triangulation helps make new hypotheses–} results drawn from many areas
Limitations of case studies
-observer bias & subjective opinions i.e Freud + Little Hans
-based on one example so cannot generalise to other times, cultures, people places etc= unrepresentative
-cannot be replicated
-triangulation is time consuming
Pilot studies
-small scale version of the actual investigation, used as a trial run.
-may include direct feedback from ppts
-adjustments can be made after
-can save time/money in the long run
Assessing reliability: test-retest
-administering the same test or questionnaire to the same person on a different occasions
-results obtained should be similar
-normally with questionnaires, but can be used with interviews too
-sufficient timing so ppts can’t recall their answers but their attitudes/abilities haven’t drastically changed
Test-retest: Beck et al (1996)
-studied the responses of 26 outpatients on 2 separate therapy sessions 1 week apart + found a correlation of .93
-high test-retest reliability of the DAS(depression and anxiety scale)
-coefficient +.80 and above= reliable data
Intra-rater reliability
-measure of how consistent an individual is at measuring a constant phenomenon (test-retest)
Inter-rater reliability
-how consistent different individuals are at measuring the same phenomenon i.e content analysis
Inter-observer reliability
-may involve a small-scale pilot study to check that the observers are applying behavioural categories in the same way
-record data independently and compare results
Difference between inter-rater/observer/ interviewer
-different research methods that all measure the consistency of different individuals
Improving reliability: questionnaires
-test-retest methods–} comparing 2 sets of data should produce a +.80 correlation
-low test-retest reliability= may require questions to be deselected or rewritten
-i.e replace some open questions for a more fixed-choice approach
Improving reliability: Interviews
-use the same interviewer every time or properly train every interviewer i.e AVOID leading questions
-easy to maintain consistency in structured interviewers
Improving reliability: Observations
-behavioural categories are operationalised (measurable and self evident)
-categories shouldn’t overlap, otherwise observers will use their own judgement–} inconsistent results
-low reliability= need to train observers in techniques and make sure categories are agreed upon
Observational research: Meltzoff + Moore
-controlled recorded observation of caregiver - infant interactions
-each observer scored the tapes twice so that both intra-observer and inter-observer reliability could be calculated
-all scores were above .92
Observational research: Ainsworth’s strange situation
-found .94 agreement when rating exploratory behaviour–} strength of the strange situation
Improving reliability: experiments
-standardised, consistent procedures
What is validity?
-refers to whether a psychological study produces accurate results
-refers to how true or legitimate something is an explanation of behaviour
Internal validity
-includes whether the researcher has managed to measure what they intended to measure
-whether the IV produced the change in the DV
-demand characteristics + confounding and extraneous variables
External validity
-extent to which findings can be generalised beyond the research setting in which they were found
-affected by internal validity
i.e temporal, ecological and population validity
Assessing validity
-face validity–} whether a test or measure appears ‘on the face of it’ to measure what it is supposed to (eyeballing or expert checking the method etc)
-concurrent validity–} when the results obtained are very close to, or match, those obtained on another well-established test
-close agreement +.80
Features of a science: Objectivity
-researchers must keep a ‘critical distance’ during research—} detached from emotions/personal bias that can influence the data
-associated with a high level of control(lab experiments)
-used to establish general laws
-EXAMPLES:
• F-scale used pre-written list of questions to measure obedience to authority
• Pavlov used an objective lab experiment to establish laws on classical conditioning
• brain scanning techniques
Counterpoints to objectivity
-Some theories were established using subjective, qualitative methods i.e:
• Freud’s theory of the unconscious mind is based on therapy sessions with his clients(dream analysis etc)
• case studies—} Danelli & EB, Phineas Gage
Features of a science: The empirical method
-emphasise the importance of collection based on direct, sensory experience
-objectivity is the basis of this
-uses experimental/observational methods
EXAMPLES:
• Baddeley used lab experiment to determine STM coding
• Asch’s line experiment to assess conformity
• Ainsworth used a controlled observation to distinguish attachment types in babies
Counterpoints to the empirical method
-the approaches differ in their alignment with empiricism:
• Humanistic psychologists focus on person-centred approaches + loose theories
• Psychodynamic’s focus on case studies
Features of a science: Replicability
-findings of a ‘trusted’ scientific theory must be repeatable across a no. of contexts and circumstances
-helps determine the validity + reliability of research
-research must have a standardised procedure so that other researchers can use it to verify their work
-EXAMPLES:
• Ainsworth’s Strange Situation has been replicated across cultures
• Pavlov used a standardised procedure to establish laws on classical conditioning
• Asch’s line study
• Milgram’s study on obedience has been repeated in his different variables
Counterpoints to replicability
-most animal studies cannot be replicated due to ethical issues:
•Harlow and Harlow
•Skinner’s rats
•Decoursey et al
•Ralph et al
Features of a science: Falsifiability
-Karl Popper argues that scientific theories should hold themselves up for hypothesis testing + possibility of being proven false
-even when a scientific principle had been tested successfully repeated it was not necessarily true
-an alternative hypothesis must always be accompanied by a null hypothesis
Features of a science: Theory construction and hypothesis testing
-theory construction occurs through gathering evidence via direct observation
-a hypothesis can be tested using objective methods
-if alternative hypothesis is accepted, the theory will be accepted
-if the null hypothesis is accepted, the theory will be revisited or rejected
-deduction= if the process of deriving new hypothesis from an existing theory
Steps of hypothesis testing
-theory—} hypothesis—} observation—} confirmation(falsify)
Features of a science: Paradigms
-clear, distinct concept accepted by most people in a scientific field i.e evolution, planets orbit the sun
-according to Kuhn, psychology has too many internal disagreements to be a science
-however there are set theories i.e STM, LTM, the nervous + endocrine system, synaptic transmission etc
Features of a science: paradigm shifts
-Kuhn suggested that progress within a particular science occurs when there is a scientific revolution—} group of researchers begin to question the accepted paradigm
-occurs when there is too much contradictory evidence to ignore i.e used to be accepted that the sun revolves around astronomers challenged this with empirical evidence + it then became accepted that the earth orbited around the sun
-EXAMPLES:
• MSM, WMM, holistic vs localisation, role of the father, homosexuality(used to be criminalised + in the DSM)
What is nominal data?
-most basic level of measurement
-data is put into tally charts/ discrete categories
-very little info can be derived from it(only a ‘head count’)
-cannot be ordered/ranked
-considered the lowest level of measurements because it only shows us the mode of the data
What is ordinal data?
- data has a set order/scale to it(is RANKED)
-distances between numbers/options are uneven/unknown
-test scores ‘at least ordinal’—} for any measurement where you cannot guarantee equal distance between data
-can find the mode + median
-can be described as ‘unsafe data’ because it relies on subjective methods i.e questionnaires and there are no equal intervals between the rankings
What is interval data?
-this is when scores have equal intervals between them + are mathematical scores i.e time and distance
-zero does not mean nothing in internal data, it is another no. on the scale as we can go into the minus numbers
-interval data will have a measurement i.e degrees celsius, mins etc
What is statistical testing?
-used to determine whether a difference/correlation is statistically significant
(more than the probability that the results are due to chance)
-determines whether we accept or reject the null hypothesis
What does significance in a test indicate?
-that the IV has caused the change in the DV
-not significant= the change/difference didn’t vary greatly or the change in the DV is due to chance
Probability
-the likelihood of something occurring
-psychologists want to be at least 95% sure that the results are significantly different and not due to chance–} they therefore use a significance level
What is a significance level?
-the point at which a researcher can claim to have discovered a large enough difference/correlation to accept the alternative hypothesis(and reject null)
-the usual significance level is p ≤ 0.05
(the probability that the results are due to chance is less than or equal to 0.05)
The null hypothesis(Ho)
-states that there is no difference between the conditions
-the stat test determines whether we accept or reject the null hypothesis
Experimental/alternative hypothesis
-can be one-tailed(directional): direct affect of IV on DV
-can be two-tailed(non directional): there will be a difference/relationship between…
What are the 3 D’s?
-they determine what statistical test you choose
- test of difference or correlation?
- experimental design? (independent measures /unrelated or repeated measures /related)
- the data? (nominal, ordinal or interval)
What is the pneumonic to remember the stat test table?
Chicken Should Come
Mashed With Sweetcorn
Under Roast Potatoes
What does the stat test pneumonic stand for?
Chi-squared, Sign test, Chi-squared
Mann-Whitney, Wilcoxon, Spearman’s Rho
Unrelated t-test, Related t-test, Pearson’s r
What is the exception with parametric tests?
-the bottom row(related, unrelated and Pearson) are all parametric tests–} more powerful + robust than the other tests
-regardless of the 3D’s, if the assumptions of a parametric test is met, use interval data
What are the assumptions of a parametric test?
-data must be interval level
-the data should come from a normal distribution (NOT SKEWED)
-data should have equal variance between each group(similar SDs due to the same dispersion)= homogeneity of variance
What is the measure of central tendency + dispersion for the 3 levels of data?
-nominal= mode
-ordinal= median & range
-interval= median & SD
How can we check for statistical significance?
-the calculated/observed value(the result of the stat test) is compared with the critical value(what tells us whether to accept/reject alternative)
-each stat test has its own set of critical values
What are the 3 steps to finding the critical value to use?
-is the test one-tailed or two tailed?
-find the value of N(the no. of ppts)*
-what is the level of significance or p value? (would be 0.05 unless another value is in the question)
*for some tests the degrees of freedom has to be worked out: Pearson’s r= (N-2), Related t-test=
(N-1), Unrelated t-test=
(Na + Nb - 2)–} ppts in condition a and ppts in condition b
How should you answer a question about checking for significance?
-the calculated value… is more/less than or equal to the critical value of… with P at… for a one/two tailed test, therefore the alternative hypothesis is accepted and the null rejected/ the alternative hypothesis is rejected and the null accepted
What are type I and type II errors?
-type I= null rejected and alt accepted when it should have been the other way around(optimistic error)
-type II= null accepted and alt rejected when it should’ve been the other way around(false negative)
Why do these kinds of errors occur?
-occur due to the ‘alpha value’ (or the p-value)
-the standard alpha value of p< 0.05 balances the types of errors that may occur
- lenient alpha= significance level is too high (p<0.1)
- strict alpha= significance level is too low (p<0.001)–} normally used for socially sensitive research or research trialling drugs as it must be very precise
Normal distributions
-symmetrical frequency of measurements to form a bell-shaped curves
-the mean, mode and median are all at the middle of the curve(all the same or very similar value)
-the mode will always be at the peak of the distribution because it has the highest frequency
-the ‘tails’ of the curve extend outwards but never touch the horizontal x-axis (more extreme scores are always theoretically possible)
Positive skew
-where most of the distribution is concentrated towards the left of the graph—} a very difficult test produces a positive skew as most people achieve a low score
-mode is the lowest value and mean is the highest value(extreme scores affect the mean, so abnormally high scoring candidates increase the mean)
-slide= positive
Negative skew
-where the bulk of the scores are concentrated to the right, resulting in a long tail of anomalous scores on the left
-very easy test produces a negative score—} mode is the highest value as the most frequent score will be high
-mean is pulled to the left as it is affected by extremely low scores
-negative= hill
What is peer review?
-all aspects of the written investigation being scrutinised by a small group of experts in the particular field—} should conduct an objective review + be unknown to the author
What are the main aims of peer review?
-to allocate research funding—} peer evaluation helps decide this + they could co-ordinate with govt-run funding organisations
-to validate the quality + relevance of the research i.e hypothesis, methodology, stat tests being used
-to suggest amendments or improvements to the report
Evaluating peer review
- anonymity= some reviewers may use their anonymity as a way of over criticising rivals as they are in competition for research funding
- publication bias= natural tendency to want to publish ‘headline-grabbing’ findings to increase their credibility or to publish positive results
- burying groundbreaking research= researchers may want to maintain the status quo and be critical of research that contradicts popular views
Implications of role of the father research for the economy
-recent research suggests. that the father may fulfil a different but valuable role from the mother—} both parents are equally capable of providing the emotional support necessary for healthy psychological development
-may promote more flexible working arrangements within the family
-couples share childcare responsibilities across the working week—} can maximise their income and contribute more effectively to the environment
Implications of the development of treatments for mental disorders for the economy
-patients are able to be assessed quickly and gain quick access to treatments i.e anti anxiety drugs
-referrals can be made my GPs for psychotherapies i.e CBT + SD
-means that people with mental disorders are able to manage their conditions effectively and return to work
