4.2.3 Research Methods Flashcards
the experimental method
- looks at how variables affect outcomes
- e.g. Bandura’s Bobo doll experiment looked at how changing the variable of the role model’s behaviour affected how the child played
laboratory experiments
- an experiment conducted in an artificial, controlled environment
- the researcher has high levels of control over all variables and environmental factors within the experiment to ensure only the IV changes between conditions
- standardised procedures are used for this control
- e.g. asch’s conformity experiments
lab experiments - strengths
- cause and effect conclusions are more possible due to the high levels of control
- the use of a standardised procedure means the research is replicable, which increases reliability
- high internal validity as the IV may be seen to affect the DV
lab experiments - limitations
- demand characteristics may be an issue as ppts know they’re in a study which may alter their behaviour, affecting the validity
- often lacks ecological validity due to the artificial nature of the procedure
- often lacks mundane realism, i.e. results can’t be generalised to real-world behaviours
field experiments
- an experiment carried out in a naturalistic, real-world setting, e.g. a school
- an IV is manipulated by researchers and the DV is measured quantitatively, as in lab experiments
- e.g. bickman’s study of the effects of uniform on obedience
field experiments - strengths
- likely to have higher ecological validity as it’s a real-life setting
- ppts are less likely to show demand characteristics as they’re less likely to know what’s expected of them
- high levels of mundane realism, i.e. results are more likely to be able to be generalised to real-world behaviours
field experiments - limitations
- harder to assign ppts, so changes may occur due to ppt variables rather than what the researcher’s measuring
- harder to control extraneous variables within the experiment, which could affect measurements of the DV
natural experiments
- where the researcher doesn’t manipulate the IV as it varies naturally
- the researcher takes advantage of the naturally changing IV to measure effects on the DV
- ppts are usually allocated to conditions randomly
- e.g. research into white blood cell activity before and after students take exams
natural experiments - strengths
- allows research in areas that controlled experiments can’t research, due to ethical or cost reasons
- high external validity as it’s conducted in a natural setting
- low demand characteristics as natural behaviours are being displayed
natural experiments - limitations
- hard to establish causal relationships as other variables aren’t controlled so may have an effect
- lack of reliability as it’s unlikely to be able to replicate the same situation to repeat the test
- ethical issues as deception is often used, making informed consent difficult, and confidentiality may be compromised if research is conducted in an identifiable place
quasi experiments
- where experimenters can’t randomly allocate ppts to conditions as the IV is natural and can’t be changed as it’s a particular feature of ppts
- e.g. gender, the existence of a mental disorder, etc.
quasi experiments - strengths
- often carried out in controlled conditions
- higher ecological validity as research is often less artificial than lab studies, so results are more likely to be able to be generalised
quasi experiments - limitations
- ppts aren’t randomly allocated, so confounding variables may affect results, e.g. where the ppt lives
- hard to establish causal relationships as the IV isn’t being directly manipulated
naturalistic observation
- observations made in a real-life setting
- e.g. setting up cameras in a school to see how people interact in those environments
controlled observations
- observations made in an artificial setting set up for the purposes of observation
- e.g. zimbardo’s stanford prison experiment
covert observation
- ppts are unaware they’re being observed as part of a study
- e.g. setting up hidden cameras in an office
- higher validity as it rules out demand characteristics
- unethical as ppts don’t have informed consent
overt observation
- ppts are aware they’re being observed as part of a study
- e.g. zimbardo’s prison study
- ethical as ppts have given informed consent
- social desirability is likely, where ppts present their ‘best selves’ to the researcher
- demand characteristics are more likely, which impacts validity of results
participant observation
- when the researcher is actively involved in the situation being observed
- ppts may not realise the researcher isn’t really one of them
- e.g. zimbardo playing the role of a prison warden in his own study
- researcher is able to build a rapport (relationship) with ppts, so they’re more likely to have open conversations and act in a natural way
- researchers can become too involved with ppts which makes their interpretations of ppts’ behaviours biased
non-participant observation
- when the researcher isn’t involved in the situation and remains separate from ppts
- e.g. in bandura’s bobo doll study, the observers didn’t interact with the children being observed
- researcher is more likely to remain objective whilst observing and recording the ppts behaviour
- researcher is unable to build rapport (a relationship) with ppts, so they’re less likely to open up fully / show fully natural behaviours
self-report techniques
when ppts provide info about themselves, either through questionnaires or interviews
questionnaires
- a standardised list of questions answered by all ppts in a study
- questions can either be open or closed;
- open; allows ppts to write their own answer, so often seen as having more validity, and it qualitative data
- closed; ppts have to choose from a fixed set of responses (multiple choice / yes or no), which gives quantitative data
questionnaires - strengths
- closed questions give quantitative data which is easier to analyse and spot patterns
- questionnaires are standardised, so studies can be easily replicated which can strengthen reliability
questionnaires - limitations
- open questions provide qualitative data which can make analysis difficult
- ppts may lie in their answers, e.g. on controversial topics
- differences in interpretations of questions
- ppts are less likely to provide full detail in written answers, so interviews may be better suited for gaining detailed info
- responses may be biased, e.g. towards people who have a lot of spare time, as they’re probably more willing to complete it
structured interviews
- questions are standardised and pre-set
- the interviewer asks all ppts the same questions in the same order
- means the interviewer doesn’t need intensive training to ask questions
- answers are easy to compare as the same questions were asked
- ppts’ responses can’t be followed up with extra questions to add more detail
- interviewer effects - their appearance or character may bias ppts’ answers, e.g. females may be less comfy answering questions on sex by a male interviewer
unstructured interviews
- the interviewer discusses a topic of interest with the ppt in a less structured and more spontaneous way, pursuing avenues of discussion as they come up
- interviewer is able to build a rapport with the ppts, which is more likely to allow for honest answers and higher levels of validity
- answers can be followed up for more info / extra detail
- interviewer has to be highly trained and ready to come up with suitable questions on the spot
- lack of quantifiable data as every interview will be different, making it harder to compare
- interviewer effects
semi-structured interviews
- combination of set questions, with the ability to add in extra questions to gain more info
- easy comparison as the same questions are asked
- more info can be gathered from extended questions
- rapport can be built as the interviewer can ask more questions and relax the interviewee
- interviewer has to be highly trained and ready to come up with suitable extending questions
correlations
- analysis of relationships between co-variables
- two co-variables are measured and compared to find a relationship, so variables aren’t manipulated
- e.g. age and reaction time
- the relationship between co-variables can be analysed visually through scattergrams, or numerically through the correlation co-efficient
scattergrams
- these can provide 3 outcomes;
- positive correlation; when one co-variable increases, so does the other
- negative correlation; when one co-variable increases, the other one decreases
- zero correlation; when there’s no relationship between variables
correlation co-efficient
- represents both the direction and strength of the relationship between co-variables as a number between -1 and +1;
- a perfect positive would be +1
- a perfect negative would be -1
- no relationship would be 0
- both positive and negative correlation coefficients can be described as weak, moderate or strong, depending where they fall between the numbers
correlational analysis - strengths
- able to show relationships between variables
- data is already easily available for researchers to analyse
- as data is readily available, it’s unlikely for there to be any ethical issues
- allows predictions to be made when looking at relationships between co-variables
correlational analysis - limitations
- they don’t show causation, i.e. unable to show which variable impacts the other
- extraneous relationships with other variables may affect the co-variables and the outcome
- only works for linear relationships; unable to work for curvilinear relationships
content analysis
- an indirect observational method used to analyse human behaviour
- it’s used to analyse qualitative data by transforming it into quantitative data through coding units, e.g. the amount of times a swear word is said in a film
- can be used for data in different formats, e.g. interview transcripts, films, audio recordings, etc.
content analysis procedure
- data is collected
- the researcher reads through / examines it to familiarise themselves with it
- the researcher identifies coding units
- the data is analysed by applying the coding units
- a tally is made of the number of times that coding unit appears
content analysis - reliability
- after conducting a content analysis, the researcher will need to test reliability, and they can use;
- test-retest reliability; run the content analysis again on the same sample and compare the results - if they’re similar, it shows reliability
- inter-rater reliability; a second rater (researcher) conducts the content analysis with the same set of data and compares them - if results are similar, it shows reliability
content analysis - Waynforth & Dunbar (1995)
- they conducted one by analysing lonely heart adverts in newspapers to see if men and women were looking for different things in relationships
- they looked at 881 adverts
- they found men aimed their adverts at younger women and mentions their resources more than attractiveness
- women aimed theirs at older males and mentioned attractiveness more than resources
- they concluded from their content analysis that men and women did want different things in their relationships
thematic analysis
- a qualitative method that allows researchers to identify, analyse and report themes
- an alternative to content analysis
- researchers attempt to identify deeper meanings in data by organising, describing and interpreting it
thematic analysis procedure
- the researcher familiarises themselves with the data
- codes are created to identify the data and important features of it
- the data / codes are examined to look for themes and identifying patterns
- the researcher reviews themes and patterns to see if they can explain the behaviour and answer the research question
- the researcher then names and defines each theme
- they make a formal write up of the research analysis and the themes found
- themes must emerge from the data and not be pre-described by the researcher
content and thematic analysis - strengths
- reliability is established as it’s easily replicable
- high levels of external validity as the data is taken from the real world and not made just for the study
- data is readily available so results are usually generalisable to other real world situations
- quick, easy and non-invasive to perform as it doesn’t need contact with ppts
- can be used to verify results from other research
content and thematic analysis - limitations
- researcher bias may occur as they have to interpret the data
- may lack validity as the data hasn’t been collected under controlled conditions
- also lacks causality for this reason
- results may be flawed due to the over representation of events and using readily available material, e.g. negative events usually have much more coverage than positive, which may skew the data to show an invalid representation of behaviour
case studies
- detailed and in depth investigations of an individual, a group of people, or an event
- researchers may use observation, questionnaires, interviews, etc. to collect data
- they usually provide qualitative data as they’re studies of a single subject
- this data can be turned into quantitative data through different analysis techniques
- researchers use data to build a case history of the subject and then interpret this to draw conclusions
- snapshots; case studies conducted over a short period of time
- longitudinal studies; follow ppts’ behaviour over a long period of time
case studies - strengths
- holistic approach as the whole individual and their experiences are considered
- allows researcher to study unique and otherwise difficult to research behaviour, e.g. it’s unethical to remove half a toddler’s brain just for an experiment, but if such a procedure is medically necessary anyway, then researchers can use this as an opportunity to learn more about the brain
case studies - limitations
- results aren’t generalisable or representative as it only looks at single examples
- researcher bias as researchers may be biased in their interpretations when drawing conclusions from the case studies
- information / details can be easily missed
- difficult, if not impossible, to replicate
scientific processes
looks at how scientific studies are organised and reported, and covers ways of evaluating a scientific study
aims
- a precise statement about what the researchers are investigating and why
- e.g. ‘to investigate the effect of SSRIs on symptoms of depression’
hypothesis
- a testable prediction of what the researchers expect to happen
- it needs to contain the IV and the DV which must be operationalised (you must have a way of testing them)
- e.g. subjects are more likely to comply when orders are issued by someone wearing a uniform
types of hypothesis
- directional hypothesis (one-tailed); states what the experimenter thinks will happen
- non-directional hypothesis (two-tailed); when the experimenter isn’t sure what will happen and is ‘sitting on the fence’
- experimental / alternative hypothesis (H₁); a prediction that changing the IV will cause a change in the DV, i.e. a directional or non-directional hypothesis
- null hypothesis (H₀); a prediction that changing the IV won’t have an effect on the DV
sampling
- used to select the ppts who will take part in the study
- the sample is selected from the target population (everyone that the study is supposed to apply to)
- the researcher uses the sample to draw from the target population then generalises the findings to them
sampling techniques
- random
- systematic
- stratified
- opportunity
- volunteer
random sampling
- selecting people in a way that everyone has a fair chance of being selected, e.g. by pulling names out of a hat
- it’s an unbiased selection, so is more likely to be a representative sample
- as results are fairly representative, results can be generalised to the target population
- time consuming and impractical as it’s not always possible for the entire target population to want to participate
- it may be non-representative as one specific group may be coincidentally selected randomly, e.g. all males, which wouldn’t be a true representation of the population
systematic sampling
- selecting every nth person from a list to form a sample
- unbiased selection, so more likely to be representative, and so results are more able to be generalised
- not always truly random or representative, as the list may involve hidden periodic traits, e.g. if every 10th person is a 19 year old, then they’re the only person in the sample which isn’t a true example of the target population
stratified sampling
- a small-scale reproduction of the target population
- involves dividing and categorising the population by characteristics important to the study, e.g. age, gender, etc., then working out what % of the population is in each group, then randomly sampling them according to these percentages
- i.e. if 20% of the whole population is aged 0-18, then the representative sample will randomly select 20% of 0-18 year olds from that group
- the sample is representative of the target population, which also makes it easier to generalise results
- selection is unbiased as it’s based on the subgroups in society
- time-consuming to know the subgroups, divide the ppts, then select them to match the population
- the researcher requires knowledge of the subgroups / categories of the population which may not be available
opportunity sampling
- selecting those who are the most convenient, willing and able to take part, e.g. by approaching people in the street and asking them to take part
- used in natural experiments as the researcher has no control over who’s being studied
- a quick and easy way to get info as it’s using people who are readily available to use
- can’t generalise results as the sample is likely to be unrepresentative, e.g. if you conduct it in a school, it will exclude people who work in jobs elsewhere
- it’s a self-selected sample as ppts can agree or decline to join in the study
volunteer sampling
- ppts volunteer to take part in a study, e.g. often by replying to adverts
- variety of ppts as they’re choosing to take part
- easy way to get info as it’s using people who’re willing to be involved in the study
- less likely to have people who want to jeapordise the study
- volunteer bias as ppts can offer differ from non-ppts as they chose to take part, so results won’t be able to be generalised
- demand characteristics are an issue as volunteers are often eager to please their researcher
pilot studies
- small-scale practice investigations carried out prior to the research to identify any issues which could arise
- they identify problems in the design, method or analysis, allowing them to be altered and fixed which ensures a higher level of validity
- ppts can also help identify issues and propose changes to prevent demand characteristics
- they also identify if there’s a chance of significant results being found, so help to see if the research will be worth it
repeated measures design
- when the same ppts complete each of the experiment conditions
- ppts are competing with themselves, producing related data
repeated measures design - strengths
- no group differences as the same ppts are used, so ppt variables won’t affect the measure of the IV
- less ppts are needed as they take part in all conditions and produce more data per ppt