- Created by: OliviaAnneSumner
- Created on: 06-11-17 17:51
*Allocate participants to conditions.
*Change in IV, researcher records/measure effect on DV.
*1 experimental group, 1 control group- compare.
*Lab, Field, Quasi, Natrual.
*Control as many variables as possible.
*Standardised procedures, same instructions.
*Know they're taking part in experiment, may not know true aim.
+High control of variables- accurate, objective.
+Easily replicated- check validity.
+Determine cause/effect- high int. validity.
+Easier to use complex equipment.
-Experimenter bias- influence participants (expecations) / act according to experiment demands (demand characteristics).
-Low Ex. validity- high control, artificial tasks in intimidating setting (artificial behaviour) difficult to generalise (mundane realism).
*Natrual setting, direct control of IV and allocation of participants to conditions.
*Unaware of participation in experiment.
+Determine causal relationships if confounding variables avoided.
+More typical behaviour- greater external validity.
-Low internal validity- hard to have full control over all variables.
-Ethics- informed consent, deception.
-Hard to obtain a lot of information, if researcher attempts to obtain a lot, participants would become aware they were in study- lose natrualness.
*Natrually occuring event used for research purpose, doesn't manipulate IV- occurs natrually and would happen regardless if researcher was/wasn't there.
+Not response to demand characteristics- unaware in study.
+High external validity- involve study of real life issues.
+See effects on behaviour of IV that would be highly unethical to manipulate.
-Hard to precisley replicate- conditions never identical.
-Hard to control extraneous variables, low int. validity.
-Not randomly allocated, observed differences may be due to differences in participant types not IV.
-Sample bias- lacks allocation to conditions, group differences.
*IV isn't determined by experimenter, based on difference that already exists (gender).
+Demand characteristics reduced (if participants unaware they're taking part, act natrually).
+High mundane realism- if task is more natrual.
-Unethical- unaware they're taking part, can't debreif, lack informed consent, invasion of privacy.
-Lacks population validity- no control over participants, sample may be biased- unrepresentitive.
-Difficult to replicate.
Non-experimental. Studied participants endaged in behaviours being studied/recorded.
Consent in Observations
Participants unaware they're being observed, should be in public- ethical.
+Removes participant reactvity, increase validity.
-Ethically questionable- no informed consent to record behaviour.
Observe after giving consent, researcher clearly visable.
-Participants may act as extraneous variables- corrupt study on purpose.
-Observer effect- behave differently, know they're being observed.
Nature of Observation
Behaviour observed under controlled contitions, 1+ variable controlled. Aware being observed.
+High control over extraneous variables.
-Low ex. validity, results hard to generalise.
Research carried out in natrual settings, all aspects of environment free to vary.
+High ex. validity, generalised.
-Risk of uncontrolled extraneous variables.
Role of Researcher in Observation
Investigator part of group being studied.
+Richer insight into study, more observational data.
+Lack objectivity, researcher starts to associate with participants.
Investigator observes from distance.
+Objective, unaffected by subjective thoughts/emotions.
-May lack rich data, info. observed from distance.
*Simplify target behaviours, main focus- scientific method.
*Quantify observations, predetermined behaviours/sampling method.
*Systematic sampling method- event sampling (count no. of times target behaviour occurs in target group), time sampling (record behaviour at pre-established time).
*Create behavioural categories- checklist, behaviour to be studied defined, made observable/measurable.
-Includes all ways target behaviour could occur, observable.
-Must be precise, measurable, exclusive (clear focus on what they're looking for), record in objective way, analysed easy (quantifiable).
Write down everything seen.
*Provides rich qualitative data.
Observer bias- research may not record all behaviour, may be interpreted subjectivley.
Obtain participant's own opinion/beliefs/thought of given topic.
*Interviews / Questionnaires.
Researcher ask question in face-face situation.
Structured- questions read to participants, responses recorded, all experience identical interview, same questions in same order.
Unstructured- less controlled, informal discussion, topic decided prior to interview starting, can explore areas of interest which arise (flexibility), can expand on responses.
*Semi-structured- combination, set of predetermined questions, respondent free to expand on responses. Free to ask follow up questions if necessery.
+Complex/sensitive issus discussed.
+Quantitative/qualitative data analysed.
+Structured- easily replicated.
-Interviewer effects, may bias answers.
-Unstructured interviewers should be highly trained, hard to find.
-Ethics- don't know true purpose, may reveal mre than they wish to.
-Might not be able to put into words what they want to say.
*Participants able to directly provide info about themselves.
*Generate large amounts of info- cheap/quick.
*Only suitable for literate participants, willing to spend time filling it in- biased.
*Contain 2 types of question- closed/open-ended.
*Researcher determines possible answers, responses often via tick boxes/circle answers.
*Best to collect factual data, easy to quantify/analyse.
*Questions may lack realism- respondent forced to give one of provided answers.
*No restriction on answer, qualitative information, greater in depth.
*Harder to analyse.
+Large samples, researcher doesn't have to be present (postal)- gather alot of responses.
+Quantitative/qualitative data produced.
+Replication easy if closed questions used.
-Misunderstanding of questions effects results.
-Low response rate- unrepresentitive.
-Closed questions- not suitable for studying socially sensitive issues.
-Social desirability bias- lie to present themselves in positive light on sensitive questions.
-Response bias- complete questionnaire in similar way, tick to agree without reading content.
Constructing Questionnaires and Interviews
*Questions always adress aims, length important- less willing to participate if time consuming.
*Questionnaires often use Likert scale, rate level of agreement. Not easy for respondents to judge, select middle value- hard to analyse.
*Consider analysis- if question isn't going to be analysed don't include it.
*Provide incensitive for completing- more willing to participate.
Affect responces recieved.
Gender/Age- topic of sensitive/sexual nature- big impact on results recieved (Wilson et al 2002).
*Ethnicity- interviewers greater difficulty interviewing people of different ethnic group. Word et al- white participants spent less time interviewing black job applicants.
*Personal Characteristics/Adopted Roles- some easier to talk to than others- use of language, accent, appearance can effect how they're interpreted by interviewer.
*Consider Open/Closed Questions- closed- options, easy to collate, display collected data and conpare specific responces. Restricted responses- get certain info, quantitative data.
Open- respondents interpret question, develop detailed response- rich info, qualitative data.
Type of Interview
Question order- start with biographical ones, interesting- encourage them to continue. All essential questions in first half.
Avoid Jargon- familiar with all words or may be reluctant to complete, if using complex terms include definition.
Avoid leading questions.
Avoid double-barrelled questions.
*Avoid emotive questions- bias response recieved?
*Don't be vague/ambiuous- clear questions.
*Don't make assumptions.
Descriptive statistical technique, measure relationship between 2 variables (co-variables)- high variables on one variable associated with igh values on another (+ correlation) or how high values of one variable are associated with low values of another (- correlation).
+ Correlation- variables increase together, relationship graphically represented on scattergram.
*- Correlation- inverse relationship, one variable increases one decreases.
*Assess extent to which co-variables are correlated using correlation co-efficient. +1 (perfect + correlation) to -1 (perfect - correlation). Increases as calculated coefficient moves away from 0, closer to + or - 1.
Correlation Analysis A03
+Allows us to study hypothesis, can't directly be examined.
+Measure many variables- relationship between them without directly manipulating behaviours.
+Provide precise quantitative measure of relationship between 2 variables.
+Make predictions once correlation found.
+Quick- ethical, doesn't manipulate behaviour.
-Hard to interpret results, inaccurate conclusions common.
-Impossible to establish cause/effect.
-Variables other than one of interest may be operating- third variable problem.
-Only works for linear relationships.
-Variables measured- no diliberate changes made, impossible to determine cause/effect.
*Indirectly observes presence of certain words/images/concepts within media. Carried out on secondary data.
*Coding, second stage data to be analysed may be large- info needs to be categorised into meaningful words and examined.
*Quantify/analyse presence, meanings, relationships of words/concepts. Make inferences.
Content Analysis Process
1.Sample of materials to be analysed examined, decide on sampling method, content they will include, how materials are selected.
2.Familiarise with/study material, may do piolet study.
3.Categories created, data classified/organised- reflect purpose of research.
4.After frequencies counted for different categories, researcher looks for themes common to many categories.
5.Researcher may ask how themes identified within categories apply to human behaviour as a whole- compare/conclude.
Content Analysis A03
+High ecological validity, based on researcher's observations.
+Takes into account personal experiences/perspectives.
+Easily replicated, assesses reliability.
+Easy to gain data from vast number of different sources.
+Avoids ethical issues.
-Different observers interpret behaviours differently, behaviour bias,
-Time consuming- subject to interpretation.
-Descriptive- doesn't reveal underlying reasons for behaviours/attitudes.
-Not performed under controlled conditions, doesn't show causality.
*Focuses on qualitative aspects of material, analyse data to identify patterns within.
*Organises/descripes/interprets data, identified themes become categories for analysis, performed through coding.
*Identifies ideas within data, theme- idea which continues to crop up. Analysis may involve comparision of themes.
*Once reasearcher is satisfied themes cover most aspects of data, they collect new set of data to test validty of themes. If they explain data adequatley, researcher writes final report including direct quotes to illustrate themes.
Thematic Analysis Process
1.Familiarisation with data- read through intensely, immerse in content.
2.Coding- generate codes, identify features of data important to answering research question.
3.Search for themes- check potential themes against data, see if they explain data/answer research question. Themes may be refined, splitting/combining/discarding.
5.Define/name themes- complete detailed analysis of each theme, create informative name for each.
6.Writing Up- combining info gathered from analysis.
*Detailed account of one person/institution/event.
*Provide rich qualitative data, research often focused on particiular aspect of behaviour- numerous methods used to study behaviour- observations, interviews.
*Can be longitudinal.
*Explanations of behaviours given in written descriptive ways, subjective reports used- includes thoughts/feelings/beliefs on particular issues.
Case Studies A03
+Rich qualitative data.
+Used to research rare behaviours/experiences.
+High external validity.
+May provide enough information to contredict theory.
+Aim to use objective/systematic methods- scientific.
+Range of research methods increases reliability- triangulation, double check results.
-Involves one person- generalise?
-Difficult to replicate.
-Researcher bias- lack impartiality- biased data.
-Retrospective data- may be unreliable.
*Small scale practice studies, allow researcher to check research, adapt problematic aspects before rolling out larger scale. Avoid unnecessary work- reduce wasted time/money.
*Participants may suggest improvements to reduce problems- demand characteristics.
*Used in non-experimental methods, self-report. Helpful to try out questions in advance, remove confusing questions. Check coding systems/train observers.
*Best way to check everything will run smoothly.
*Less viable with natrual/case studies- events/participants are rare, wasteful to sacrifice sample.
Key Features of Science
Definable subject of matter- Paradigm.
Theory must be constructed from which hypotheses are derived/tested.
*Concepts must be falsifiable.
*Must use empirical methods of investigation which can be replicated (objective).
*Must be general laws which govern human behaviour (nomathetic- generalisation).
Objectivity and the Empirical Method
*Aim to be objective in investigation.
*Keep critical distance in research.
*Not allow personal opinions/biases to discolour data they collect or influence behaviour of studied participants (researcher bias).
*High control- lab study, high internal validity, nomothetic.
*Objectivity is basis of method.
*Emphasise importance of data collection based on direct, sensory experience.
*Experimenral/observational methods are good examples of this.
*Theory can't claim to be scientific unless empirically tested and verified via. experiment/observation.
*Replication determines validity/reliability of finding.
*Is it an accurate measure of truth?
*If a scientific theory is to be trusted, findings from it must be repeatable across numerous contexts and circumstances.
*Replicating findings over different contexts/circumstances- findings can be generalised.
*Use objective method, standardised procedures/measurements, operationalised variables.
*1930- Popper- a scientific theory is falsifiability.
*Genuine scientific theories should hold themselves up for hypothesis testing and the possibility of being proven false.
*Even when a scientific principle had been successful and repeatedly tested it was not necessarily true, just hadn't been proven false yet.
*Drew clear line between good science (theories constantly challenged) and pseudosciences (couldn't be falsified) i.e. Freud's psychodynamic approach.
*Scientific theories which survive most attempts to falsify become strongest- not because they ar true, because they haven't been proved false yet.
*Psychologists avoid saying 'proves', use 'seems' / 'suggests'.
*Alternate hypothesis must always be accompanied by null hypothesis- open to fact that theory may be false.
Theories in Science
*Set of general laws/principles that have the ability to explain particular event/behaviour.
*Satisfy key principles in science- ORDER and DIRECTION. Scientists believe events in our world are ordered, can be predicted.
*Theories organised facts, find regularities/patterns- condense them into general principles.
*Provide basics for research, theory construction occurs through gathering empirical evidence from direct observation, from basis of theory it is possible to make hypothesis/prediction [supporting/disproving theory] that can be scientifically tested.
Theory Construction / Hypothesis Testing
*Possible to make clear/precise predictions on basis of theory- hypothesis testing. Essential component of theory- can be scientifically tested.
*Theories suggest numerous possible hypotheses.
*If hypothesis testing supports the theory it will be strenghtened.
*If testing refutes the theory, it will be revised/revisited/rejected.
*Process of deruving new hypothesis from existing theory- deduction.
7. No? repeat 1,2,3,4,5,6. Yes? Report.
An agreed upon set of theoretical assumptions about a subject and it's method of enquiry that are acceoted by most people in scientific field.
*Psychology isn't a paradigm- we have conflicting ideas and methods (approaches).
*Kuhn- subject can only be called a science if there is an agreed gloabal theory (ie. evolution, accepted by all biologists).
Three Stages of Science
*Kuhn- argued there are 3 distinct stages in development of a science.
1.Pre-Science: no paradigm exists, much debate about what subject is and its theoretical approach.
2.Normal Science: generally accepted paradigm, can account for all phenomena related to subject and can explain and interpret all findings.
3.Scientific Revolution: evidence against old paradigm reaches certain point, new paradigm shift. Old paradigm replaced by new one.
Why is psychology a pre-science?
*Kuhn- there is no one global unifying theory in psychology.
*Psychology categorised by many theoretical approaches, aspects of which are conflicting. Uses several conflicting methods of enquiry.
*Progess within science occurs when there is a scientific revolution, group of researchers begin to question accepted paradigm- too much contredictory evidence to ignore.
*Ie. accepted that sun revolves around earth, 17th century- work of various astronomers challenged this with empirical evidence, accepted earth orbited the sun (copernican revolution).
*Kuhn argues psychology has not undergone any paradigm shifts.
*Purpose of study- makes research more focused. Clarifies what researcher is trying to discover.
*What researcher intends to investigate.
*Includes IV and DV.
*What researcher expects will happen, precise, testable statement.
*Use directional hypothesis when previous research suggests that it is possible to make prediction/when previous research using directional hypothesis is being replicated.
*Clear about what they aim to find.
Experimental- prediction of what experimenter thinks will happen to DV when IV changes
Alternative- all hypothesis that aren't null.
*All experimental hypotheses are alternative, all alternative aren't experimental.
Directional (1-tailed): predict direction of results.
Non-Directional (2 tailed): states there will be difference, doesn't state direction of results.
Null Hypothesis: IV has no effect on DV, purpose of experimental studies is to compare merits of experimental hypothesis to null. Propose null hypothesis so we can reject it, imply that we can accept experimental/alternative hypothesis.
Manipulation of Variables
*Research focuses on relationship between variables, if researcher wants to establish cause/effect they need to use experimental method, identify IV and DV.
*Control group and experimental condition to compare effects of IV.
*If researcher wants to look at association between variables, use non-experimental method.
*IV and DV used only when experimental hypothesis used.
Operationalisation of Variables
*Define variables, i.e. effects of alcohol on reaction time:
Operationalised IV- number of units of alcohol consumed.
Operationalised DV- speed of reaction to particular stimulus.
*All studies must operationalise variables accuratley and objectivley to ensure study is reliable and valid- know what they're looking for, easier to replicate.
Control of Variables
IV manipulated, DV free to vary- all other variables must be controlled/removed. Kept constant so we can assume only variable causing change in DV is IV, other aspects- extraneous variables.
Extraneous Variables- any other variable other than IV which may have effect on DV if not controlled. Should be identified at start of study, don't vary systematically with IV, effect all levels of IV.
*Possible to control extraneous variables- test participants in same environment, standardised instructions, all tested by same researcher.
*Confounding Variables- change systematically with IV, confound results.
1-Participant Variables- age, intellegence, personality controlled across different groups.
2-Situation Variables- setting of experiment controlled- light, temp, noise.
3-Experimenter Variables- researcher's personality, appearence remain stable.
*Participant reactivity extraneous variable (tries to make sense of situation).
*Involve participants guessing aim of study, try to please researcher by giving them results they think they want/attempt to annoy researcher, give results they don't want.
*Acting.responding unnatrually, avoid being considered abnormal. Results artificial, less useful- lack validity.
*Single blind procedures reduce impact, ensure participants don't know what condition they're in.
*Physical characteristics may influence results, appearance, age, ethnicity.
*Personal characterisitics- tone/accent- may be picked up, interpreted in a way which effects behaviour.
*May collect/interpret data in biased manner, aware of aims of study.
*Double blind procedure- conteract effects of investigators, both participant/researcher unaware of which condition they're allocated to.
*Avoided in observation- 2 independent researchers. records compared/assessed for inter-rater reliability. Film behaviours- permanent record, checked in peer review process to review/confirm ratings.
Reducing Effect of Extraneous/Confounding Variable
1.Randomisation: use of chance whenever possible, reduce researcher's influence on design of investigation. Control investigator effects, participants allocated randomly, word lists generate randomly. Repeated measures design- order conditions completed has to be randomised for each participant.
2.Standardisation: all participants exposed to same environment, information, experience. All procedures standardised- instructions. Non-standardised changes in procedure don't occur, not extraneous variables.
*Population- group of people, share given set of characteristics, researcher wishes to draw conclusion from. Research obstains sample, target population too large to allow all to be tested, expensive, impractical.
*Sample- group of participants that take part in research, sample drawn from target population, representative- can be generalised. Samples should be sufficient size- represent variety of individuals, not too large to make it uneconomical.
*Large Samples- time consuming, expensive.
*Small Samples- less chance of producing meaningful effect, sample bias greater with small groups.
*Small sample from large target population- biased.
*Every person in target population has equal chance of being selected. Researcher has list of all members in target population, random generation, select required number of participants.
+Free from researcher bias- no control over who is selected.
+Representitive- high pop. validity.
-Represented sample- not guarenteed, sub groups may be overrepresented/not selected.
-People may refuse to participate when contacted.
*Every nth member of target population selected. Sampling frame produced, list of people in population followed by sampling system to obtain sample.
+Free from resercher bias.
+Fairly representitive sample obtained.
-Not unbiased unless interval selected for sampling is chosen randomly.
*Represents composition of people in certain sub groups (strata) within target population. Population divided into sub-sets [gender], random sample taken from these. Ensures sample is representitive of population.
+Avoids researcher bias.
+Represents strata composition- generalisation.
-May lack 100% representation- individuals can't be accounted for.
*Researcher selects anyone willing to take part, asks people who is most conveinent for them to ask.
-Unrepresentitive of target population.
-Suffers from researcher bias- has control over who is selected.
*Participants put themselves forward for inclusion, initial ad in magazine/newspaper/internet asks for volunteers to take part in research. May place questionnaires, ask people to return answers.
+Minimal input from researchers.
-Atypical members of target population respond- volunteer bias, hard to generalise findings.
-Only people who see ad have chance of being selected- generalisability.
-Manstead and Semin- volunteers tend to be more sensitive to demand characteristics, more likley to comply with them. Low ext./pop. validity- type of people who volunteer are different to target population.
*Balance advantages/disadvantages of different designs. Researcher considers precise nature of experimental task, how variables will be controlled and availability of participants.
*Independent groups, matched participants, repeated measures.
*Participant selected for one condition, allocated using randomisation.
+Less chance of order effects.
+Less chance of demand characteristics.
+Same task/materials can be used in both conditions, greater control- standardisation.
-Differences between groups- individual differences.
-Need twice a many participants.
*Participants chosen for one group, participants in both groups matched for some relevant factor(s). Monozygotic twins- perfect matched pairs. Participants pre-tested to assess suitability as match.
+Less likley to be group differencs.
+Less chance of order effects.
-Time consuming to accurately match participants on all variables.
Each participant appears in both groups, same in both groups.
+No individual differences between groups.
+Fewer participants required.
-Order effects- perform better in second condition, learn someting from first condition about experimenter/task- do worse (tired/bored).
Counterbalancing- half of participants begin in experimental condition other half begin in control, swap- balance out order effects.
-Demand characteristics- guess purpose of study.
-Additional materials need to be produced.
*Primary aim of psychology- improve quality of human life- human participants, duty to respect rights/dignity of all participants.
*Moral codes laid down by professional bodies to ensure members adhere to certain standards of behaviour when conducting research.
*Follow moral principals of conduct, designed to protect participants/reputation of psychology.
*BSP- list of ethical guidelines (tells researcher what they should be doing to conduct research in ethical manner). All practicing psychologists must follow these.
Dilemma between what researcher wants to do and rights of participants.
Informed Consent: Participant always agree to take part in study, told aims of research, nature of procedure, purpose of role. Researchers can make case to withold info when they 'don't want to give the game away'.
Deception: Shouldn't withold info or mislead them about true nature of study to encourage consent/get more valid results. Sometime unavoidable, must:
- Ensure alternate procedures which avoid deception are impossible.
- Consult with colleagues about how deception may effect participants.
- Reveal deception immediately after research is completed- debrief.
Right to Withdraw: Allowed to leave at any point during study, including retrospectively (data removed from study/destroyed).
Protection from Psychological/Physical Harm: Protect participants from harm, risks no greater than exposed to in normal lifestyles. Contact investigator after research if they have worries/concerns, researcher detects/removes consequences of their research.
Privacy and Confidentiality: Keep info private, studies report won't reveal info which makes them identifiable, provide them numbers not names.
Dealing With Ethical Issues
Cost and Benefit Analysis: Assess viability of research proposal, value/ground breaking nature of research, possible cost- in terms of effect on participants or reputation to psychology.
- Issue consent letter, detail all info that may affect decision to take part in study.
- Children under 16- parent signature required.
- Preserve objective of study, obtain presumptive consent.
- Prior general consent taken to take part in numerous studies including one involving deception.
- Retrospective consent taken after participants have done study, have right to withhold data.
- Debrief given after about true aims of study, deception disclosed, explained justified. Attempts made to undo negative effects of research. Reminded of right to withdraw, told how to find out results of study. Independent groups design- informed of condition they didn't take part in.
- Role play used in situations where participants may be effected by participation.
Assessment of scientific work by other specialists in same field, ensure research intended for publication is high quality. Small group assess work before publication in journal- objectively.
*Assessors unknown to author.
*Form of self-regulation, maintain standards of quality, improve performance, provide credibility.
- Decide whether funding for propsed research is awarded.
- Validate quality/relevance, all research principles adhered to.
- Suggest improvements/amendments before publishing, withdraw if deemed inappropriate.
- Independently scrutinise research report, consider work in terms of validity, significance, originality
- Assess appropriateness of methods/design used.
- Editor of journal makes final decision on whether to accept/reject based on reviewers comments/reccomendations: Unconditionally Accept research article / Accept in event author will improve in certain ways. / Reject- encourage revision, invite resubmission. / Outright rejection.
Peer Review A03
+Prevents dissemination of irrelevant findings, unwarranted claims, unacceptable interpretations, personal fiews, fraud.
+Improves quality of research.
+Successful- improve researcher's reputation/chances of getting future further funding, institutions where research takes place gains improved reputation.
+Published research taken seriously- scrutinised.
+Increases probability of weakness/errors identified.
-Anonymity- criticise rival researchers work due to limited research funding.
-Publisher publish work that is headline grabbing, potrays journal in positive light- research not meeting criteria risks being dismissed.
-Critical of work which contredicts own mainstream theories, favourable to work that maintains status quo or supports current work- rate of change impeded in given discipline.
Implications of Psychological Research for The Eco
Attachment- Bowlby: babies need constant care of mother for healthy psychological development [stay at home mothering], later evidence shows substitute childcare doesn't have detrimental effect on social development- mum can return to work, remain economically active. Emotional care important in early development- looked after children develop healthily, productive members of society, improve world economy,
*Psychopathology- Treatment/ability to work as effective members of society, studies compare effectiveness of different types of therapies, may show people with disorder less likley to suffer relapse after CBT- more expensive than drug therapy but long-term more economically sound, less time off work.
Memory- Relate to more efficient use of public money- cognitive interview facilitates accuracy of eyewitness reporting, better use of police time/resources.
*Biopsych- Research into S/W cycle- how to be productive but safe at work (Chernobyll), when most effective time to use drugs is- use less, save NHS resources.
*SI- Research into social change- understanding of SI used to improve behaviour/benefit economy. Give up alcohol benefits health, attend work, less likley to have time off sick, doesn't drain NHS resources, less people attend A&E for alcohol related injuries- treatment.
Extent to which tests test what it claims to, does data collected give true picture of what we're studying?
*Use control group- assess if changes in DV are due to effect of IV.
*Standardise procedures- minimise impact of participant reactivity/investigator effects.
*Single/double blind techniques.
*Operationalise behaviour categories.
*Maintain anonymity- more willing to answer honestly.
Incorporate lie scale into questionnaires, control for social desirability bias.
Types of Validity
-Issues within experiment, any effects observed directly due to IV manipulation. Results internally valid if they're unaffected by confounding variables. Must be:
- No investigator effects- musn't effect results due to characteristics/biased interpretation.
- No demand characteristics-participant musn't guess purpose of research/respond to please researcher or intentionally give wrong answer.
- Standardised instructions- all given same info.
- Random sample- every member of target pop. equal chance of being selected.
-Results generalised to situations/samples outside/beyond experimental setting? Only becomes clear when study is found to apply/not to different situations- hard to test. Field/natrual experiments/natrualistic observations- high, generalised.
- Can results be generalised to others [population validity]?
- Can results be generalised to other settings [ecological validity]?
- Can results be generalised to other times [temporal validity]?