Research Methods

Independent Variable: researcher manipulates to determin a direct effect.
Dependent Variable: affected by changes in IV and researcher measures effect of IV on it. Reasearcher creates conditions to see the effect, other conditions are experimental condition.
Extraneous Variable: other than DV that affect IV - must remove them to stop them coming
Confounding Variables: other than IV that influence the DV

Lab exp: carried out in high controlled environment, eliminate extraneous v, IV's manipulated. +: high level of control, easy to determin cause and effect, can be replicated very easily -: artificial so lacks MR and EV, chance of demand characteristics

Field exp: carried out in real world or natural setting, IV still manipulated for effect on DV
+: More MR & EV, less demand characteristics, cause and effect is established aswell        
-: Harder to replicate, less control over extraneous variables, less control of sample

Natural exp: researcher takes advantage of naturall IV. +: High EV and useful when impossible to replicate  -: Less control over extraneous variables, harder to replicate, cant determin cause and effect

Correlational Anaylsis: A technique for analysing data which measures relationship between two quantitative variables

Positive: As one increases, other increases       Negative: As one increases, other decreases No Correlation: Variables have no effect on each other

+1 = Strong positive         -1: Strong negative        0=No correlation

Adv: establish strength of relationship, allows researchers to investigate things that couldnt be manipulated for ethical reasons and predictions can be made about other variables 

Disadv: doesnt demonstrate cause and effect, doesnt measure curvilinear.

Ethics: No informed consent, violate right to withdraw

Non participant: researcher doesnt get involved with interactions of participants.        Participant: researcher directly gets involved with interactions of participant.                      Covert: psycologists are undercover and group dont know theyre being watched              Overt: psychologists reveal true identity so demand characteristics can occur
Natural: researcher observes in own environ, no manipulation variables                                 
+: High EV, less demand if covert, rich info, good for manipulating variable when unethical   
-: Cant control extran variables, observer bias and cant determin cause of behaviour           Ethics: private, confidentiality, withdrawal

Controlled: researcher observes in controlled environ so is manipulation of variables           
+: Cause and effect infered, extraneous variable control, rich complex data obtained
-: Low EV, demand characteristics, observer bias                                                                  Ethics: informed consent, right to withdraw, confidentiallity

Observer bias: observers know purpose of the study so they see what they want to see. Have 2 observers to correlate = inter-rater reliability

Research methods where participants give info about selves without interuption by reseacher

Interview: questions are asked face to face.        Structured: all participants asked same questions in the same order = quanitative           Unstructured: informal in depth convo exchange = qualitative                                         Semi Structure: Half half                           ADV: Struct: less interviewer effects, less training.   Unstruct: rich in detail, can expand       DISADV: Struct: cant follow up answers, formal sit may inhibit honest answers    Unstruct: increase interviwer effects, training needed and hard to analyse answers                         Ethics: Privacy, confidentiality, withdrawal, consent

Questionnaires: written set of questions to answer, on attitudes, opinions and behaviour  Closed: requrire a fixed response answer = quantitiative     Open: require a expanded and open ended response = qualitative                                                                                       ADV: quick, easy, cheap to collect, analyse and replicate, increase reliability, time efficient DISADV: low response, social desirability, misunderstand, biased sample, limit depth/analysis Ethics: Privact, confidentiallity, consent

In depth detailed investigation of individual, biological, behavioural data. Give great detail.

ADV:

• lots of rich detail in one CS than loads of studies - Genie
• MR and EV is high
• Allows us to investigate behaviour thats unethical and impractical to replicate
• Influential in proving theories and suggesting hypothesis

DISADV:

• Cant generalise to ethnicity, age, gender as specific to one individual - Genire and Tom
• Hard to replicate due to circiumstances
• Researcher bias - Freud sex
• Based on subjective info where small data is reported

Aim: first step! A precise statement about purpose of study and what it intends to find out

Operationalise variables: second step, chose IV & DV! Making variables easy to measure.

Hypothesis; third step! Precise statement about expected outcome of investigation. If difference in DV and IV its due to chance.

Null: IV has no effect on DV     Alternative: IV will have an effect on the DV.    Splits into two: Non-directional(2tail): doesnt state direction of differences e.g eat choco will affect mood Directional(1tail): States direction of difference e.g eat choco will significantly rise mood

Pilot Study: small scale preliminary studies which make it possible to trail procedues so that you can make a chance before spending time and money on large scale

They are how participants have been organised

Independent Groups: diff participants used in each condition, randomly allocated to remove individual differences                                                                                                             ADV: order effects wont occur, demand characterisitcs reduced, save time                               DISADV: more participants needed, different results can be due to individual differences

Repeated Measure: Each participant is tested in all the conditions of experiment               ADV: Same no people so no indiv differences, less participants needed                                   DISADV: demand characterisitcs, time consuming, order effects

Matched Pairs: same participants used in all conditions with repeat measure but participants in groups are matched on characteristics so same gender                                                       ADV: less order effects, less demand characteristic, less individual differences                        DISADV: twice as many more participants, time consuming, hard process

Sample: representative of target pop from which its drawn and have same characteristics.

Target Pop: group who researchers are studying and want to generalise results to

Random: every member of pop has equal chance of being selected so its unbiased          ADV: samples likely to be representative so can generalise                                                 DISADV: hard to get full details of target pop, not all members want take part making it biased

Opportunity: selecting participants who are readily avaliable and willing to take part            ADV: easiest and most practical for larger samples                                                              DISADV: sample wont be representative of target, some feel obliged to take part(ETHICS)

Volunteer: people self select to participate, they advertise for participants        
ADV: good for specific areas of research                                                                            DISADV: unrepresentative giving volunteer bias

Reliability: Consistency of results from investigation. Replication = same results, is reliable Internal: test is consistent within itself  External: test is consistent over time (re-test)

Validity: measuring what it intends to measure                                                          Internal: outcome of a study is a direct result of manipulation of IV                                  External(Ecological): Extent to where findings of study can be generalised

Control of Extraneous:to prevent coming confounding which make study lack validity    Participant: characteristics of participant which affect DV                                                Situational: Factors in environment where study is conducted which affect DV                   Experimenter: Factors to do with experiment which affect DV

Demand characteristics:studies enable participants to guess purpose and whats expected. It leads to right or wrong results to annoy or please the researcher. Cause social desirability. Overcome by singleblind technique where dont know hypothesis/condition.

Investigator Effects: Investigators inadvertly influence results or cause bias interpretation. Use double blind technique, no one knows hypothesis or the conditions

Ethics must be considered before research is conducted. British Psychological Society published a code of ethics.

Informed Consent: inform participants about objectives for consent. Parents give for <16. Deception: Witholding of info or misleading participants is unacceptable if they show unease when know theyre deceived. Intentional should avoid unless scient/medical reason. Presumptive: consent gained from similar background so asume others are willing. Retrospective: asking for consent after participanting, may not consent despite participating Prior general: participants agree to be deceived without knowing when it occur, affect behave If participants are deceieved must told straight after and give the opportunity to withdraw data. Determin how to avoid it, make sure sufficient info given &consult withdraw. Debrief: explained what study was about, should be left differently to the state they started. Protect: protect from physical and psychological harm, no greater than in real life                Withdrawal: Made aware they can leave whenever, hard in covert and can remove data Confidentiality: data not disclosed, no/intials used, anonymity when no name can traced. Advice: Investigators obtain evidence of harm that participants unaware of, must tell Colleagues: If colleague not in line with BPS must say something

Naturalistic: consider behavioural categories(specific behavioural caterogies defined at start), recording data, pilot study, avoid oberserver bias and adress ethics. You record the data visual video or photos or audio recordings or spot note taking. Pilot study is used to establish behavioural categories and iron problems eg camera placement. Oberserver bias avoid by clear categories, more than one observer or doubleblind technique. Ethics are privacy, consent, confidence & withdraw.Must happen when expected.

Questionnaire: consider qualitative/quantitative, how to word q's, no and layout of q's, sample, ethics, pilot study. Qualitative is open ended and quantitative is closed. Scales can be used and average is calculated. Q's should look professional and have consice instructions. Should be clear and easily understood, not leading, double barreled or emotive. Start with simple q's and move on to more probing q's. Must be interesting for motivation, short. Ethics are withdraw and confidence. Pilot gives idea of unclear q's.

Interviews: need to consider structure, sample, recording data, wording, interviwer and ethics. Structure or unstructure. Sample must be representative of target pop and experimental design. Record by notes, tape or video. Must be clear and not leading, misunderstanding can be corrected. Consider interviewer, they must know when to talk and non verbal communication is important. Ethics are withdraw and confidential.

• Nominal Data: data in mutually exclusive catergories. Its discrete. e.g. name and school. Ordinal: data in rank order but with unequal intervals. Interval: ordinal with equal interval.
• Bar chart: nominal, vertical bars with space inbetween. Bar is diff catergory and in any order. Histogram: ordinal and interval. Vertical lines of equal width with no space inbetween. Single or data group values are used. Frequency is the area of bar. Frequency Polygon: alternative to histogrm, shows two sets of data. Scattergram: relationship between two variables. One on y and one on x axis. Shows positive, negative or no trend, closer points stronger relation.
• Mean: adding all scores and dividing by number. Most accurate and takes into account all data. Can be distorted by outliers and may not be an actual value in the data set. 
• Median: ranking scores in order and using mid value. If even its mid of the two in the middle. Used on ordinal, interval, easy to calculate and unaffected by extreme. Not all scores used, unrepresentative if clustered in one area. Mode: most frequent value. Easy to calculate and unaffected by extreme, nominal data. But tells us no info on others scores and not good if more than one mode or small sets. Range: high-low. Good as quick and easy buts its disorted by outliers. IQR: Spread of middle 50%. Quick, easy, not affect by outlier but innacurate if extreme. SD:average value scores differ from mean. All scores, hard to calculate, interval only
• Conent Analysis: systematic research technique to analyse qualitiative data. Researchers decide categories and create an encoding system, this helps record a tall of categories. Pilot study and interraterreliability looked for. Adv: rich detail info, valuable insight   Disadv: bias and hard to analyse