Research methods Flashcards
(160 cards)
1
Q
Experimental method: Aims
A
-Developed from thoeries
-General statments that describe the purpose of an investigation
e.g To investigate whther drinking energy drinks makes people more talkative
2
Q
Experimental method: Hypotheses
A
Directional hypotheses: People who drink water become more talkative than people who dont
Non directional hypotheses: People who drink water differ interms of talkativeness compared with people who dont
-Relationship between variabes
3
Q
Experimental method: Doing an experiment
A
e.g 2 groups of people, 10 in each
-Participants in one goup have water and the other have juice
-Then record how many words each participat says in a 5 minute period immediately after drink
4
Q
Experimental method: Deciding which hypotheses to use
A
-Directional hypotheses when a theory of the findings of previous research studies suggest a particular outcome
-Non directional hypothesses when no theory or previous reseach or earlier findings are contradictory
5
Q
Independent variable
A
-Manipulatives in an experiment
6
Q
Dependent variable
A
-Measured change or effect
7
Q
Levels of the IV
A
-Control condition e.g no water/ drink juice
-Experimental condition e.g water
8
Q
Operationalisation of variables
A
-Making variables testable
-e.g 300ml of water
9
Q
Research issues parts
A
Extraneous variables
Confounding variables
Demand characteristics
10
Q
Extraneous variables
A
Any variable other than the IV that may affect the Dependent variable if t is not controlled e.g age, lighting in lab
-Known as nuisance vairables
-Dont systematcally vary with the IV
11
Q
Confounding variables
A
-Systematically change with the IV
-Brings a second unintended IV e.g being excited or not
12
Q
Demand characteristics
A
-Participant reacivity from clues
-Please U effect : Act in a way they think is expected and over perfrom
-Screw U effect - Underperfom to sabotage results
13
Q
Investigator effects
A
Effect of investigators behaviour (conscious or unconscious) on the DV
-e.g design of the study, selectrion of and interction with participants during research process
-e.g Leading questions
14
Q
Randomisation
A
-Reduce the effect of extranous/confouding vairabes on the outcome
-Chance method to reduce the researches unconscious biases when designing an investigation
-Controls investigator effects
15
Q
Standardisation
A
-Using exactly formalised preocedures and instructions for all participants ina research study
-Use standardised instructions read to each participants so non standardised changes do not act as extraneous variables
16
Q
Participant variables
A
Any individual difference between the people taking part that may interfere that may interfere with outcome of the investigation
17
Q
Situational variables
A
Any aspect of the experimentasenvironment that may interfere with the outcome of the investigation
18
Q
Types of experimental designs
A
Independent groups
Repeated measures
Matched pairs
19
Q
Experimental design
A
How participants are arranged in relation to the different experimetnal conditions
20
Q
Independent groups
A
Participants are allocaed to different groups where each group represents one experimental condition e.g experimental or control condition
21
Q
Repeated mesures
A
-All participants take part in all conditions of the experiment
22
Q
Matched pairs design
A
Pairs of participants are first matched on some variables that may affect the dependent variable
-Then one member of the pair is assigned to Condition A and the other to condition B
-e.g matched on IQ
-Controls for the confounding vairable of participant variables
22
Q
Matched pairs design
A
Pairs of participants are first matched on some variables that may affect the dependent variable
-Then one member of the pair is assigned to Condition A and the other to condition B
-e.g matched on IQ
-Controls for the confounding vairable of participant variables
23
Q
Random allocation
A
An attempt to control for participant variables in an independent groups design which ensures each participant has the same chance of being in one condition as any other
24
Counter balancing
Attempt to control for the effects of order in a repeated measures design
-Half participants experience the conditions in one order, and the other half in the opposite order
25
Independent groups design evaluation
-Mean differences may be due to participant variables
-Acts as confounding variable, reducing the validity of the findings
-Dealt with Random allocation
-Less economical than repeated measures as each participant contributes to a single result only
-Twice as many participants would be needed to prodce equivalent data than repeated measures
-Increases time and money spent on recruiting participants
-
-Strength is that order effects are not a problem whereas they are for a repeated measures design
26
Repeated measures evaluation
-Each ppt has to do at least 2 tasks and the order may be significant
-Order effects dealt with counter balancing
-Can cause boredom that may cause deterioration in performance
-Or performance may increase through practice
-Deamand characteristics more likely the more they repeat
-Strength: Participant variables are controlled thus higher validity and fewer participant are needed, ecological
27
Matched pairs evaluation
-Participants only take part in a single condition so order effects and demand characteristics are less of a problem
-Participants can never be matched exactly as seen when identical twins used, there are still important differences in between them
-Time consuming an expensive, especially where pre-test is required so it is less econmical than other designs
28
Types of experiment
Laboratory experiments
Field experiments
Natural experiments
Quasi experiments
29
Laboratory experiments
-Highly controlled experiments
-Not always lab, just well controlled conditions
30
Laboratory experiments strengths
Strength: High control over confounding variables and extraneous variables
-Means the researcher can ensure that any effect on the dependent variable is likely to be the result of the independent variable, thus high internal validity
-Replication is also more possible than in other types of experiment because of the high level of control
-Ensures that new extranous vairables are not introduced when repeating an experiment
-Replication is used to check resuts of a study to see whether finding is valid and not just one off
31
Laboratory experiments limits
-May lack generalisability due to artifical settings unlike real life, low external validity
- Demand characcteristics
-Tasks i labs lack mundane realism
32
Field experiments
-IV manipulated in a natural , more everyday setting
33
Field experiment strength
Higher mundane realism
PArticipants not being aware may be unaware they are being studied thus high external validity
34
Field experiment limit
-Loss of control of Confounding variabes and extraneous vairablles
so hard to establish cause and effect between IV and DV and precise replication is often not possible
-Important ethical issuses due to cannot consent to being studied and may cosittute an invasion of privacy
35
Nautral experiments
-Change in Iv is studied but not manipulated or controlled by researcher and would have happened even if researcher is not there
-Researcher records the effect on a DV they have decided on
-e.g whether child is in hospital at age 5 or 10
-DV may be naturally occuring e.g exam results or may be devised by the experimenter then measuredi n the field or a lab
36
Natural experiment strength
-Provides opportunity for research that may not otherwise be undertaken for practical or ethical reasons e.g Romanian oprhans
-High external validity because they involce study of real world issyes and problems as they happen such as the effects of a natural disaster on stress levels
37
Natural experiment limit
-Rare, reducing opportunites for research
-May also limit the scope for generlisinf findings to other similar situations
-Participants not randomly allocated to experimental conditions meaning researcjer may be less sure whether the IV affected te DV
-Some conducted in lab may lack realism and demand characteristics are an issue
38
Quasi experiments
-IV based on an existing difference between people e.g age or gender
-Unlike natural, IV cannot be changed
-DV may be naturally occuring e.g exam results or may be devised by the experimenter and measured in the field or lab
39
Quasi experiment evaluation
Strength: carried under controlled conditions and shade some strengths of a lab experiment
-limit: cannot randomly allocate participants to conditions and there may be confounding variables
-Both quasi and natural experiments, the IV is not deliberately changed by the researcher and we therefore cannot claim the IV has caused any observed change
40
Raondom sampling
-Complete list of all members of the target population
-Names on list assigned a number
-Lottery method used e.g picking numbers from a aht or computer software
41
Systematic sampling
-Sampling frame of target population listed int oe.g alphabet order
-nth number
-May begin with a radomly determined statr to reduce bias
42
Stratified sampling
-Sample reflects the proportions of people in certain subgroups
-Identifying different strata that make up the population e.g race
-Randomly selct from the race
43
Opportunity sample
-Asks who ever is aruond at the time of their study e.g on the street
-Because reprasentative samples of the target population are hard to obtain
44
Volunteer sample
-Self selection
-E.g advert or raising hands
45
Random sample evaluation
-Unbiased meaning confounding or axtraneous variabes are equalaly divided, enhancing internal validity
-Difficult and time consuming e.g complete list of target population is hard to obtain
-Sample may still be unrepresantative e.g the odds of having 50 white women called julie
-Participants may refuse to take part
46
Systematic smaple evaluation
-Objective , no bias
-Time sampling and participants may refuse to tak part ,resulting in volunteer sample
47
Stratafied sampling evaluation
-Accurately reflect the compositioon of the population, thus representative
-Identified strata cannot reflect all the ways people are different, so copmplete representation of the target population is not possible
48
Opportunity sample evaluation
-Convenient and less costly in money and time because no list required and no need to divide populaton in to different strata
-Bias: unrepresentative as it is drawn from a very speevific area so findings cannot be gernerlised to the traget population
-Researcher has complete control of the slection of participants and may avod peopel they dont like te look of, researcher bias
49
Volunteer bias evaluation
-Easy, minimal imput from the resarcher so less time consuming
-Participants more engaed
-Volunteers attracts a profile of person, those who try to please the research affecting extent of generalisation
50
Types of sampling
Random
Systematic
Stratified
Opportunity
Volunteer
51
Ethical issues
Informed consesnt
Deception
Protection from harm
Privacy and confidentiality
52
Informed consent
-Aware of aims, procedure, their rights (including the right to withdrawl partway) and what their data will beu sed for
-May make study meaningless becaues behaviour will not be natural
53
Deception
Delibertely misleading of witholding infromation
-Ppt how hvae not recieved adequate information or lied to cannot say they have given infromed consenst
-Exceptoin if it does not cause distress e.g not telling there is another group drinking another drink
54
Protection from harm
Should not be placed at any risk
-Physical and psychological harm
-e.g Embarrased, stress or pressure
-To protect from harm ppts can withdraw
55
Privacy and confidentiality
Right to control infromaition about themselves
-If privacy infromation is invaded the confidentiality shoul be protected
-Confidentialty refers to our right to have any personal data protected
-Extends to the area where the study took place such that istitutions or geogrpahical loactions are not named
56
Ways of dealing with ethical issues
BPS Code of conduct
Dealing with infromed consesnt
Dealing with deception and protection from harm
Dealing wit hconfidentiality
57
BPS code of conduct
_BPS Code of ethics
-Ethical guidlines
-If breached wont be sent to prison but may lose their job
-Guildlines are implemented by ethics comittess in researchers who use a cost benefit approach to determine whether researc proposals are ethically acceptable
58
Dealing with informed consesnt
-Consesnt letter deatailing all relevent infromation
-Signed
-Ujnder 16, parental consesnt is required
59
Dealing with decepetion and protection from harm
-Debrief, mae aware of true aims of invesigation and details not supplied such as existaence of other groups of experimental conditions
-Told what their data will be used for and right to withdraw and right to withold data especially in retrospective conesnt
60
Dealing with confidentiality
-If persnal details are held they must be protected
-Usual to simply record no ersonal details i.e maintain anonyminity
--Refer to particpants with numbers of initials when writing the investigaion
-During briefing and debriiefing, ppts are reminded their data will be protected throughout the process and told data will not be shared with other researchers
61
Cost benefit analysis
Weigh up costs and benefits of research propossals to decide whether a researchh study should go ahead
Benefits e.g Valur o the research
Costs e.g Damaging effect on participants or reputation to psyhology
62
Alt Ways of getting consesnt
Presumptve consent- Similar group of people are asked if study is acceptable, if agrees then consent of orignal participants is presumed
-Prior general consesnt- Participants give their permission to take part in a number of different studies, including decpeption . By consenting , participants are consenting to being decieved
-Retrospective consent : Consents asked during debriefing
63
Pilot sutdies and aims
Snal scale trial run of the atual investigation
-Handful of participants not total number t oroad test the procedure and check the investigation runs smoothly
-Can be used for self report methods e.g questionnaired or interviews to rweord confusing questions
-In observational studies a pilot study checks coding systems before the real investigaton is undetaken
64
Single blind procedure
-Participant not told aim or if theres other conditions or what conditions theyre in
-Any infromation that migh create expectations are not reelaed until the end of the study to control confounding effects of demand characteristics
65
CDouble blind procedure
-Neither participants or researcher is aware of aims of the ivestigation
-Important for drug trials
-Researcher doesnt know which drugs are real and which are placebos
-If they dont know what each participant is recieveing then expectations cannot influence participant behaviour
66
Types of observation
Naturalistic and conrolled observations
-Covert and overt observatins
-Participant and non participant observations
67
Nauturalistic vs controlled observations
Natural: Setting or context where target behaviour would usually occur
Controlled : e.g Two way mirros
-Control of confounding/ extraneous variables
-Specifically deigned envornment
68
Covert/overt observations
CovertL Unaware of the focus of studied and ehaviour is observed in secret
-Should be in public so observations are ethical
-Overt is when participants known their behvaiouris observed, given consent
69
Participant and non participant observations
Particpant : observer becomes part o f the research
Non participant: Researche rremains seperate
70
All observations evaluation
-Capture the reality of what people do including unexpected behaviour
-People may act unatually
-Observer bias = Observers interpertrations may be affected by their expectations . Reduced by using more than 1 observer
-Cannot demonstrate casual relatioships though it may be used in experiments in detecting cause and effect relationships
71
Natural vs controlled observtions evaluation
-Natural = high external validity and finings often generalised to everyday life
-But lack of control means replication is difficult and uncontrolled confounding/ extraneous variables
-Cootrolled= findings cant generlise to real life
but cofounding/ extraneous variables lesso f a factor and so replication of the observation becomes easier
72
Covert vs overt evlautaion
-Covert- demand characteristics removed, icnreasing internal validity
-Ethics questioned
-Overt= more ethical but demand hcaracteristics
73
Participant vs non participant observations
Particiapant = Researcher gets increased insight of lives through experience, increasing external validity
-But may identify to strongly and lose objectivity
Non participant = Objective psychological distant so less danger of adopting a local lifesyle
-May lose valuable insight to be gained in a participant observation as too ffar removed from people and behaviour htey are studying
74
Inter observer reliabiltiy
- 1 researcher = bias
-May miss importnat details or only notice events that confirm their opinions or hypothesis
-Observers familairise themselved with the behavioural categroies to be used
-They then observe the same behaviour at the same time, e.g part of the small scale pilot stud y
-Observers compare the data they have recorded and discuss any differences in interpretations
-Then observers analyse data from the study
-Interobserver reliability is caluculated bycorrelting each pair of observations made and an overall fugure is prduced
75
Issues in observational design
Ways of recording data
Behavioura categories
Sampling methods
76
Ways of recrodig data
Unstructured observation : Writing everything they see down, rich in data
Used when observations are smal in scale and have few participants
-However may be too much ging on so must simplify target behaviours using behavioural categories thus A Structured observation
77
Behavioural categories
-Necessary to break the target behvaiour up into a set of behavioural categores or behaviour check list
-Similar to operationalisation
-Behaviours must be observable
78
Sampling methods
Contunious recording of behaviour is a key feature of unstructured observations in which all instances of a target behaviour are recorded
-For complex behaviorus this is not parctical
-Event sampling = Counting the number of times a particular behaviour occurs in a target individual or group
-Time sampling = Recording behaviour within a pre established time frame
79
Structures vs unsctructured evaluation
Structured = Behavioural cateories make the reording of data easier and more systematic
-Data produced is numerical thus qunatitative meaning analysing and comparing behaviour is more easy
-Unstructured observations produce qualittative data which may be more difficult to record and analyse
Unstructured however produce more richness and eth of detail
-Thoug there is risk of observer bias because may only record behvaiours that catch their eye and may not be importnant or useful
80
Behavioural categories
Categories must be observable, measurable and self evident
-Shoud not be a dustbin category where different behaviours are deposited
-Categories should be exclusive and not overlap e.g smiling/grinning
81
Sampling methods observation
Event sampling is useful when target behvaiour is infrequent an could be missed if time sampling ws uesd. But if specified event is too complex, the observer may overlook important details
-Time smpling is effectie in reducing the number of observations that have to be made
-Thus data may be unrepresantaive of the observation as a whole
82
Self report techniques
Questionnaires and inerviews
83
Types of questionnaires
Open - No fixed range of answers , qualitative data
Cosed - fixed number of responses
Quantitative e.g yes/no / 1-10
-Easy to analyse but lacks depth and detail
-Closed that produce quantitative data can be oncverted into qunaitiative e.g amounts of yeses
84
Types of interviews
Structured interviews
Unstructured interviews
Semi-structured interviews
85
Structured interviews
-Predeterimed set of questions
86
Unstructured interviews
-No set of questions
-General aims and topic discussed and interaction is free flowing
-Encouraged to expand and elborate on their answers
87
Semi stuctured interviews
-List of questions but inervieweres are free to ask follow up questions based on revious answers
88
Questionnaires evluation
-Cost effective , large amounts of data as can be distribute to large numbers of people, can be completed without researchers present
-Data is straightforward ot analuse especially with fixed choice clsoed questions
-Data leads to statistical analysis
-May not be truthful
-Demand characteristic called social desirability bias
-Response bias e.g always yes or favoured end of a rating scale
-Becaus too quickly completed and fail to read questions properly
-Acquisience bias - yes bias
89
Interviews evaluation
Structured = Straiggjtforawrd to replicate and reduces differences inbetwen interviewers
-Cannot deviate from the topic or explain their questions and this will limit richness of data and limit unepected information
Unstructured= More flexible
but increases interviewer bias
-Not straugthforward, may habe to sift trhough much irrelavant infromation and drawing firm conclusions may be difficult
-May lie for social desirability but a skilled interviewer should establish sufficient rapport with ppt so when sensitive nad peronal topics are discussed ,repsonses are mroe truthful
90
Closed questionnaires types
-Likert scales : 5 points for agreement e.g 1 strongly agree, 5 strongly disagree
-Rating scales : Same thing but feelings e.g enetertaining
-Fixed choice option e.g tick all those that apply that apply to them e.g For what reasons do you watch tv , options are entertainment, to escape , to be frightneded
91
Designing interviews
-List of questions the interviewer tends to cover, shold be standardised to reduce contaminating effect of interviewer bias
-Intrviwer mmay take notes or interview is recorded and analysed later
-Group inerviews in clinical settings
-ONe to one should be quiet room and no other people to increase honesty
-Should begin with neurtal questions to make interviewee feel relaced and comfortable and establishin rapport
-Should be reminded several times their answers will have stict confidentiality, importnant for personal topics
92
Writing good questions parts
Overuse of jargon
Emotive language and leading questions
Double barrelled questions and double negatives
93
Overuse of jargon
Jargon = Techincal terms only familiar to hose within a specialised field or area
-Too complex for other people so must be simple and easily understood questions
94
Emotive language and leading questions
-Specific emotions should be replaced with neutral alternatives
-Leading questions such as 'sint it obvious' or 'when did you last breake the speed limit' , assuming ppt has done so
95
Doube barrelled questions and double negatives
Double barrelled =2 questions in one, where repsondents may agree with one half and not the other
e.g Teachers are underpaid and should have to strike
-Double ngatives e.g I am not unhappy in my job - agree or disagree
-Difficult to decipher
96
Correlations
Illustrat the strength and direction of an association between two or more co variables (things being measured)
-Correlations plotted on a scatergram
-One co variable is represented on the x axis and the other on y axis
97
Types of correlation
Positive correlation
Negative correlation
Zero correlation
98
Difference between correlations and experiments
In an experiment the researcher controls or manipulates the independent variable to measure effect of dependent variabe
But in correlations there is no manipulation of one variable and not possible to establish cause and effect . Despite positive correlation between caffeine and anxiety level, we cannto assume caffiene was the cause of anxiety
99
Correlation strengths
-Quantifiabe measure of how 2 variables are related
-May syggest ideas for future research if variables are strongly related or demonstrate an interesting pattern
-Starting point to assess possible patterns between variables before researchers commit to an experimetnal study
-Quick and economical to carry out , no need for a controlld environment and no manipulation of variabes
-Data can be secondary meaning less time consuming than experiments
100
Correlations limit
Lack of experimental manipulation and control means studies cna only tell us how variables are related but no why
-No cause and effect, so we do not know which co variable is causing the other to change so lack of direction
-Intervening variable/ third variable causes relation between 2 variables tested e.g high pressured jobs make anxiety and also needing to work long hours to remain alert
-Relatoonships sometimes presented casual when they arent especially by media
-People from broken homes likely to become criminals
-But 3rd variable of poverty causes broken home but also crimes
101
Primary data
Original data collected specifically for the purpose
-First hand data from aparticipants themselves
e.g data from expriment, questionnaire or observation
102
Secondary data
-Collecetd by someone else than the person conducting research
-Data that already exists
-Subject to statisical testing so significance is known
-Sexondary data includes data located in journal articles, books or websites
103
Meta analysis
-The process of combining findings on a particular topic , same aims /hypothessis
-Creates a conclusion
-Aim is to produce an oerall statistical conclusions (effect size) based no a range of studies
-Not a review
-Prone to publish bia ,leaving out not relevant stuies or with negative results
104
Qualitative data evaluation
-Rich detail, can fully report thoughts feelings and opinions, external validity
-Difficult toanalyse, not summarised statistically so patterns and compairsons between data are hard o identify
-Coclusions rely on the subjective interpretations of the researcher and these may be subject to bias , where researcher has preconceptions about what he/she is expecting to find
105
Quantitative data evaluation
-Simple to analyse so comarisons between groups are easily drawn
-More objective and less open to bias
-but more narrow in meaning and less detail, may ail to represent real life
106
Primary data evaluation
-Data rightfully recieved from main target infromation the researcher requires
-However takes time and effort
-Planning prepreations an dresources
107
Secondary data evaluations
-Inexpensive, minimal effort, may find desired information already exists so no need to conduct parimary dat collection
-Data may seem valuable but is acrualy outdated or incmplete
-Data may not match the researchers needs or objectives ,challengign validity of conclusions
108
Mean
Average
-Representative as whole
-However easily disotrted by extreme values
109
Median
Middle
-EXtreme scores dont affect it and is easy
-But extreme values are ignored and higher and lower ends are ignored
110
Mode
-Most frequnt score
-May have 2 modes (bi-modal)
-Several modes is not useful
-USed for data in categories alot
111
Measures of dispersion
Range
Standard deviation
112
Range
Highest value minus lowest value and adding 1
-Easy to calculate
-Only takes in 2 most extreme valuses and may be unreprasensttive
-Influenced by outliers
113
Standard deviation
-Single value that tells us how far scores deviate from the mean
-The larger the standard deviation the greater the spread within a set of data
-If it is large in an experiment it suggests not all ppts wer eaffected by the Iv in the same way because data is widely spread . May be a few anomalous results
-Low SD reflects data is clustered around the mean meaning ll participants responded in a familiar way
-More precise than range as it include al lvalues within the final calucultion
-Howver can be disorted, like the mean, with a signle value
-Also extreme valuses may not be revealed unlike with the range
114
Presentation of quantitative data
Table / summary table
Bar charts- When data divided into categories known as discrete data
Histograms - Bars touch eacjother showing x axis data is continuous rather than discrete
-Scatter grams for associaios btween co variables
-Line graphs for continous data , IV
115
Types of distribution
Normal distributoions
Skewed distributions
116
Normal distribtuons
- Bell curve
-Most people located in the middle area of the cruve with very few people at the extreme ends
-Mean median and mode all at the midpoint
-Tails never touch the x axis as more extreme scores are always possible
117
Skewed distributuon
-Positive skew -
-Negtive skew
118
Positive skew
eg hard test
119
Negative skew
120
Statsitcal test
Determining whether hypotheses should be accepted or rejected
-By using a statistical test we can find out whether differences or relationships are significant or likely to occur by chance
121
The sign test
A statistical test used to analyse the difference in scores between related items. Data should be nominal
122
Peer review role
-BEfore a piece of research can be a part of a journal, it must be subject to a process of peer review
-Involves all aspects of the written investigation being scrutinised by a small group of experts in the paritcular feeled
-Experts hsould conduct an objective review and be unknown to author or researcher
123
Peer reviews aims
1) To allocate research funding
Whether or not to award funding for a proposed research project
2) To validate the quality and relevance of research
Assessed for quality and accuracy, the formulation of hypotheses, the methodology chosen, the statistical tests and the conclusions drawn
3) To suggest ammendments or imporvments
Rwviweres may suggest minor revisions of the work and therefore improve the repott or may conclude the work is innaporoprate for publication and should be withdrawn
124
Peer review evaluation
Anonymity used to crticise rival researchers because researchers are in direct competition for limtied research funding
-This is why some journalsd favour a system of open reviewing where names of reviewers are made public
Publication bias: They want to publish headline grabbing findigns to increase credibility of their publication and prefer to publish positive ones (flile ddrawer problem)
-MEans research that dont meet this criteris is ignored , creating a false impression of the current state of psychology
-Buring Groundbreaking research
Supressing opposition to mainstream theories wishing to maintain the status quo in particular scientific fields
-Reviewers tend to be critical of researcch that contradicts their own view and more favourable t that which matches it
-Established scientists are more likely to be chosen as reviewers meaning new innovative research that challenges the established order are more likely to be used than correlating findings
-Peer review may have the effect of slowing down the rate of chnage within a partiular scientific discipline
125
Implications of psychologyical research for the economy
Attachment in the role of father , modern families, mother is higher earner and so works longer hours. Modern parents are bttwer equipped to maximise their income and contribute more ffectively to the economy
The development of treatments for mental disorders
Absence of work lost 15 billion quid a year .1/3 of absences due to mental health disorders
Treatmeant of these means these people can return to work
126
Descriptive statistic
Graphs, tables and summary statistics e.g measures of central tendency and measures of dispersion
-Used to identify trends and analyse sets of data
127
Inferental statistic
-USe fo statistical tests which tell researchers whether the differences or relationships they hav efound are statistically signifcant or not
-Help decide which hypothesis to accept and which to reject
-A creelantion coefficint is calucualted using a statistical test and is an iferental statistic
128
Correlation coefficient
A number between -1 and and +1 that represents the direction and strength of a relation ship between 2 co variables
129
Case studies
-In depth detailed analysis of an individual,group, insituton or event
-Analysis of unusal individuals or events e/g rare sidorder
-cases tudies may focus on more typical cases such as an elderly persons recollections of their chidhoods
-Conducting a case involeves qualitative data
-May construct a case history using interviews, observations, questionnaires, or a combination
-Person may even be subject to experimental or psychological testing to assess what they are or not capable of, and this may produce quantitative data
-Case studies tend to be longitudinal and may involve gathering more data from family and friends of the individual aswell as the person themselves
130
Case studies evaluation
Strength- offer rich detailed insights that shed light on unusual and atypical forms of behvaiour
-Preferred ti nire superficial forms of data such as experiment or questionnaires
-Also understand of 'typical functioning' e.g HM
-Generate hypotheses for future study and one contradictory instance may lead to the revision of an entire heory
-
-Generalisation hard with small samples
-The infromation that makes into final report is based on subjective and interpratation of researcher
-Personal accounts from participants mya be innacurate and have moemory decal especially wiht chidhood storiies
-Means evidence from case studies have some low validity
131
Content analysis
Type of observational research where people are studied indirectly via communications they have produced
-e.g conversation or speech/rpesentation, written froms like texts or emails or books magazines, tv prgromams or films
-Aim is to summarise and desrcibe this communication in a systematic way so overall conclusions can be drawn
132
Content analysis parts
Coding & Quantitiative data
Thematic analysis & Qualitative data
133
Content anaysis : Coding and quantitative data
Coding is the inital stage of content analysis
-Some data sets may be very large e.g transcripts of several dozen lengthy interviews
-They must categorise this infromation into meaningful units
-e.g Couting the number of times a word or phrase in the text to produce quantitative data
-e.g News paper reports analysed number of time derogatory terms for people with mental issyes are used e.g crazy or mad
-e.g TV adverts examined to see how often men and women are depicted in professional jobs or familial roles
134
Content analysis: Thematic analysis and qualitative data
-Identification of themes
-Theme- explicit or implicit idea that is recurrent
-More descriptive than coding units
-Mental healht issues may be representated in newspapers as a threat to well being of children ir as a 'drain of resources of the nhs'
-Such themes may be developed into broader categories such as 'control' or 'sterotyping' or 'treatmnet' of people with mental health issyes
-Once themes developed cover nost aspects of the data they are analysing, they may collect a new set of data to test the validity of themes and categories
-Assuming these explain the new data adequately, the researcher will write up the final report, typically using direct quotes from the data to illustrate each time
135
Content analysis evaluation
-Can deal with ethical issues associated with psychological research
-Material already exists in public doman, no ssues obtaining permission
-Have high external validity and may access data of a asensitie nature provided authors cnosent to use
-Flexible due to qualitative and quantitative data
_People studied indirectly so communications are analysed outsdie of the contect which it occurre
-There is a danger that the researcher may attribute opinions and motivations to the speaker or writer that were not intended originally
-Modern anlysts are clear about how ther own biases and preconceptions influence the research process and often make referecne to these as part of their fina report
-However may still lack objectivity especially when more decriptive forms of thematic anayssis are employed
136
Reliability
A measure of consistnecy
-If particular measuremnt is made twice and produces the same result then that measurement is desrcibed as reliable
-Must not be a change of measurement
137
Ways of assessing reliability
Test - retest
Inter-observer reliabiltiy
138
Test -retest
-Administernig same test or questionnaire to the same person or people on different occasions
-If reliable then results should be same or similar
-Commonly used with quesitionairres and psychological tests such as IQ and interviews
-Msut be suffiecicent time between test an retest e.g survery to forget questions , but not too long that attituteds change
-In questionaire or test, 2 sets of scores could be correlated to make sure they are similar
-If correlation is significant and positive then the reliability of the measuring instrument is assumed to be good
139
Inter-observer reliabiltiy
-Observational research open to subjectivity and bias and unreliabiltiy
-Should conduct observations in at least 2 people
-Inter observer reliability must be established
-Involves pilot study of the observation to ensure observers are pplying behvaioural categories in the same way , or comparison repoted at the end of a study
-Data collected by 2 observers should be correllated to assess its reliability
-Applies to Content analysis: Inter-rater reliability
aswell as interviews (inter -interviewer reliability _
140
Measuring reliabiltity
Measured using correlational analysis
-In test-retest and inter-observer reliability the 2 sets of scores are correlated
-The correlation coefficient should exceed +,80 for reliability
141
Improving reliability in questionnaires
-Test retest
-Should be a corellation that exceeds +.80
-Low test retest reliabiltiy may require some items to be deselected or rewritten
-eg replacing open questions with closed
142
Imroving reliabiltiy in interviews
-Use same interviewer each time
-If not they must be properly trained
-Not asking too leading or ambiguous questions
-Structured interviews
143
Improving reliabiltiy in observations
-Making sure behavioural categories have been operationalised and they are measurable and self evident
-If not then observers will differ in results and inconsistnet
-Reliability is low then observers may need furhter training
144
Improving reliability in experiments
Standardised procedures
145
Ecological validity
Type of external validity - generlising the findings to other settings - to everyday life
-e.g list of words tested
146
Temporal validity
Whether findgins hold true over time
147
Ways of assesssing reliabiltiy
Face validity
Concurrent validity
148
Face validity
Whether a test,scale or measure measures what its suposed to measure
-can simply eyeball or passing to an expert to check
149
Concurrent validity
When results obtained are close to or match those obtained on nother well established test
150
Improving validity in questionnaires
Lie scale within the questions to assess the consistency of a respondents response and to control for the effects of social desirability bias
- also assuring data will remain anonymous
151
Improving validity in observations
Overt not covert
making behavioural categories not broad
152
Improving validity in qualitative research
Qualitative have higher ecological validity because of detail associated with case studies and interviews
-Must demonstrate the interpretive validity of their conclusions : extent to which the researchers interpretation of events matches participants
-Demonstrated through coherence of the researchers narrative and the inclusion of direct quotes from participants
-Validity enhanced through triangulation - the use of a number of different sources as evidence e.g data from interview with friends family personal diaries , observations etc
153
Type I error
When null hypthesis is rejcted but it should have been accepted
-Happens when significance level it too lenient e.g 10% rather than 5%
154
Type II error
When null hypothesis is accepted but shoudlve been rejected
-When significant levels are too stringent e.g 1% rather than 5%
155
5 features of science
Paradigms & paradigms shift
Theory construction & Hypothesis testing
Falsifiability
Replicibality
Objectivity and the merpical method
156
Paradigms & paradigms shift
-Kuhn suggestets paradigms distinguish between scientific disciplines and non scientific disciplines
-Social sciences lack a universally accepted paradigm, best seen as pre science
-Psychology has too much internal disagreement to be ascience
-Paragdigms shift - change of accepted paradigm
157
Theory construction and hypothesis testing
Theory construction = gathering evidence from direct observation
-Hypothesis testing - systematic and objective methods to deterimine support
Devising a new theory from an existing is deduction
158
Falsifiability
- Proposed by Popper - argued that the key criterion of theory is its ability to hold up for hypothesis testing and the chance to be proven false
-Thoery of falsification - retained possibility of being proven false
- Good sciences can be falsified and constantly challenged whereas pseudosciences cannot be falsified - strongest theories withstand falsification and challenges
- Hypotheses should come as both null and alternative
159
Objectivity and the empirical method
- Researchers must maintain objectivity i.e. maintain a critical distance - Personal opinions shouldn't affect data or behaviour of ppts
- Lab experiments tend to have the most control so are the most objective
- Objectivity is basis of empirical method - theory cannot be scientific if it hasn't been empirically tested and verified
