Research Methods Flashcards
Experimental Method
=manipulation of IV to measure effect on DV
- Aim=general statement of what the researcher intends to investigate
- Hypothesis=clear, precise, testable statement that states relationship between variables to be investigated
=Directional- positive/negative direction
=Non-directional- states is a relationship but not specific - Operationalisation=clearly defining variables in terms of how they can be measured
Research Issues
- Extraneous variables=any variable, other than IV, that may affect DV
- Confounding variables=kind of EV but varies systematically with IV
- Demand characteristics=any cue from researcher or situation that may reveal true aims of study to participants who will manipulate their behaviour accordingly
- Investigator effects=any effect of investigator’s behaviour on research outcome
- Randomisation=use of chance methods to control for effects of bias when designing materials and deciding orders
- Standardisation=using exactly same formalised procedures and instructions for all participants
Experimental Designs- AO1
- Independent groups;
=participants allocated to different groups where each group represents one experimental condition
Random allocation=control participant variables - Repeated measures:
=all participants take part in all conditions of experiment
Counterbalancing=control order effects when half experience one condition and vice versa - Matched pairs:
=pairs of participants matched on variable that may affect DV- one member assigned to A and B
Experimental Designs- AO3
- Independent groups:
~ participant variables=reduces validity- use random allocation to combat
~ less economical than repeated measures
+ order effects not a problem - Repeated measures:
~ order effects- use counterbalancing to combat
~ boredom / deterioration in performance
~ demand characteristics more likely
+ participant variables controlled=higher validity
+ fewer participants needed so more cost-effective - Matched pairs:
+ order effects and demand characteristics less of a problem
~ participants can never be matched exactly
~ matching time-consuming and expensive
Types of Experiment- Laboratory
- Conducted in highly controlled environment- manipulates IV and records effect on DV
+ high control confounding and extraneous variables=high internal validity
+ reliable as replication more possible due to high control
~ artificial so lacks generalisability=low external validity
~ demand characteristics more likely
~ low mundane realism
Types of Experiment- Field
- Takes place in natural setting within which IV manipulated and effect on DV recorded
+ high mundane realism as environment more natural=high external validity
~ lack of control of CV’s and EV’s=cause and effect more difficult to establish
~ ethical issues as consent and privacy
Types of Experiment- Natural
- Change in IV not brought about by researcher but would have happened anyway and effect on DV recorded
+ provide opportunity for research not otherwise undertaken
+ high external validity as real-world issues as evolve and happen
~ reduced opportunity for research=limit scope for generalisability
~ participants not randomly allocated
Types of Experiment- Quasi
- IV not determined by anyone- variables simply exist
+ controlled conditions=reliable
~ cannot randomly allocate participants=confounding variables
~ cannot claim IV caused any observed change
Sampling- Overview
- Population=group of people focus of researcher’s interest, from which smaller sample drawn
- Sample=group people take part research investigation grawn from target population=representative=generalisable
Sampling- Random
=all members target population equal chance being selected- ordered list of target population and lottery method used
+ potentially unbiased- EV’s and CV’s divided equally=high internal validity
~ difficult and time-consuming to conduct
~ unrepresentative sample
~ selected participants refuse to take part
Sampling- Systematic
=every nth member of target population selected- sampling frame produced and sampling system nominated
+ objective- no influence over who is chosen
~ time-consuming and participants may refuse to take part
Sampling- Stratified
=composition of sample reflects proportions of people in subgroups within target population- identify strata/proportions worked out/selected using random sampling
+ representative as reflects composition of population=generalisable
~ strata cannot reflect all ways people different so complete representation not possible
Sampling- Opportunity
=selecting anyone who happens to be willing and available at the time
+ convenient and cost-effective
~ unrepresentative of target population=not generalisable
~ researcher bias as control over participants
Sampling- Volunteer
=participants select themselves to be part of the sample
+ easy and less time-consuming
+ participants are eager and willing as it was their choice
~ volunteer bias- attracts certain demographic of individual
Ethical Issues
- Informed consent=making participants aware of aims of research, procedures, rights and what data used for- participants judgement without feeling obliged
- Deception=deliberately misleading or withholding information from participants during study
- Protection from harm=should not be placed at any more risk than would experience in day-to-day lives
- Privacy and confidentiality=right to control information about themselves
Ways of Dealing with Ethical Issues
- Informed consent:
consent letter detailing all info affect decision to participate
=presumptive consent- similar group give consent
=prior general consent- give consent to number of studies include deception - Deception and protection from harm:
Full debrief- true aims of study
=what data used for and reminded of right to withdraw
=offer advise regarding counselling - Confidentiality:
Maintain anonymity- numbers/initials
=reminded of protection of data throughout process and will not be shared
Pilot Studies and Single/Double-Blind
- Pilot study=small-scale version of investigation takes place before real one conducted- check everything works and allow any changes if necessary
- Single-blind procedure=participants not known aim of research or what condition they’re in
=not revealed until the end to control for CV’s and demand characteristics - Double-blind procedure=neither participants nor researcher knows aims of study or conditions
=used in drug trials with placebos - Control groups and conditions:
=experimental for control and comparison
Observational Techniques- Naturalistic and Controlled
- Naturalistic obs=watching and recording behaviour in the setting within which it would normally occur
- Controlled obs=watching and recording behaviour in a structured environment
=strange situation - Evaluation:
+ special insight into behaviour
+ naturalistic=high external validity as findings generalised
+ controlled=reliable due to control
~ naturalistic=low reliability as lack of control
~ potential CV’s and EV’s
~ controlled=not generalisable
~ observer bias
Observational Techniques- Covert and Overt
- Covert obs=participants behaviour is watched and recorded without their knowledge or consent
- Overt obs=participants behaviour is watched and recorded with their knowledge and consent
- Evaluation:
+ covert=removes demand characteristics increasing internal validity
+ overt=more ethical
~ covert=less ethical
~ overt=demand characteristics
Observational Techniques- Participant and Non-Participant
- Participant obs=researcher becomes member of group whose behaviour they are watching/recording
- Non-participant obs=researcher remains outside of group whose behaviour they are watching/recording
- Evaluation:
+ participant=increased insight increasing external validity
+ non-participant=maintain objective psychological distance
~ participant=lose objectivity as identify too strongly
~ non-participant=lose valuable insight
Observational Design- Ways of Recording Data
- Unstructured observation=write down everything that they see
=rich in detail
=small-scale with few participants - Structured observation=simplify target behaviours to be focus using behavioural categories
- Evaluation:
+ structured- easy to record and systematic
+ structured- quantitative so easy to analyse
+ unstructured- data rich
~ unstructured- qualitative so difficult to analyse
~ unstructured- observer bias
Observational Design- Behavioural Categories
=when a target behaviour is broken up into components that are observable and measurable (operationalisation)
- Evaluation:
+ data more structured and objective
( categories should be clear and unambiguous/all forms of target behaviour included/categories should not overlap)
Observational Design- Sampling Methods
- Event sampling=target behaviour is first established and researcher records every time it occurs
- Time sampling=target time frame is established and researcher records everything that occurs each time
- Evaluation:
+ ES- useful when target behaviour could be missed by TS
+ TS- effective in reducing number of observations have to be made
~ ES- may overlook important details
~ TS- observations unrepresentative of observation as whole
Self-Report Techniques- Questionnaires
=set of written questions used to asses person’s thoughts and/or experiences
- Open questions=no fixed range of answers- qualitative
- Closed questions=fixed number of responses- quantitative
- Evaluation:
+ cost-effective- large amounts of data quickly and distributed to many people
+ data straightforward to analyse
~ demand characteristics- social desirability bias
~ response bias- replying in similar way
Self-Report Techniques- Interviews
=person asks set of questions to assess interviewee’s thoughts/experiences
- Structured=pre-determined set of questions asked in fixed order
- Unstructured=no set questions- certain topic but free-flowing
- Semi-structured=list of questions but can ask follow-up questions
- Evaluation:
+ structured- reliable as standardised
+ unstructured- more insightful
~ structured- limited richness of data
~ unstructured- interviewer bias
~ social desirability bias
Self-Report Design
- Designing questionnaires:
=likert scale- strongly agree/disagree
=rating scale- numbers represent strength of feeling
=fixed-choice option- tick boxes - Designing interviews:
=standardised interview schedule to reduce interviewer bias
=quiet room / warm-up neutral questions / reminded of strictest confidence - Writing good questions:
=avoid overuse of jargon
=avoid emotive language and leading questions
=avoid double-barrelled questions and double negatives
Correlations
=relationship (strength and direction) between two co-variables- plotted on scattergram
- Positive correlation=one co-variable increases and so does the other
- Negative correlation=one co-variable increases and the other decreases
- Zero correlation=no relationship between co-variables
- Evaluation:
+ useful measure of how two variables related
+ quick and economical to carry out
~ correlation not causation
~ an intervening variable may be responsible for the relationship
~ sometimes misused/misinterpreted
Types of Data- Qualitative and Quantitative
=Qualitative-non-numerical
=Quantitative-numerical
- Evaluation:
+ qual- data rich
+ qual- high external validity
+ quant- easy to analyse
+ quant- objective
~ qual- difficult to analyse
~ qual- subjective interpretations open to bias
~ qual- fails to represent real life
Types of Data- Primary and Secondary
=Primary- data collected by the researcher themself for the purpose of their own investigation
=Secondary- data collected by someone else but can be applied to the researcher’s investigation
- Meta-analysis=number of studies identified with same aims/hypotheses- results pooled together and joint conclusion produced
- Evaluation:
+ primary- authentic
+ secondary- inexpensive and easily accessed
+ meta- high validity as large, varied sample
~ primary- time and effort
~ secondary- validity challenged due to content of data
~ meta- publication bias
Measures of Central Tendency
=general term for any measure of average value in a set of data- used in descriptive stats (graphs/tables etc)
- Mean=adding up all values and dividing by total number of values
+ more representative
~ easily distorted by extreme values - Median=central value in a set of ordered, chronological data
+ extreme values do not affect it
+ easy to calculate
~ less sensitive as values ignored - Mode=most frequently occurring value in a set of data
+ easy to calculate
~ crude measure
Measures of Dispersion
=general term for any measure of the spread in a set of scores
- Range=subtracting lowest score from the highest score (and adding one as mathematical correction)
+ easy to calculate
~ only takes into account most extreme values
~ does not indicate numbers grouped around mean or spread out - Standard deviation=tells us by how much, on average, each score deviates from the mean
+ much more precise
~ can be distorted by extreme values
Presentation of Quantitative Data- Tables and Graphs
- Summarising data in a table:
=when tables appear in a results section, they have been converted to descriptive stats
=include summary paragraph explaining numbers and drawing conclusions - Bar charts:
=difference in mean values easily seen
=nominal data (categorical)
=separate bars as separate conditions - Histograms:
=touching bars as continuous data
=area of bar represents frequency - Scattergrams:
=represents strength and direction of relationship between co-variables
Presentation of Quantitative Data- Distributions
- Normal distribution=symmetrical spread of frequency data that forms a bell-shaped curve
=mean, median, mode at highest peak - Skewed distributions=spread of frequency data that is not symmetrical, where the data clusters to one end
=positive skew- long tail is on positive(right) side of peak and distribution concentrated on left
=negative skew- long tail on negative(left) side of peak and distribution concentrated on right
Peer Review and Psychological Research and the Economy- the Role of Peer Review
=the assessment of scientific work by others who are specialists in the same field, to ensure that any research intended for publication is of high quality
- Main aims:
=allocate research funding
=validate quality and relevance of research
=suggest amendments - Evaluation:
=anonymity- of reviewer for more honest appraisal- some use as means to criticise other researchers as they are in direct competition
=publication bias- prefer to publish significant findings to increase their credibility- false impression of psychology
=burying groundbreaking research- suppressing opposition to mainstream theories- slows down rate of change
Peer Review and Psychological Research and the Economy- Implications for the Economy
=the state of a country/region in terms of the production and consumption of goods and services
- Attachment research into role of the father:
=showed both parents equally capable of providing emotional support for healthy development- promoting flexible working arrangements
=modern parents better equipped maximise income and contribute more effectively to economy - Development of treatments for mental disorders:
=absence from work costs economy £15 billion/year
=1/3 absences moderate health disorders
=patients assessed quickly and receive swift access to treatment
=can manage conditions effectively and return to work- economic benefit is considerable
A2- Case Studies
=in-depth investigation, description and analysis of a single individual, group, institution or event
- Qualitative data usually- case history of individual using interviews/questionnaires etc
- Longitudinal with potential additional data from family and friends
- Evaluation:
+ rich data
+ contribute to understanding of typical functioning
+ generate hypotheses for future study
~ not generalisable
~ subjective selection and interpretation of researcher=low validity
A2- Content Analysis
=enables indirect study of behaviour by examining communications that people produce
- Coding and quantitative data:
=coding- the communication to be studied is analysed by identifying each instance of the chosen categories- counting how many times something appears - Thematic analysis and qualitative data:
=thematic analysis- inductive and qualitative approach to analysis involving identifying implicit/explicit ideas within the data
=themes will emerge once the data has been coded - Evaluation:
+ navigates ethical issues well
+ communications have high external validity
+ produces qual and quant data
~ indirect study so analysed outside context in which occurred
~ lack objectivity
A2- Reliability- Ways of Assessing
=how consistent a measuring device is
- Test-retest reliability:
=assessing the same person via the same means on two different occasions
=must be sufficient time in between
=if correlated above 0.8 then reliable - Inter-observer reliability:
=extent to which agreement between two observers involved in observations
=same behavioural categories but record own data separately
=if correlated above 0.8 then reliable
A2- Reliability- Improving Reliability
- Questionnaires:
=test-retest
=correlation coefficient exceeding 0.8
=low test-retest then changed/rewritten- replace open questions with closed - Interviews:
=inter-interviewer
=interviewers properly trained so no leading questions
=more easily avoided in structured interviews - Observations:
=inter-observer
=behavioural categories properly operationalised
=categories should not overlap
=low inter-observer then further training - Experiments:
=standardised procedures
A2- Validity- Types
=extent to which observed effect is genuine- measure what supposed to and generalisable
- Internal validity=whether observed effects due to manipulation of IV
=threat to this is demand characteristics - External validity=whether research can be generalised to other settings, populations, eras
=ecological- extent to which findings generalised to other settings/situations
=temporal- extent to which findings generalised to historical times/eras
A2- Validity- Ways of Assessing and Improving Validity
- Face validity=when a measure is scrutinised to determine whether appears measure what supposed to
- Concurrent validity=extent to which psychological measure relates to existing similar measure- results similar to other test
- Experiments:
=control group- internal
=standardised procedures- internal
=single/double blind procedures- internal - Questionnaires:
=lie scale- control social desirability bias- internal
=assuring anonymity - Observations:
=minimal intervention- ecological
=behavioural categories - Qualitative research:
=depth and detail- ecological
=interpretative validity- extent to which interpretation of events matches that of participants
=triangulation
A2- Probability and Significance
=Probability=measure of the likelihood that particular event will occur- 0=statistical impossibility and 1=statistical certainty
=Significance=how sure we are that a difference or correlation exists
- Alternative hypothesis=directional/non-directional
- Null hypothesis=no difference between conditions
=stats test determines which hypothesis accepted and rejected - Significance level=point at which claim discovered large enough difference/correlation to claim an effect has been found- 0.05 / 5%
= p ≤ 0.05 = due to chance
A2- Use of Stats Tables and Type I/II Errors
- Calculated value compared to critical value from table
=one-tailed (directional) or two-tailed (non-directional)
=number of participants/degrees of freedom
=level of significance - 0.05=standard level of significance
0.01=more stringent level- studies with human costs (drug trials)
0.1=more lenient as significance not found at lower level - Type I error (false positive)=null rejected, alternative accepted but should be other way around
=0.1 - Type II error (false negative)=null accepted, alternative rejected but should be other way around
=0.01
=0.05 balances risk
A2- Choosing A Statistical Test
- Used to determine whether difference/correlation found in an investigation is significant
- Levels of measurement
Nominal=categoric
Ordinal=rate/scale
Interval=numerical with sizes- measurementDifference Correlation Un Re N Chi Sign Chi O Man Wil Spear I Un Re Pea Central tendency Dispersion Nominal Mode N/A Ordinal Median Range Interval Mean S.D
- Parametric test
Interval data
Normal distribution
Homogeneity of variance
A2- Mann-Whitney U
- Requirements:
Difference
Unrelated
Ordinal - U=where both groups meet on critical value table
A2- Wilcoxon T
- Requirements
Difference
Related
Ordinal - T=participant value in critical table
A2- Unrelated t-test
- Requirements
Difference
Unrelated
Interval - df=Na+Nb-2
A2- Related t-test
- Requirements
Difference
Related
Interval - df=N-1
A2- Spearman’s Rho
- Requirements
Correlation
Ordinal - rho=participant value in critical table
A2- Pearson’s r
- Requirements
Correlation
Interval - r=df=N-2
A2- Chi-Squared
- Requirements
Difference/correlation
Unrelated
Nominal - df=(rows-1)(columns-1)
- Draw contingency table
2x2 = 2 groups and 2 variables
But extra column and row for totals
A2- Features of a Science
- Paradigm and paradigm shift- Kuhn:
=shared set of assumptions
=psychology has multiple paradigms whereas natural sciences have one universal=not scientific
=paradigm shift when contradictory evidence replaces old paradigm due to scientific revolution - Theory construction and hypothesis testing:
=gathering evidence via direct observation
=hypotheses are tested to determine whether supported and scientific - Falsifiability- Popper:
=concepts only scientific if can be proven false by the evidence
=psychology allows for falsifiability through alternative and null hypothesis - Replicability- Popper:
=only scientific if findings repeatable across different contexts and circumstances
=leads to validity, reliability and generalisability - Objectivity and empirical method:
=biases cannot influence data or behaviour of participants- more control is more objective
=importance of data collection based on direct, sensory experience