Research Methods Flashcards
Experimental Method - AO1
Aim - purpose of the investigation.
Hypotheses - testable, operationalised statement.
Directional or non-directional - identifying a difference/correlation or not. Choice depends on previous research.
Variables - AO1
IV and DV - IV manipulated, DV measured.
Levels of IVs - experimental or control.
Operationalisation - making variables measurable.
Research Issues - AO1
Extraneous variables - nuisance variables but randomly distributed.
Confounding variables - vary systematically with IV.
Demand characteristics - when participant guesses aims of experiment and changes behaviour.
Investigator effects - influence of researcher on DV/design decisions.
Randomisation - chance methods to reduce bias.
Standardisation - ensuring all participants have the same experience.
Experimental Designs - AO1
Independent groups - participants in each condition of experiment are different.
Repeated measures - all participants take part in all conditions.
Matched pairs - similar participants paired on participant variables, allocated to condition A or B.
Experimental Designs - AO3
Independent groups
- Participant variables not controlled (use random allocation) may still be differences
+ No order effects
Repeated measures
- Order effects (use counterbalancing)
- Demand characteristics
+ Participant variables controlled
Matched pairs
+ No order effects and variables controlled. Little demand characteristics.
- Time and cost consuming.
Types of Experiment - AO1
Lab - IV manipulated in controlled setting. Participants go to researcher.
Field - IV manipulated in natural setting - researcher goes to participants.
Natural - IV changes naturally. DV/setting may be in a lab.
Quasi - IV based on an existing difference between people, effect on DV is recorded.
Types of Experiment - AO3
Lab
+ High internal validity - control over extraneous variables.
+ Replication more possible (support for findings)
- Low internal validity (DC’s) and low external validity (generalisability.)
Field
+ Higher external/ecological validity (more authentic)
- Lower internal validity (less control)
- Ethical issues (consent not possible)
Natural
+ High external validity (real-world problems)
- Limited opportunities, no random allocation
- No manipulation of IV (no cause and effect)
Quasi
- Issues with lab experiments as often in lab.
- No manipulation of IV or random allocation
Sampling Methods - AO1
Random sample - equal chance of selection
Systematic sample - Selecting every nth person from a list
Stratified sample - sample reflects proportions of strata in a population
Opportunity - choosing who is available
Volunteer - participants self-select
Sampling Methods - AO3
Random
- Potentially unbiased - control CVs and EVs
+ Time consuming, may not work
Systematic
- Time consuming, those selected may refuse
+ Objective/bias free
Stratified
- Time consuming - cannot account for all subgroups
+ Most representative of population
Opportunity
+ Quick and easy
- Unrepresentative
Volunteer
+ Easy and participants engaged
- Volunteer bias - responsive to cues.
Ethical Issues and How to Deal With Them - AO1
Informed consent - advise participants what is involved. Reveals aims.
Signed consent form. Presumptive. Prior general consent. Retrospective. Family.
Deception - misleading participants/withholding information. OK if not distressing.
Debriefing. Right to withdraw. Counselling.
Protection from harm - psychological/physical risk factors.
Deal with in same way as deception.
Privacy and confidentiality - right to control and protect personal data.
Use numbers not names. Data not shared with other researchers.
Pilot Studies and Single/Double Blind - AO1
Check procedures and techniques, make changes.
Single-blind - participants aren’t aware of aims until end.
Double-blind - neither participants nor individuals conducting research know the aim/condition.
Types of observation - AO1
Naturalistic - behaviour observed where it would normally occur. No control over variables.
Controlled - some control over environment - manipulation of variables.
Covert or overt - observing with/without participants’ knowledge.
Participant or non-participant - join group or remain an outsider.
Types of observation - AO3
All observations
- Observer bias and no causal relationships.
Naturalistic
- Low internal validity (no control)
+ high external validity (everyday life)
Controlled
+ High internal validity
- Low external validity (unless covert)
Covert
+ Low demand characteristics
- Ethical issues
Overt
- Behaviour may be different.
Participant
+ Increased external validity
- May lose objectivity
Non-participant
+ More objectivity
- Less insight
Observational Design - AO1
Record everything - unstructured. Categories - structured.
Behavioural categories - target behaviours broken down into observable components.
Sampling methods
Time-sampling - count at timed intervals.
Event-sampling - count events.
Observational Design - AO3
Structured
+ Numerical - easier to analyse.
Unstructured
- Observer bias
- Qualitative data hard to analyse
Behavioural categories
- Must be observable - no overlap.
Event
+ Useful for infrequent behaviours
- Misses complexity
Time
+ Less effort
- May not represent whole event
Questionnaires - AO1
Pre-set list. Can be used to measure DV.
Open and closed questions - fixed-choice or not.
Questionnaires - AO3
Questionnaires
+ Distribute to many people
+ Fixed choice - easy to analyse
- Social desirability response bias
Closed and open questions
+/- Effects ease of analysis
Interviews - AO1
Structured interviews - pre-set questions, fixed order, face-to-face.
Unstructured - no formula, just general topic. Questions based on responses.
Semi-structured - some pre-set questions with follow-up.
Interviews - AO3
Structured
+ Easy to replicate
- Interviewer can’t explain or elaborate
Unstructured
+ Flexible
- Increased interviewer bias
+ Social desirability bias reduced by rapport.
Correlations - AO1
Types - positive, negative, zero. Strong or weak. Measured from -1 to 1.
Difference between experiment and correlation.
No manipulation of IV - co-variables. No cause and effect.
Correlations - AO3
+ Quick and economical - using secondary data.
- Can’t demonstrate cause and effect - may be a third variable.
- Often bidirectional (which variable is affecting which)
Qualitative vs. Quantitative Data - AO1
Qualitative - written - non-numerical. Descriptions of thoughts/feelings.
Quantitative - expressed numerically.
Qualitative vs. Quantitative Data - AO3
Qualitative
+ Richer data. Greater external validity.
- Difficult to analyse/subjective.
Quantitative
+ Can statistically analyse
+ Less bias
- Narrower in meaning
Primary vs. Secondary data - AO1 + AO3
Primary
Collected first-hand for purpose of investigation.
+ Targets relevant info
- Time consuming
Secondary
Collected by someone else (eg. meta analysis.)
+ Inexpensive and easy
- Variation in quality - eg. outdated.
Meta-analysis - AO1 + AO3
Collection of multiple studies and statistical analysis of them. Secondary data. Larger number of studies the better.
+ High validity
- Publication bias/file drawer problem.
Measures of central Tendency - AO1 + AO3
Mean
Add up, divide by N.
+ Most sensitive and representative.
- Distorted by extreme values.
Median
The middle value
+ Less affected by extremes.
- Not sensitive.
Mode
Most frequently occurring result
+ Relevant to nominal data
- Crude, unrepresentative
Measures of Dispersion - AO1 + AO3
Range
Subtract lowest from highest (add 1.)
+ Easy to calculate
- Unrepresentative if there are extremes
Standard deviation
How much scores deviate from mean
+ More precise than mean
- Distorted by extreme values
Presentation of Qualitative data - AO1
Tables - raw scores in a grid.
Bar charts - discrete categorical data.
Histogram - continuous data rather than discrete, no space between bars.
Scattergrams - shows strength and relationship between co-variables.
Normal distribution - bell curve. Mean, median and mode all the same. Tails never touch 0.
Skewed distributions - negative leans right, positive leans left. On positive, goes mode median mean.
Peer review - AO1 + AO3
Main aims:
Allocate funds
Validate quality of research.
Suggest improvements
- Anonymous - may permit unjustified criticisms by rivals
- Publication bias - file drawer problem creates false impression of current knowledge
- Burying ground-breaking research - maintains status quo
Case Studies - AO1 + AO3
Detailed analysis of an unusual individual or event. May also be ‘typical’ behaviours.
May involve a case history - qualitative but also quantitative data. Often longitudinal
+ Insight into unusual cases, e.g. HM provides understanding of normal functioning. Generate hypothesis for future studies
- Generalisability issues - small samples. Conclusions based on subjective interpretations and subjective data from participants.
Content Analysis - AO1 + AO3
Content analysis - form of observation where communication is studied indirectly. Data must be categorised into meaningful units and operationalised.
Thematic analysis - recurrent themes that keep ‘cropping up’ are identified and supported by quotes.
+ Fewer ethical issues, high external validity, flexible
- Information may be studied out of context and be subjective. Reflexivity aims to address issue of bias.
Ways of assessing reliability - AO1
Reliability - does the measurement produce the same result when the thing being measured has changed.
Test-retest - same test is administered to the same person on different occasions and results compared.
Inter-rater/inter-observer - observers compare data in a pilot study or at the end of study to make sure behavioural categories are consistently applied.
Correlation coefficient of at least +0.80
Ways of improving reliability - AO1
Questionnaires
Change to closed questions - less interpretation, less ambiguous.
Interviews
Avoid leading questions or ambiguity, ensure interviewers are trained.
Observations
Behavioural categories operationalised, more training may be needed.
Experiments
Standardised procedures ensures consistency
Types of validity - AO1
Validity - whether a psychological test, observation etc., produces a result that is legitimate.
