Research Methods Flashcards
Aims
Developed from theories, general statements that describe the purpose of an investigation
Hypothesis
Clear testable statement that states the relationship between the variables and predict the outcome of a study before it starts
Directional hypothesis
Identifies direction of difference between two conditions
Directional
Non-directional hypothesis
States a difference between conditions but the nature of the difference is not made clear
Used when there is no prior evidence or results are contradictory
Non-directional
IV
Independent variable, variable that experimenter manipulates, or naturally changes so effect on DV can be measured
DV
Dependent variable, variable measured against the IV
Operationalisation
Variables being manipulated should be clearly defined and measurable
Extraneous variables
Variables that do not vary systematically with the IV, can often be controlled before experiment starts
Confounding variable
Variables that do vary systematically with IV, can’t be sure what is effecting DV
Essentially extraneous variables the experimenter failed to control for
Demand characteristics
Participants interpret cues from the experimenter (investigator effects) and research situation. They may change their behaviour as a result
Randomisation
The use of chance to reduce researcher’s influence on the design of the investigation
Standardisation
Making sure all ppts are subject to the same standardised instructions and experience
Random allocation
Attempt to evenly distribute ppt difference across experimental conditions in independent group designes
Counterbalancing
Attempt to control for order effects in repeated measures experiment, half the group experiences conditions in one order, and the other half in the opposite order
Experimental design
The different ways in which the testing of ppts can be organised in relation to the experimental conditions
Experiment
Where a change in IV effects the DV and results are measured/recorded
Types of experiment
Lab experiments
+ve and -ve
Takes place within controlled environment, researcher manipulates IV and records effect on DV. High control of extraneous variables
+ve High contol means study can be replicated without introducing more extraneous variables
-ve Low mundane realism, tasks carried out by ppts in lab exp may not represent real life
Types of experiment
Field experiments
+ve and -ve
Takes place in natural setting, researcher manipulates IV and records effect on DV
+ve Higher mundane realism than lab exp
-ve Ethical issue, ppts aware not be away they are being studied, no informed concent + invasion of privacy
Types of experiment
Natural experiments
+ve and -ve
IV changes without researcher’s influence, DV is simply recorded
+ve High external validity, involves study of real life issues as they happen - eg natural disaster on stress levels
-ve The situation may be rare in occurrence, limits ability to generalise results to other situations
Types of experiment
Quasi experiments
+ve and -ve
Iv is not manipulated but is based on existing differences between ppts, such as age
+ve Carried out under controlled conditions similar to lab exp
-ve Random allocation is not possible, may be confounding variables
Random sample
+ve and -ve
Every member of the target population has an equal chance of being selected.
+ve free from researcher bias, researcher has no influence over who is selected, prevents them picking those who would support hypothesis
-ve more likely to produce sample that is not representative, Eg 20 females from ldn and 1 boy
Systematic sample
+ve
Every nth person is chosen from a list.
+ve Avoids researcher bias
+ve Provides representative data
Stratified sample
+ve and -ve
The proportions of people in population sub-groups (strata) are reflected in the sample.
+ve Made to accurately reflect groups within population, highly representative sample makes generalising results possible
-ve odentified strata can not reflect all the ways people are different, cant be perfectly representative
Opportunity sample
+ve and -ve
Whoever is available at the time of sampling will be included.
+ve Convenient as saves researcher time and effort less costly than random sampling eg
- ve Results cant be generalised to whole pop as sample is drawn from one place eg street
- ve Researcher has control over selection of participants, researcher bias as they wont select people they dont like the look of
Volunteer sample
+ve and -ve
Ppts self select through an ad eg
+ve Easy to obtain sample, minimal effort required from researcher, time efficient
-ve Volunteer bias is a problem - may attract certain type of people eg helpful, curious. Restricts how far results can be generalised
Define ethical issue
When a conflict arises between the rights and dignity of a ppt and the needs of the researcher to produce valid/worthwile data beneficial to society
Ethical issue + solution
Informed consent
Ppts should be made aware of the aims, procedures and their right to withdraw before agreeing to take part
If informed consent cant be obtained:
- get prior general, retrospective or presumptive consent
Presumptive consent
Similar group to research group is asked for consent, if they agree, research group’s consent is presumed
Prior general consent
Ppts are shown a list of studies to consent to, one which involves deception, by agreeing they are essentially agreeing to be deceived
Retrospective consent
Ppts are asked for consent during debriefing, once they have already taken part in study
Ethical issue + solution
Deception
Ppts should not deliberately be misled, some deception is acceptable
-full debrief is required at the end
Ethical issue + solution
Confidentiality
Participants’ anonymity and privacy must be respected
-names should not be recorded
Ethical issue + solution
Protection form harm
Participants should not be exposed to any more risk than they would be in everyday life.
-participants should be offered counselling as part of the debrief
Pilot study
Small scale version of the experiment/questionnaire/observation and will usually involve a small sample size
Allows researcher to identify issues in exp design, saving them time and money in large scale study
Single blind preocedure
Ppts are not made aware some of the details of the investigation, like what condition they are in to reduce demand characteristics
Attempts to control confounding effects of demand characteristics
Double blind procedure
The ppt nor the person conducting the study is made aware of the aim, to reduce demand characteristics + investigator effects
Controlled observation
+ve and -ve
Watching and recording behaviour within a structured environment, some variables are manipulated
+ve Extraneous variables become less of a factor, replication of study becomes more possible
-ve Findings may not apply to everyday life, not generalisable to daily life
Covert Observation
+ve and -ve
Covert: behaviour observed without awareness/consent of ppt
+ve Low ppt reactivity, less chance of demand characteristics
-ve Ethically questionable as no consent for observation has been given
Participant Observation
Researcher may choose to take part in the observation with ppts, increases objectiveness
Observational design
Structured observation
Only records behaviour in reference to pre-set behaviour list, can lead to key behaviours being missed/ignored
Observational design
Un-structured observation
Observer records everything they see, this often can lead to too much data to record
Effective behaviour categories for observations
Behavioural categories should be measurable, not overlap and avoid ‘dustbin categories’ within which many behaviours are recorded
Observational design
Event/time sampling
Event: how often a behaviour occurs
- Useful for infrequent behaviour
Time: how often behaviour occurs within pre-established time
- May not represent whole behaviours
Kinds of data
Qualitative
+ve and -ve
Non-numerical description of ppt’s thoughts and feelings
+ve Rich in detail, high external validity
-ve Hard to analyse, conclusions may be too subjective
Self-report techniques
Questionnaires
Pre-set questions used to record thoughts and feelings towards a matter
Self-report techniques
Open/closed questions
Open: no fixed range of answers, more detailed
Closed: Fixed range of answers, lacks detail
Self-report technique
Structured interviews
Structured interview involves pre-set questions asked in a fixed order
Similar to questionnaires, but fewer respondents
Self-report technique
Semi-structured interview
Some pre-set questions but follow up questions are allowed
What should a questionnaire/interview not have
Questions should not include too much jargon, be leading or use emotive language, be double-barrelled or include double-negatives
Types of closed questions
Likert scale
Rating scale
Fixed choice option
Participant extraneous variables
Situational extraneous variables
Ppt variables: Age, gender, personality
Situation variables: Lighting, background noise, time of day
Types of design
Independent groups
+ve and -ve
Two separate groups do two separate conditions in the experiment
+ve order effects are not a problem, ppts less likely to guess the aim
-ve Ppts in different groups are not the same, personality variables makes it unclear if the IV is the only thing affecting the DV
To control for this use random allocation
Types of design
Repeated measures
+ve and -ve
All ppts take part in all conditions of the experiment
+ve Ppt variables are controlled for, this is also economically beneficial as less ppts are needed (less pay)
-ve Order affects may cause ppts to feel bored or fatigued by second task, this could affect performance
Types of design
Matched pairs
+ve and -ve
Ppts are matched on a variable relevant to the experiment, pair is then split up, one in each group
+ve All ppts take part in one condiion, order effects and demand characteristics are less of a problem
-ve Matching may be time consuming and expensive, more so if pre-tests are required
Sampling techniques
Methods used to select people from the population
Generalisation
Extent to which findings can be broadly applied to the population, high generalisability when sample is representative of population
Naturalistic observation
+ve and -ve
Observing and recording behaviour in the setting within it would naturally occur, all aspects of the environment are free to vary
+ve High external validity, behaviour occurs naturally, results can be generalised to day-day life.
-ve Lack of control over research situation makes replication harder
Investigator effects
The unconscious influence of the researcher on the research situation
Non-participant observation
observer may remain separate from the group, increases objectivity
Overt observation
+ve and -ve
Observing ppts with their knowledge
+ve Ethically acceptable, informed consent has been given
-ve Could lead to demand characteristics, behaviour and therefor results are affected
Self-report technique
Un-structured interviews
General topic questions developed based on responses
More flexibility
self-report design
Questionnaires
Likert scales, ranking scales, fixed choice option
Kinds of data
quantitative
+ve and -ve
Data expressed numerically
+ve Easy to analyse, less bias
-ve Narrow in scope may miss out lots of detail
Kinds of data
Primary
+ve and -ve
Collected first hand from the ppts by researcher for purpose of investigation
+ve High validity, targets relevant information
-ve Requires time and effort
Kinds of data
Secondary
+ve and -ve
Collected and analysed by someone other that researcher who conducted experiment
+ve Inexpensive, easy to access data
-ve May be outdated and incomplete
Descriptive stats - Measures of central tendency
Mean
+ve and -ve
All them all up, divide by number of categories/groups
+ve Most sensitive and representative of all results
-ve Easily distorted by anomalies
Descriptive stats - Measures of central tendency
Median
+ve and -ve
The middle value
+ve Not affected by extreme values
-ve Less sensitive than the mean
Descriptive stats - Measures of central tendency
Mode
+ve and -ve
Most frequently occurring result
+ve Easily to calculate
-ve Unrepresentative of all data
Descriptive stats - Measures of dispersion
Range
+ve and -ve
Minus lowest value from highest (add one if they ask to account for numerical inaccuracies)
+ve Easy to calculate
-ve May be unrepresentative of data set
Define generalisability
The extent to which findings can be broadly applied to the population
Correlations
Mathematical technique where a researcher investigates an association between two variables – called co-variables
Correlation coeffecient
Number between –1 - +1 that represents the direction and strength of a relationship between two co-variables
Descriptive statistics
Things such as graphs, tables and summary statistics (measures of central tendency and dispersion). Used to identify trends and analyse sets of data
Inferential statistics
Refers to the use of statistical tests
Case studies
An in depth investigation, description and analysis of a single individual, small group or institution
Content analysis
Research technique that enables the indirect study of behaviour by examining communications that people produce through various media, for instance TV, emails and texts
Reliability
Refers to how consistent the findings from an investigation or measuring device are
Test-retest
Method of assessing reliability of a questionnaire or psychological test by assessing the same person on two separate occasions
Inter-oberserver reliability
Extent to which there is agreement between two or more observers involved in observations of a behaviour
Validity
Extent to which an observed effect is genuine does u measure what it was supposed to measure (internal) and can it be generalised beyond the research setting within which it was found (external)
Ecological validity
Extent to which findings from a study can be generalised to other settings and situations
Temporal validity
Extent to which findings from a study can be generalised to different periods in time and eras
Face validity
Basic form of validity where the measure is scrutinised to determine whether it appears to measure what is it supposed to
Concurrent validity
Extent to which a psychological measure relates to an existing similar measure
Probability
Measure of the likelihood of a particular event occurring. 0 indicates statistical impossibility and 1 indicates statistical certainty
Significance
Statistical term that tells us how sure we are of a difference or correlation existing. A significant result means the researcher can reject the null hypothesis
Calculated value
A number resulting from calculating a statistical test
Critical value
The numerical boundary between acceptance and rejection of a null hypothesis
Paradigm
Set of shared assumptions and agreed methods. Within a scientific discipline
Paragdim shift
Significant change in the dominant underlying theory within a scientific discipline
Theory
Set of general laws or principals that have the ability to explain particular events or behaviours
Falsafiability
A principal stating a theory can not be considered scientific unless it admits the possibility of it being proven false
Replicability
Extent to which scientific procedures and findings can be repeated by other researchers
Objectivity
When all sources of personal bias are minimised to avoid distorting/influencing the research process
Emperical method
Scientific approached that are based on the gathering of evidence through direct observation and experience
Calculated value is ≥ critical value
- Chi squared
− Spearman’s rho
− Pearson’s R
We accept alternative, reject null hypothesis
Calculated value is ≤ critical value
- Sign test
− Mann Whiteney - Wilcoxon
We accept alternative, reject null hypothesis
