Research Methods Flashcards
Operationalised hypothesis
Clearly defining variables in terms of how they are measured
Eg after drinking 250ml of monster participants say more words in the next five minutes than participants who drank 250ml of water
Extraneous variables
Any variable that could affect the dv
Doesn’t change with Iv
Eg Age
Confounding variable
Changes the DV
Changes with the IV
For example excitement affects chattiness
Experimental designs
Independent groups
Repeated measures
Matched pairs
Independent groups
One group 1 condition other group different condition
L- less economical
S- no order effects can’t guess aims
Repeated measures
All pp experience both conditions
L- order effects affect performance
S- pp variables are controlled
Matched pairs
Pp matched on pp variables which may effect DV assigned to different conditions
S- control confounding variables/ demand characteristics
L- pp can’t be matched exactly pp variables may remain
Types of experiment
Lab experiment
Field experiment
Natural experiment and Quasi experiment
Lab experiment
Highly controlled environment
S- high control internal validity
L- lack generalisation artificial behaviour lack ex validity
Field experiment
IV manipulated in every day setting
Eg nurse and drugs obey study
S- high mundane realism high ex validity
L- loss control of CV and EV - harder to establish IV effect
Natural experiments
Researcher has no control over IV
Eg sex of participant = IV
S- provide opportunities for research
L- cannot randomly allocate unsure IV causes DV
Quasi experiment
Iv is pre-existing
Eg IV= gender
S= controlled conditions high internal validity
L= cannot claim Iv has had an effect confounding variables
Sampling
Systematic
Random
Stratified
Opportunity
Random
List of people assign number sample using random generator
S= unbiased
L= may be unrepresentative
Systematic
Every nth member of population
S= objective no influence who is chosen
L= time consuming
Stratified
Sample reflects proportions of people from different subgroups within population
S= representative sample
L= not perfect never reflect every difference
Opportunity
Select those willing and available
S= convenient
L= unrep and researcher bias
Volunteer
Pp select themselves to be part of the study
S= easy min input
L= volunteer bias certain profile lack gen
Ethical issues
Informed consent
Deception
Protection from harm
Privacy
Observations
Naturalistic / controlled
Covert/ overt
Participant / non participant
Naturalistic/ controlled
Naturalistic - watching behav in setting of normal occurrence
S= high ex validity
L= lack of control
Controlled - in structural environment lab
S= controlled
L= low ex validity
Covert/ Overt
Covert- pp watched without knowledge
Overt- pp watched with knowledge
Co L= ethics questioned
Participant/ non participant
Increased insight natural behaviour - ex validity
Lose objectivity
Observational design
Unstructured
Structured
Unstructured
Write down everything you see
Structured
Write down specific target observations - catergories
Behavioural categories
Target behav broken up into components observable and measurable
Sampling methods
Event sampling - counting behaviour continuously
Time sampling - recording behaviour in specific time intervals
Self report
A person is asked to explain own feelings or experiences related to a topic
Evaluation
L- Desirability bias pp may try to give the “correct” answer to be socially accepted for example
S- may be more useful than observing alone
Peer review
Assessment by others who are specialists in same field to ensure data is of high quality
1. Allocate research funding
2. To validate quality and relevance of research
3. Suggest amendments or improv
Eval
Anonymity criticise rival researchers
Publication bias - publish headline grabbing articles data disregarded
Burying groundbreaking research to maintain status quo contradict established research
Meta analysis
Process of combining findings of studies on a particular topic to provide an overall statistical conclusion
Features of reports
Abstract=summary of entire research project
Method= sample design materials and procedure described
Results= quant - stats analysis qual content analysis
Discussion= evaluation
References
Pilot studies
Small scale version of an investigation that takes place before real one is conducted
Aim is to check procedures materials etc work
Allow researcher to make changes
Positively skewed
Mode median mean
peak closest to y axis
Mode farah positive day if wins the Y marathon Mean stinks
Negatively skewed
Mean median mode
Mode furthest from y axis
Sad day when mode farah is beaten by both median and Mean in the Y marathon
Type 1 error
When null hypothesis is incorrectly rejected
False positive
Type 2 error
When null hypothesis is incorrectly accepted significance level is too low
False negative
Case studies
An in depth investigation description and analysis of an individual, group or institution
Qualitative data occasionally quantitative
Longitudinal
Eval
Offer rich detailed insight on atypical forms of behav eg HM Hypotheses for future study possibly leading to rev of new theory
Generalisation issue v small sample size subjective interpretation of researcher question case study validity
Correlation coefficient
+0.8 within each other
Face validity
Appears to measure what it measures
Concurrent validity
Agreement between 2 different assessments/ psych research
Ecological validity
The extent to which findings can be generalised to other settings ex validity
Temporal validity
Whether findings hold truth over time
Eg Freuds theory lacks temporal validity as penis envy is outdated reflects patriarchal Victorian era
Ways of assessing validity
Face validity appears what it is supposed to measure - eyeballing
Concurrent validity - particular test is demonstrated where results should be similar eg new intelligence test may compare iq score of well established test before
Improving validity
In experiments- use control standardise procedures use single bind or double bind to avoid to minimise investigator effects reduce demand charac
Questionnaires - incorporate a lie scale to control social desirability bias
Observations - covert observation and refine behavioural catergories
Qualitative research - interpretative validity direct quotes etc and triangulation using many sources as evidence
Paradigm
Shared assumptions and agreed methods within a scientific discipline
Paradigm shift
Result of scientific revolution when there is a significant change in theory
Theory construction
Process of developing explanation for causes of behaviour by gathering evidence and determining a theory
Falsifiability
Principle that a theory cannot be regarded as scientific if variables cannot be proved untrue eg Freud
Empirical method
Scientific approaches that are based on gathering of evidence through direct observation and experience
Investigator effects
Any effect of the investigators behaviour on the research outcome may include design of study or interaction with pp
Standardised script used
P<0.05
There is a less than 5% likelihood that results occur if there is no real difference between conditions
