research methods Flashcards
aims
a general statement describing the purpose of a study
hypothesis
statement made at the beginning of a study - describing the relationship between variables
independent variables
the variable changed by the researcher
dependent variable
the variable which is affected by the independent variable, it is measured by researchers#
operational variables
clearly defining each variable before research starts, increases internal validity
extranious variables
a variable that can effect the DV that isnt the IV
confounding variables
extraneous variable that only effects the DV
demand characteristics
participants work out the aim of the experiment and manipulate their responses to either please the experimenter (please-U effect) or sabotage the experiment (screw-U effect)
investigator effect
investigator gives unconscious cues to participants manipulating their behaviours
randomisation
randomizing as many parts of the experiment as possible to reduce biases
standadisation
standandising the procedure so it is the same for each participant
independent groups
2 seperate groups are used, one completed the experimental condition, the other the control condition
AO3
- requires lots of participants
- can create a confounding variable if the participants are not randomly allocated to a group
repeated meausures
1 groups of participants that experiance both the experimental and control conditon
AO3
- order effect - can effect results - more demand characteristics - accounted for with counter balencing
matched pairs
prrticipants paired up on their ability of a specific variable - one member of each group allocated to the experimental group the other to the control group - prevents participant variation from becoming confouning
AO3
- participnats never matched exactly
- large sample
laboratory experiment
conducted in a controlled enviroment
AO3
+ confounding and extraneous variables easy to control - high internal validity
+ increased replication due to high levels of control
- may not be generalized to real life
- unnatural behaviours - know they are being observed
- low mundane realism
field experiemnt
like lab but set in a everyday setting where the task naturally occurs, IV manipulated artifically AO3 \+ mundane realism - less control of CV and Ev - ethics of consent
natural experiments
like field experiments but IV varies naturally
AO3
+ high external validity - rela life situations
+ opotunity for potentially unethical research if done artificially
- rare to find correct conditions
- unable to randomly alocate groups - biased groups
quasi-experiment
field but IV based on exisiting diffrence ie age AO3 \+ often controled conditions - lab - cannot randomly alocate participants - hard to determin cause and effect
population
all the people a peice of research applies to
sample
a subset of the population selected to take part in an experiment - often biased or urepresentitve
random sample
all members of the population can be selected - assigned a random number - selected by randomly generating them AO3 \+ unbiased - difficult and time consuming - unrepressentitve
systematic sampling
population ordered - every nth person selected to make up sample AO3 \+ objective - time consuming - p's may refuse to take part
stratified sampling
population split into strata (sub populations) a sample of each is selected proportionate to that of the population to make up the sample
AO3
+ representitve population
- cannot represent all strata
opportunity sample
selects anyone from the target population that is available AO3 \+ cheap and easy - unrepresentative - often biased
volunteer sample
self selections - participants volunteer to be part of sample
AO3
+ easy
+ participants more engaged
- volunteer bias - certain type of perosn likely to volunteer
informed consent
P should know what they are getting into before participating in research - should know aim, the procedure, their rights ie right to withdraw, and what their data will be used for - should be a signed consent letter - if p under 16 parent or guardian must sign consent
deception
misleading or withholding info from P - should be a full debrief after experiment especially if deception is necessary - should be reminded of right to withdraw
protection from harm
P’s should not be placed at any more psych or physical harm than they would be in their every day life - counsiling should be offered to all P’s
privacy and confidentiallity
P’s should have ability to control who can hold and how much data - protected by GDPR - reminded of right to withdraw info - anonymity maintained in report ie initials or p numbers
pilot study
small scale trial of experiment - road test procedure hypothesis and methods
single-blind procedure
participant does not know the aim of a study - protects against demand characteristics
double-blind
neither participants or researcher knows the aim of the experiment - protects from double blind and researcher effect
control groups
a baseline study which allows researchers to compare the results of the experimental group against it
naturalistic and controlled observation
naturalistic - observation in a location where the behaviour would naturally occur
controlled - certain aspects of the location or task is manipulated to control EV or CV
AO3
naturallistic have high external validity - easy to generalize to everyday life - hard to recreate (lack of control) - no control of CV and EV - controlled oposit
covert and overt observations
covert - participants do not know they are being observed - done using concelment or cameras
overt - participants know they are being observed - have given informed consent before hand
AO3
covert observations eliminate all demand characteristics and ensure behaviors are natural - ethical issues with right to privacy - overt oposit
participant and non-participant observation
participant - researcher is part of the group - allows for first hand accounts and self report data - can bias data
non-participant - observer separate from group - more objective
AO3
participant observation gives researchers insite into through process behind behaviors - but can become subjective (going native) - non participant oposit
general observation AO3
+ evidence of how P’s acted - self report not always accurate
- observer bias - observer interprets behaviors differently
- cannot demonstration a cause and effect relationship
inter observer reliability
best practice for multiple observers to observe the same trial - all observers report on the same coding scheme - the data is then compared, the closer data is the higher the inter observer reliability is, indicating unbiased observations
behavioural catagories
instances of a behaviour are broken down into set catagories - essentially operationalizing what is being observed
AO3
if categories not clear they can cause observers to make inferances - dusbin catagories can bias results - catagories that are similar or overlapping can cause confusion and poor inter-observer reliablity
sampling method
continuous recording - all instances of a target behavior are recorded
event sampling - number of times a behaviour occurs is recorded
time sampling - records behaviour in a set time frame
AO3
event sampling useful when event infrequent but may over look details if it is complex
time sampling may be unrepresentative of event (too many or too little occurrences) - reduces number of observation
questionnaires
a set list of questions the participant responds to from which thoughts and feelings can be assessed
open questions - a prompt without a set list of answers - produced qualitative data
closed questions - have a fixed set of answers - quantitative data
AO3
+ low cost and fast
+ quick to produce and analyse
- demand characteristics - social desirability bias
- response bias - in closed questions always answering around the same point on a scale
interviews
structured - interviewee is asked a set list of questions
unstructured - more like a convosation - no set questions
semi-structured - a set list of questions but interviewer is free to ask follow up questions
AO3
structured interviews
+ easy to replicate
+ standardized
- limits depth of and unexpected data
unstructured interviews
+ unexpected data and depth
- interview bias
- can produce irrelevant data
designing closed questions for questionnaires
likert scales - the p’s response indicates their agreement with a scale
rating scales - response indicates their strength of feeling on a topic
fixed choice options - list offers options participants select all that apply
designing interviews
interview schedule (list of questions) should be standardized should make interviewee comfortably, away from others, and a rapport should be built
writing good questions
limit technical jargon
minimize emotive language and leading questions
double barreled questions removed - p’s could agree to half
remove double negatives
correlations
the strength and direction of the r/s between two + co-variables
positive - as one rises the other also rises
negative - as one rises the other falls
cannot be determined though experiments where the IV is artificially varied
AO3 correlations
+ useful for showing nature and strength of relationships
+ provide basis and ideas fro future research
+ quick and easy
+ can use secondary data
- only shows how not why
- does not determine cause and effect
- can lead to miss interpretations of r/s especially in media
quantitative and qualitative data
qualitative data - lexical rather than numerical - often gathered through interviews adn questionnaires
AO3
+ richer data with greater insight into participant
- difficult to analyse
qualitative data - numerical data - often from experiments - includes yes/no answer questionnaire questions
AO3
oposit of qualitative
= less likely to be biased
primary v secondary data
primary data - specifically collected for the research being conducted - arrives directly from participants
AO3
+ custom collected for research
- time and effot
secondary data - data collected for another piece of research - already exists in the psychology world AO3 \+ inexpensive and easy to access - quality and accuracy unknown - may not match exact needs
meta-analysis
a form of secondary data which collates the data of several previous studies investigating the same aims/hypothesis - draws conclusions from all data
Ao3
+ larger more varied samples
- publication bias - researchers only select studies they know will support their study, leaving out others
skewed distrabutions
positive skew - peak to the left
negative skew - peak to the right
mode->median->mode order of meassures of centeral tendancy from peak to tail
peer review
a review of all areas of the research conducted by a small group of peers in the particular field, they give an objective view on whether the research is valid or not - the research should not know who is going to review the research and vice versa
aims
allocates funding
validate quality and relevance of research
suggest amendments or improvements
peer review AO3
- anonymity - reviewers may critisise research unjustly if they do not like a researcher or are in competition for funding
- publication bias - journals tend to publish positive or attention grabing articles - ignores other research and gives unjust view of psychology
- burying groundbreaking research - reviewers more likely to critisise papers which go against their view or the widley accepted opinion
implications on the econemy
research can affect financial prosperity
role of the farther - changed from fun parent responsible for play to an equal car giver responsible for emotional and cognitive development - mother can now be primary earner / farther stay at home dad - more couple earn more so contribute more to the economy
treatment of mental disorders - absence from work cost ~ £15b yearly in the UK 1/3 caused by mental health - research into treatments can cause a smaller cost each year with better treatments
case study
a detailed and in-depth analysis of unusual, events, individuals, or institution - commonly produces qualitative data
AO3
+ detailed with rich insight
+provides understanding of typical function
+ often used to generate hypothesis for future studies
- small sample makes genralizations hard
- info published based on researchers inferences and selections
- memory not snap shot of events
content analysis
observations made from a participants communications - sumerises communications and draws a conclusion
coding - categories of data, the number of time each category appears in the content is recorded - produces quantitative data
thematic analysis - the themes that occur in a communication is recorded - qualitative data
AO3
+ avoids ethical issues - content already occurs in society
+ both quantitative and qualitative
- indirect
- context removed
- subjective - unconscious researcher biases
reliability
a measure of consistency - if a measurement produces the same result multiple times it is consider reliable - measured with correlation coefficent shoulf be +0.8
test-retest - the same test should be administered to the same participants on different instances - same/similar answers indicate high reliability - a period of time long enough for the P to not be able to recall previous answer must be left
inter-observer reliability - multiple researchers observe the same trial and score it using the same coding - if results similar reliability high
improving reliability
questionnaires - test-retest should be used - questions with low score should be rewritten or deselected
interviews - same researcher should be used - structured interview best for limited variations
observations - categories operationalized well - overlapping categories adjusted
experiments - procedure standadized
types of validity
validity - a measure of if the research produces legitimate results
internal - whether the effects observed are a result of manipulating the IV not another variable
external - whether the research can be genralised beyond the setting of the research
ecological - whether the findings can be genralised to to other settings
temporal - whether the findings can be applied over an extended period of time
ways of assessing validity
face validity - informally checking the research to see if it achieves what it sets out to measure on the face of it
concurrent validity - checks if a particular test or scale produces a similar output to an already established equivalent
improving validity
experiments - use control groups - standadise the procedure - single/double-blind
questionnaires - include a lie scale (redundant questions to ensure the P responds consistently) - anominity
observations - covert observation - ensure categories are not broad, overlapping or ambigous
qualitative research - use multiple sources in case studies
choosing a statistical test
test of difference or correlation
experimental design
level of measurement
chi squared - nominal, unrelated/correltion sing - nominal, related design mann-whitney - ordinal, unrelated wilcoxon - ordinal, related spearman rho - ordinal, correlation unrelated t-test - inteval, unrelated related t-test - interva, related pearsons r - interval, correlation
data types
nominal - categorical data - discrete
ordinal - has some for of order ie scale - subjective
interval - based on numerical scales with equal units that are precisely defined ie stop clock
type I / II errors
type I - null hypothesis incorrectly rejected - false positive
type II - alt hypothesis incorrectly rejected - false negative
sections of a psychology report
abstraction - a short summery (150-200 words) of aim, hypothesis, procedure, results and concusion
introduction - a literature review of general field, relevent theories, studies and concepts - should start broad and become more precise where the aims and hypothesis are presented
method - should be specific enough to allow other researchers sample, materials recreate the study - includes design, sample, method and apparatus, procedure and ethics
results - sumarise findings of study - descriptive statistics (graphs etc), inferential statistics (process of statistical testing)
discussion - verbally sumarise the results - limitations - impilications and real world applications
refrences - details of any source material cited
paradigms and paradigm shifts
paradigms - a shared set of assumptions and methodologies
khun - psychology pre-scientific - too many internal disagreements over the core principles and assumptions to be considered a science
paradigm shift - how sciences progress - a revolution occurs where an established paradigm is questioned gaining evidence contradicting it - the old paradigm falls and is replaced with a new set of general assumptions and methodologies
theory
a set of general laws or principles that are able to explain an event - should be systematically and objectively tested through multiple hypotheses
theory construction - through gathering evidence via direct observations supporting an hunch
falsifiability
popper - for research to be considered strong it should be proven not false - strong research should be constantly proven not false through challenges with alternative theories
replicability
for a theory to be trusted it must be proven true across a range of contexts and circumstances
important for external validity and generalization
objectivity
researchers should maintain a critical distance - personal opinions should not bias data or influence participants - empirical evidence should be used
