Research Methods - paper 2 (all) Flashcards
what is an aim?
driven by theory & states the intent of the study in genreal terms
what is a hypothesis?
- prediction of an investigation - involes independant & dependant variable
- need to be directional or nondirectional
- operationalise
independant variable
characteristics that is manipulated
dependant variable
measurable variable
directional hypothesis
states the specific effect on the IV
non-directional hypothesis
states there will be a change but is not specific - when there is no previous research
null hypothosis
no difference or no relationship - prediciting there will be no effect
operationalisation of a hypothesis
make the variables measurable
extraneous variables
are all variables that are not the independant variable but could effect the depenant variable
confounding variables
a variable that is not the indpendant variable but may have effected the dependant variable (extranous variable that you cant control)
controlling extranous variables
- standardisation
- single blind
- double blind
- random allocation
- standadised instructions
- ensure anonymity
- counterbalancing
controlling extranous variables - standardisation
all aspects of the environment are the same
controlling extranous variables - single blind
making sure particpants dont know the aim
controlling extranous variables - double blind
particapants nor expirementer know the aim or conditions
controlling extranous variables - random allocation
randomly allocating particapants to the independant variable
controlling extranous variables - standardised instructions
pre written script for both groups
controlling extranous variables - ensure anonymity
make sure particapants know their results are anoymous
controlling extranous variables - counterbalancing
change order for the conditions for each particapnt
extranous variables
- situation variables
- particapnt variables
- demand characteristics
- investigator effects
- order effects
- social desirability
- placebo effect
extranous variables - situational variables
- factors in the enviroment that can unintentionally affect the DV
- more noise in one group than another
- standarndisation
extranous variables - particapant variables
- the charcteristics of a person may impact the results
- better reaction time
- random allocation
extranous variables - demand characteristics
- particapants uncounsouly changing their behaviour as they have worked out the expirenment
- one group works it out and the other doesnt
- single blind
extranous variables - investigator effect
- the expirementor uncounsiously conveys to the particapants how they should act
- standarisation
- double blind
extranous variables - order effect
- any effect that comes from doing the expirement twice
- one group has a better time because they have practised before
- counterbalancing
extranous variables - social desirability
- when particapnts want to come across a certain way to expirementors so change their behaviour
- ensure anonymity
extranous variables - placebo effect
- when taking a drug and they act in a way they think they should feel
- single blind
- double blind
qualitative data
- observation notes
- subjective
- hard to anaylise
- open questions
- cant test hypothesis
quantative data
- objective
- scores on a likert scale
- closed answers
- easy to anaylse & draw graphs
- can test hypothosis
primary data
- collected by researcher
- collected specificly for hypothesis
- longer and more difficult
- researcher has control on how data is collected
secoundary data
- quick and easy
- collected for a different hypothesis
- researcher has no control on how it is collected
observations
- naturlistic & controlled
- covert & overt
- particapant & non-particapant
- structered & unstructured
naturlistic observations
- in a setting where the behaviour would naturally occur
- all aspects of the enviroment are free
- no manipulation of the IV
+ .+high ecological validity - measures how generalisable expirmental findings are to the real world - natural - .- lacks informed consent - observing in a natural enviroment means that they dont know its happening so cant give consent
controlled observations
- carried out in a lab/controlled enviroment
- researcher decideds and controlls everything
- standardised procedures
- .+ standardised procedures - same process
- .- lacks ecological validity - not genralisable to everyday life
covert observations
- particapants are unaware they are being watched
- .+ high ecological valdity - behaviurs are extremely accuart
- .- lacks ethical - no informed consent
overt observations
- particapnts are aware they are being watched
- .+ ethical - infomred consent
- .- lacks validity - change behaviour due to social desilibty
particapant observations
- resesrcher joins in as a particapnt ( undercover)
- .+ more insight - can see everything going on - high validity
- .- to involved - loose objective and make it to personel
non-particapant observations
- researcher is detached from the observations - covert or overt
- .+ find objective - stick to aim and find pure research
- .- less insight - cant always see why things happen
structured observations
- clearly defined way of measuring behaviour
- specific set stages (piloet studies first)
- .+ standardised procedures - high reliability
- .- limited - cant write something if you see anything else other than in the procedure
unstructured observations
- no set catogories to look out for
- .+ more detail - no limitations so can measure anything
- .- hard to anaylise - qualitative data
observer bias
- seeing something that yu expect to see - reseacher bias
- fixing it ;
1. observers are calibrated - trained in a piloet study , define behaviour catagories
2. results are compared - check consistently
3. do observations - seperartly
4. stats test - check similarity , standard deviations - under 0.8
event sampling
- recording observations when target behavoiour
- noting all occurance happens
time sampling
- recording behaviours in specific time period
- behaviour ingnored outside time period
questionares
- open questionares - not limted with how they respond
- closed questionares - have limited numer of repsonse (likert scale, rating scale , fixed choice)
- what questionares need:
1. no use of jargon
2. no leading questions
3. no double barrel questions
disadvantages of questionares
- response bias - social disarabilty , aquiescence bias (saying yes bc dont understand)
- hard to interpret - everyone has different meanings for certain things
- sample not reprosentive - certain type of people reply to them , extroverts , authoritarian
advantages of questionares
- quick and easy - mass produce
- no investigater effect - investigator isnt present so no more likely to be true
- right to withdraw - just dont send it in
interviews
- structured - set questions , limited awsners , same for everyone
- semi-structured - list of set questions but can elabroate
- unstructured - no specifc questions , detailed awsners
advantages of interviews
- detailed
- can clarify things that arent clear
- freely speak
disadvantages of interviews
- time consuming
- hard to anaysle
- investigator effect
expiremental type
- lab studies
- feild studies
- natrual study
- quasi study
expiremental type - lab studies
- standardised enviroment
- researcher controlls everything
- .+ high validity
- .+ scientific crediability
- .+ controls extrenous variables
- .- lacks ecological validity
- .- demand characteristics
expiremental type - feild study
- takes place outside a lab - basic scientific procedures
- .+ high ecological validty
- .+ no demand characteristics
- .- ethical issues - no informed consent
- .- not as replicable
expiremental type - natural study
- natrual enviroment
- no control of iv - no manipulations
- .+ high ecological vaildity
- .+ good ethics as no manipultion so no particapnt harm on purpose
- .+ no demand characterisitcs
- .- cant control extranous variables
- .- difficult to replicate
expiremental type - quasi study
- IV cannot be randomly assigned or manipulated
- innate characteristics of particapnts involved - gender , age , inteligance
- .+ same strenghts as what its conductions (lab,feild,natrual)
- .- cant not randomly allocate - big prob with particapnt variabels
experimental designs
- independant groups
- repeated measures
- matched pairs
expiremental designs - independant groups
- completly sperate groups - only expirence the iv once
- randomly allocate the groups
- baddley ( 1966) & bandura (1963)
- (+) no order effect
- (+) lower demand characteristics
- (-) particapnt cariables may effect dv - reduces validity
- (-) less economical - more particapnts mean more time and money
expiremental designs - repeated measures
- one group does both conditions of the iv
- wastons & Rayner (1920) & Baddley (1966)
- (+) particapnt variables are controlled
- (+) more economical - less particapants
- (-) demand characteristics
- (-) order effect
expiremental designs - matched pairs
- siminalr to independant groups but paired up on similar characteristics
- bandura doll study
- (+) reduces partipant effects
- (-) extremely time consuming
correlation
- a measure of the extent to which two covariables are related
- shown on scatter graphs
- postive
- negative
- curvileanar
- zero correlations
- correlation coefficent - numerical value that tells you the strenght of your correlation
evaulations of correlations
- .+ doesnt manipulate the variables - just lets it happen - real life
- .+ quick and cheap
- .+ varriables that may be unethical or impractical can be tested
- .- correlation does not mean causation - cant establish caustaion
reliabilty
- refers to the repability/consistancy of the findings
- internal - is the testing the same thing
- external - can i repeat the method and get the same results
validity
- refers to the creditbility of the research , are the results full reprosenting what the research set out to be
- interal - is the dv because of the change in the iv - is it genuine ?
- external - to what extent can the results be gernralised (ecological & population)
improving internal validity
- control all extranous variables
improving external validity
- experiments in a more natural enviroment
- reprosentive sample (random & stratified)
improving external reliability
- standarised instructions
- control extranous variables
- structured interviews
improving internal reliability
- do a piolet study - remove things that dont work
- operationalise catergories
- train observers
mudane realism
- a measure of external validity
- realtes to ecological validity
- when particapnts belive a task is so real they act as if its real
significant figures
- how many numbers are there
decimal place
how many numbers pass the decimal place
probability
- number of particalular outcomes / possible outcomes
- conditional probability - the probability of an event if something else occurs
nominal data
- catagorie data
- bar chart
- mode
- quantatative data
ordinal data
- data that can be put in order
- histogram
- median
- no meaning
interal data
- exact measurements
- scatergraphs
- meaning
rules of drawing a graph
- title with dv & iv
- correct graph form
- iv of x-axis & dv on y-axis
- apptoptiate scale
- labbeled axis
- graph is accurate
evaluations of mean
- .+ reprosentive - includes all the data
- .- sensitive to outlyers - could change the data
evaulations of median
- .+ not skwed by outlyers
- .- not reprosentive of all scores
evaluations of mode
- .+ wasy to calculate - quick
- .+ not skwed by oulyer
- .- not reprosentive
standard deviations
- a measure of how accurate data is
- large sd is varied
- small sd is consistent
steps for normal distributiosn
- examine visually- look to see if data is aorund the mean
- calculate measure of central tendacy - mean , mode , median
- plot frequency - plot on histogram to see if curved
sampling
- the process of selecting a represntive group from the population under the study
- genralise results
- sampling bias could occur when characteristics of target population
random sampling
- equal chance of being seletec
- assigned numbers to people and put in random generator
- (+) reprosents the populations and elimates sampling bias
- (-) diffuclt to get reprosentive sample
- (-) small minorty groups will be missed
sytematic sampling
- nth member
- sampling frame is produces
- (+) unbiased if list in randomised
- (+) easy and quick
- (-) if list is not randmised there will be a bias
stratified sampling
- strata is made - researcher will identify different groups
- certain percentage of strata used
- (+) avoids problem with misinterpration
- (+) reduces number of particaonts needed
- (-) hard to know how many of each group is needed
- (-) hvae to accses all stratas
opportunity sampling
- selecting people who are willing and availabe
- (+) quick and easy
- (+) more ethical bc researcher isnt judging people
- (-) may not be reprosentive
- (-) researcher bias
volenteer sampling
- particapnats selceting themselves to take part
- (+) large sample size
- (-) share similar characterisicts - extrovert - unreprosentive sample
infromed consent
- must outline what the research is about
- ask for their consent
- dont need it if its something that happens in everyday life
debrief
- must be thoroughly debriefed at the end of a study
- exaplin genral idea of research
- must told if they were decieved and why
protection from particapnt harm
- particapnts cant be caused distress
- be harmed pshycially or mentally
- cant make a risk greater than everyday life
deception
- should ne avoided where possible
- told something else so they dont guess the aim
- avoids demand characteristics
- true nature of the research should be told at the end
confidentuality
- data gained must be anonymous unless they give full consent for their names to be used
right to withdraw
- must be aware they can withdraw at any point
- also have final oppuritnty to leave
clive wearing - case study
- aim - to see how brain damage affects short and long term memory
- case history - lost his memory from disease
- reseach methods - interviews , experiments , observations
- findings - long term and short term memory are in differnt areas
- bad ethics - no informed consent , no debrief , no right to withdraw , already stressed
genie - case study
- aim - wether you could learn things after the critical time perdiod
- case hisorty - what her family did to her - locked in a room for 12.5 years
- research methods - experirments , observations , interviews
- finding - you cant learn grammar after the critical time period but you can learn language
- bad ethics - no infomed consent , no right to withdraw
case studies
- detailed study on an indicual/group who have gone through unusal expirences
- longitunal
- not expiremts bc there are no varibales
- get case history
what to use in a case study
- interviews - quantitaive or qualitive
- secoundary data - medical history ect
- case history - what they were like before
- observations- see cause and effetc
- experiments - see how they behave in certain situations
evaluations of case studies
- (+) gather real life data - real life situations , ecological validity
- (+) more ethical - cant manipulate , researcher things that arent ethical to do
- (+) lots of different trype of methods - postive and negatives of one counterbalance another
- (-) relability - cant replicate
- (-)loose objective - become to close with the perosn and loose aim
- (-) cant generlise
content analysis
- a method which transforms qualitive data into quantative data
- aim is to summarise the detailed data into catagories so be anaylised and conclusions can be drawn
- content - infomation in a media form that can be re watched
- tally everytime you see the groups
how to do content analysis
- first the content is re watched
- the qualative data is then broken down into catagories (coding)
- researcher then re watches the content and sees how often the concept comes up
- then uses this data to make a conculusion
thematic analysis
- keeps data qualative
- first content is watched
- the qualative data is not changed into catagories but researcher looks for recuring themes
- themes are made into broader catagories (sterotyping)
- researcher then will test the validity of the themes by collecting new data and see if it fits in
evaluations of content and thematic analysis
- .+ no practical and ethical issues since already in public domain
- .- the catagories are very subjective and people may tally things that other dont see as that
- .+ makes easier to draw conclusions
- .+ we can change the data to what we think is the best for the situation
inter rater reliability
- level of agreement between different rater
- strong = good agreement
- weak = bad agreement
- increase it; comes up with catagores seperatly , compare and agree with catagories , clearly define everthing
- checking it; conduct seperatly, compare tallies and make sure reliable , statistical test
peer review
- a process by which the research is scrutinszed by those who are experts in the same feild, this is done before being publishes and is considered in terms of validity, orginality and significance
what does peer review prevent ?
- prevents scientific fraud
- stops people from stealing work
- stops bad research
scientific fraud
- an act of deception or misreprensation in scientific work
- made up of fabrication, falsifications and plagiarism
- fabrication - intentional misreprensation of research results in making up data
- falsification - is changing or omitting data or results. it means the research is not accurate
- plagiarism - presenting someones work or ideas as your own
process of peer revies
- researcher finishes and sends mauscript to editor
- editor then sends manuscript to expert psychologists in the same feild
- peer reviwers then look for validity, creditbility and orginality
- peer reviwers then send it back to editor with recomendations
- then they accpet and publish or reject or if some scientifc standars are made it well be sent back to researcher to make improvements
strenghts of peer review
- promotes high standard - helps reputaions of psychology
- gets rid of scientific fraud - people know their work will be scrutified
- promotes scientifc process
- accurate knowledge - worthwile as shouldnt be false
weaknesses of peer review
- publications bias - if the reviwer doesnt want it out for some reason (like controdicts their work) they could influence wether its published or not
- the file drawer effect - refers to the practice of researchers filing away or not publishing studies which find nothing of statistical significance
- ground breaking research gets buried - challenging research
- done anyomously - could savatage bc want funding
h1
experimental hypothesis
h0
null hypothesis
descriptive statistics
- any statistics which describe data simply, include measers of central tendancy and measure of dispersion
inferential tests
- mathimatical tests which allow pschologists to accepy or reject their hypothesis
significance
- if not strong have to accpet it happened by chance
- p value greater than 0.8
- 0.05 means a fluke and happened by chance
type 1 error
- where the null hypothesis is rejected and the alternative one is accepted
- researcher claims to have found a significant effect when there isnt one
type 2 error
- where the null is accpeted and the alternative is rejected
- researcher claims there ws no significance where there was actually one
process of testing
- data studied using decriptive statistics
- choose which statistical/inferential test to use
- statistical/inferential test done on the data
- calculated value obtained
- calculated value compared to critical value
- is it significant
choosing statistical/inferential test
- is it association or difference - correlation/related = assocation , comparing/differnce=differnce
- design - matched pairs and repeated mesure = related test , indpeant groups = unrealted
-
level of data
* nominal - catagory/frequency data - lowest level of data - bar chart
* ordinal - order/rank/scale/score - doesnt have eqaul intervals
* interval - unit of measuremnt - standardised -line/histogram
SIMON COWEL TABLE
how to decide significance frim crictical value tables
- was the hypothesis direction or non directions
- the number of particapants (n value)
- level of significance (p value) (0,05)
how to write significance question
- significant or not
- because the calc vaule it …. less/more than the critical value of …..
- critical value is …. becuase the p vaule is …. and its one or two taled then the n value is …..
- accept or reject whicherve hypothesis
sign test
- plus or minus two variables
- remove any particapnts who have no difference
- lowest value of minus or plus is the calc value
- compare calce value to critical value - using table
- if not less than or more than or equal to then reject alt and accpet null
rearch paper
- abstart
- introduction
- method
- results
- dsicussion
- references
research paper - abstract
- like a blurb
- first section front page
- 150-200 words
- includes summary of all elements
- written last
research paper - introduction
- starts broad with lots of infomation
- ends with hypothesis and is very specific
research paper - method
- how research was carried out - another researcher could do it word for word
- design , sample , materials, prodecure , ethis
- list of start to finish
- standardised instructions ect
research paper - results
- summarise key findings with descriptive and inferential stats
research paper - discussion
- summarise findings
- discuss findings
- indentify limitaions
- wider implications discused
research paper - refernces
- reference everything you used ot help you
- how to write one
1. surname
2. date
3. book title
4. city
5. publisher
how to write consent form
- look like a form (leave space for signiture and date)
- inclue 50% of procedures and genral purpose and duration
- inclue 50% ethics - results will be confidentral , not harmed , fully debriefed, tell right to withdraw
- tell they can ask questions at any time
how to write standardised instructions
- formal language
- explain procedures relevant to particapnts
- give enough detail that they shouldnt need to ask more questions
- include a check of understanding
- helpful to remind them that they can withdrawl at anytime
how to write a debrief
- thank them
- tell aim of study
- condtions used
- explain any deception
- any ethical considerations
- who to contact about findings or to complain (make up number)
- give particapnts last change to withdraw
design an experiment
- hypothesis
- experimental type
- experimental design
- sampling
- materials
- procedure
- ethics
- data anyalyis
design an observation
- type of observation
- use of behavioural catagories
- how many observers
- how to sample
desgin a correlation
- co-variables
- first or secound hand data
- level of data
- graphs
- inferental test