research methods Flashcards
p2
1
Q
wt is content analysis
A
- indirect obs method used to analyse human behav
- investigating thro studying human artefacts (things ppl make)
2
Q
how 2 perform content analysis 4 steps
A
- decide research q
- select sample (random/systematic) from larger quantity of all possible;e data
- coding (researcher decides on categories/coding units 2 b recorded)
- work thr data (read thr sample & tally n. times pre determined categ appear)
- data analysis - can b performed on quantitative data to look 4 patterns
3
Q
wt do the coding units need 2 be for cotenet Ana;ysis
A
operationalised- as clearly defined as possible to reduce subjective interoperation
e.g directly threatening violence instead of aggressive statements
4
Q
AO3 content analysis pros
A
- “artefacts” not usually created 4 research but from real world.
- means high external validity & findings shld be generalisable 2 other real world situtatuobs
- as artefacts from real world often ez to gather sample
- other researchers shld be able to replicate content analysis using same coding units / behavioural categories and ez access to sample
5
Q
AO3 content analysis cons
A
- researcher will often need 2 interpret subjective text may lead 2 researcher / observer bias coz may interpret in way 2 support study
- data is created 4 purposes other than research as data not created under controlled situations may lack validity
e.g written historical diaries may not contain accurate record of past but contain inaccuracies
6
Q
wts thematic analysis
A
- researchers start by attempting 2 identify deeper meaning go text by reading first and allowing themes to emerge
7
Q
how 2 perform thematic analysis
A
- collect text/ turn recording into text by transcription writing it down
- read texts/ transcripts first 2 spot patterns that can be coded & collected
- reread transcriptions looking for emergent themes
8
Q
wts important ab thematic analysis
A
- u make clear themes r not predetermined by researcher but emerge from text
9
Q
AO3 thematic analysis
A
- as theories come after discovery of themes can b argued thematic analysis stops researcher bias
- higher external valid
- ez to sample
- ez replciatie
- subjective interpretation
- data not created in controlled cobiditions
10
Q
case study def
A
- range of data collected from individual group/institution.
- main data collected via interviews & obsv
- content analysis can b perfumed on written evidence & even experimental techniques can b included
11
Q
wts a short case study called
A
snapshot
12
Q
wts a long case study called
A
longitudinal case study
13
Q
case study AO3 pro
A
- as case studies r in depth mostly qualitative inves range of data collected results in realism. holistic approach 2 conducting research favoured by humanist psych coz gives depth and detail ab true exp.
- case studies can b used to generate hypothesis that can b tested experimentally.
e.g testing 4 existence of Brocas area w use of highly controlled FMRI SCANS ON large n ppts
14
Q
case study AO3 cons
A
- ppt in case study r unique, thus exact replications cant b conducted 2 check reliability of findings
-can suffer from researcher bias as they find which findings to include/exclude when writing up also may form personal rsp coz working over long time thus getting in way of main objective of study
- CA interviews form large part of case study and this depend on memory and memories recalled may b inaccurate or answers= 4 social desirability
15
Q
wts an aim
A
- a clearly phrased general statement ab what the investigator aims 2 study
16
Q
wts a hypothesis
A
- precise statement including levels of iv and dv or both covariables 4 correlational study
17
Q
wts meaning of operationalise
A
- carefully stated variables demonstrating EXACTLT how they r to be measure e.g dependent v would b “n. words recalled” not “recall”
18
Q
wts null hypothesis
A
states there r no change in measurement of DV as result of manipulating IV
19
Q
wts alternative hypothesis
A
states there is a change in measurement of dependant variable as result of manipulation in the IV
20
Q
wts hypothesis testing
A
- data collected & statistical testing conducted on data
- provides evidence if evidence is strong enough null hypothesis can b rejected & alternate accepted
21
Q
how can alternate hypothesis b written
A
either
- directional 1 tailed
- non directional 2 tailed
22
Q
wt does non directional hyp state
A
- is a diff in measurement of DV as result of manip of IV) but not the direction the results will go
23
Q
wt does directional hyp state
A
- is diff in measurement of dependant v & states which direction results will go
24
Q
example differentiating between directional and non directional
A
NON DIRECT: there is diff in number in words called by ppt in green or blue
DIRECT: there IS an INCREASE in words called by ppt in green or blue
DIRECR states how much of a diff
25
target pop def
- every member of group that investigator plans 2 study as target pop could contain millions of ppl who cant all b studies
26
random sampling
- each member of target pop has a MATHEMATICALLY equal chance of being in experiments sample
27
random sample AO3
- avoids researcher bias as researcher cannot choose ppt they want to form the sample to give preferred result
- could pick unrepresentative sample tho by chance so few members of 1 gender or no members of minority group
- can b diff and time consuming to get large target pop
28
systematic sampling def
- ppt chosen from list of target pop every nth (5th 10th or 100th) ppt chosen to form sample
29
how 2 conduct systematic sampling
1. researcher needs full list of entire pop
2. researcher reads down list selecting every nth ppt to form sample
3. process continues until sample required is chosen
30
AO3 systematic Samp;ing
- avoids researcher bias researcher cant choose ppt they want in sample
- if all existing list of target pop e.g school can b quick method of selecting sample
- method by chance could choose unrepresentative sample
- if target pop size too big then maybe diff to conduct
31
wt is opp sample
researcher directly asks available members of target pop 2 take part in research likely 2 be individuals researcher has ez access 2 and familiar w
32
how 2 conduct oppurtu Samp
1. researcher directly asks any members within target pop ti take part in research
2. any individuals who agree 2 rake part r added 2 sample till n ppt required is met
33
AO3 opportunity sample
- fastest way 2 get sample 4 psych research reducing time & costs compared 2 complex sampling methods
- coz researcher asks ppl 2 participate might have researcher bias
- likely 2 be unrepresentative as ppt chosen r type of ppl who researcher has ez access 2 e.g uni students
34
wt is volunteer Samp
ppt offer 2 partake after finding out ab research
35
how 2 conduct volunteer sampling
1. ads where likely 2 be seen by target pop e.g gym for gym rats
2. adv contain details of researcher whom will enrol volunteer into. sample where contacted
36
AO3 volunteer Samp
- ads can reach large n. ppl
- ez sample 2 collect as ppt putting themselves forward
- sample may nit b generalisable to target pop coz of volunteer bias. coz choosing ppl who have certain interest their views likely 2 b diff to target pop
37
stratified sample def
- by selecting from within strata characteristics of ppt within sample r in same proportion as found within target pop
38
how 2 conduct stratfied samp
1. strata r identified along w proportion in target pop
2. random sampling used 2 select n. ppt required from each stratum
39
AO3 strat sampl
- sample is representative of larger target pop meaning results found within sample shld be generalisable to get target pop
- ppt randomly chosen avoiding researcher bias
- not every possible characteristic can b included in stratified samp
- researcher decided which strata important which may result in bias
- time consuming to establish strata
39
repeated measures how does it work
- SAME ppt complete 2 levels or more of independent variable
40
independent groups design
- DIFF ppt complete 2 levels or more of indepden variable
- ppt randomly allocated 2 each design avoiding researcher bias
41
wt kind of data does independent group designs make
- unrelated data
42
AO3 indepdemt measures design
- if more ppt w particular characteristic r randomly assigned to 1 of groups can influence measurement on DV (extraneous v
- need double ppt compared 2 repeated measures for same data so time consuming/ expensive 2 find
pro= ppt less likely 2 work out aim in independent whereas more likely in repeated measures
43
repeated measures design
- same ppt comp in each of 2 (or more) experimental conditions
44
wt data does repeated measures design make
- related data
- each ppts score in 1 condition can b paired w data point (own score) in another condition
45
ao3 repeated measures
- order effects con= taking part in first condition influences performance in 2nd condition e.g worsen due to tirenedd or boredom
- or can figure out demand characteristics
-pro= half ppt needed compared to independent groups design
46
how 2 control orderr effects of repeated measures design study and whats the format used
1. counter balancing attempts 2 control But doesnt eliminate order effects.
USING ABBA format half ppt compl condition A first & B second
- other half of sample start w condition B then A
47
wts matched pair design AO1
- diff ppt complete in each of 2 conditions of exp
- ppt first assessed & ranked on a characteristic (e.g aggression) then top 2 ppt are randomly assigned 2 sep conditions
48
wt kinda data does matched pairs make
- related
- each ppt score (data point) in 1 condition can b paired w data point (ppt matched 2 them) in another condition
49
AO3 matched pairs
pro= reduced ppt variables as ppt r matched on relevant characteristic
pro= no order effects as ppt only takes part in 1 condition
con= takes longer 2 set up than other exp designs
- 2x as many ppt as repeated measures design
- ppt r similar but not identical so may still b ppt variables between conditions that influence dependant variable
50
wt happens 2 variables in correlational study
- want 2 measure whether variables r associated
- co variables r measurements that have been made that r compared 2 see if theyre associated
- correlational study 2 measured co variables e.g height weight r assessed 4 relationship
- relationship between 2 variables doest mean causal relationship
51
wt does positive relationship mean between co variables
when one co variable increases so does the other
52
wt does negative rsp mean between co variables
- as measurement of one co variable increases measurement of other decreases
53
wt r demand characteristics an example of
- extraneous variables
so in essay can mention this for eval
54
examples of extraneous variables
- demand characteristics
- ppt variables;es (may know more than others on task)
- situational variables (weather noise temp)
- order effects fatigue
55
confiding variables
- changes systematically between levels of IV
meaning as u change the indepdenr variable ur also changing the counfouding one
56
how 2 control ppt variables
- random allocation= removes potential bias when matching ppt together w relavnrt characteristics. but more can be matched coz of chance
- matched pair design ( coz relevant charctaeritcis r measured prior 2 stay and top 2 ppt r randomly assigned 2 separate groups)
57
how 2 control situational variables
- standardisation of procedures
makes sure all ppt have same exp aside from variation in indepdnet variable
58
how 2 control demand charatceristuds
- single or double blind trials
1. single = ppt blind 2 aims of exp
2. double= both researched and ppt blind
59
wt is a pilot study
- small scale version of main research study conducted before main study
- using same metholdofy/ data collection/ data analysis
60
wt r the aims of a pilot study
- improve quality of main research study by assessing exp of ppt in the pilot
e.g=
- unexpected extraneous variables tht need 2 be controlled
- ppt may reveal they discovered aim & acted accordingly
- in observational study= may b clear that behavioural categories r not fully operationalised
61
wt is peer review
- before publication researcher needs 2 show their written up study to other experts in their field
- unaccounted for variables may b identified that have confounded 2 the results.
- may mean study lacks internal validity + needs to be repeated
62
wts retroactive consent
1. researcher asks 4 consent after ppt has taken part in study
- if ppt says np then data destroyed
63
wts presumtive consent
researcher asks group similar 2 sample if they would agree 2 take part in research
if group agrees then researcher assumes that ppt would would also agree
64
AO3 cost benefit analysis
1. if research requires deception/ risks harm 2 ppt researcher can perform cost benefit analysis 2 assess potential harm 2 ppt and comparing this 2 potential benefits of the research 2 society
CON= often the true value of study isnt known 4 many years after a study has been conducted so can b diff to accurately conduct cost benefit analysis
65
peer review steps
1. submit paper 2 academic journal journal then sends paper to independents experts "peers"
2. peers consider quality of paper including research design & methods used. considering if there were extraneous variables that were uncontrolled. data analysis also checked aswell as deciding if results actually support conclusions draw by author
3. peer decide if they can recommend the paper for publication if they reject then they suggest changes or they jus reject
4. if accepted then journal editor makes final decision if paper will be published
66
AO3 peer review
pro= work and facts will b double checked 2 ensure anything released is factual & trustable
- quality and amount of peer review work can help 2 assess the quality of diff academic institutions by giving each institution a research rating. universities& research labs w history of producing peer reviewed articles r likely 2 be favoured by students & gain *government funding 4 future research*
CON=
- journals may struggle 2 find suitable peers w enough knowledge in new or very specialised areas of scientific search w few experts
- professional rivalry exists can result in peers rejecting papers jus so they can publish it first
- peers may feel unable 2 chLLENGE WORK in fear of negative response
67
how 2 control peer review 2 be fair & AO3
- single blind peer reviews= author doesnt know who peers r
- double blind author & peer dont know
but blinding is problematic can figure out who it Is via writing style and anonymous peers may b uncecesarily harsh
also there is publication bias towards positive results so negative results r not published
- also slow process
68
psychology and the economy examples
- psychologically healthy & well educated pop may b more productive producing more & higher quality services
- generation of tax revenue pays for public survives depend on taxing income which depends on ppl employed in producing and consuming goods 2 be able 2 continue paying taxes
69
maternal deprivation x economics
- research by Harlow bowl by ainsowrth rutter show emotional & cog develop can b damaged if deprivation in early life
so on economy= midwives use psych knowledge gained ab developmental needs of young children when giving advice 2 mothers.
- hospitals improved policies for family so more visiting hours for young children & more maternal contact
- adoption agencies prioritise quickly placements for infants w loving families
= these changes improve development of children which improves economy as they will b more productive and better adjusted 2 demands of adult life.
AND less likely 2 require costly mental health support as adults esp under NHS wink wink
70
social influence x economy
- milgram ( obedience) asch (conformity) demonstrate how situational pressures cld influence individuals behaviour
- government have used these findings 2 change behaviour of citizens
NUDGE unit BIT works within UK government to implement psych principles
- nudge unit sent letters 2 ppl who didnt pay tax lets included stats showing majority of ppl paid tax on time & Ade clear they were not part of majority group. this increased taxes being paid within 23 days by more than 1.6 mill
- projects have improved economy by increasing amount of money by increasing amount of mon available 2 gov 2 spend on other servibves b
71
Facebook x conditioning
- pavlov operant condi skinner classical conditioning
- social media x comp game companies manipulate users 2 engage w their products more by providing virtual rewards e,g likes followers and levels
- act as social rewards that result in impulsive use.
improves economy as companies display more ADVERTS making more money.
estreme compulsive behaviours can result from variable ration reinforcement giving rewards on an unpredictable schedule
- variable ration reinf0r also same technique used by casino slot machines
72
neurotransmitters x economy
- development of ssris
= cheap effective treatments 2 help reduce symptoms and help w day 2 day life.
improves economy as ppl function adequately and work so money used for businesses and increasing n. workers paying taxes
reduces costs on NHS for therapies
obvs has issues tho
73
memory x economy
- justice system. Loftus showed effect memory can have on EWT serious limitations.
thus memory research helps economy by reducing costs of imprisoning ppl who dont need 2 be in prison wrongfully
- and arresting ppl for crimes that if still committed would b costly e.g theft fraud or killing ppl in potential work force
74
wt is reliability
- another way of saying results from study r consistent. if researchers replicate study they'd get similar results
75
wt r the 2 types of reliability
external= extent 2 which measure is consistent when repeated
internal= extent 2 which diff parts of exp r consistent w themselves
76
wts split half method used for and how 2 conduct
testing internal reliability
- assesses measures that test 1 variable w multiple questions e.g questionnaires or tests of IQ
- split test intio 2 parts
- ppt complete both parts
- test strength pf correlation between 2 parts of measure
- strong correlatio= internal reliability
77
wts test resets for and how to conduct
testing external reliability
- repeat study using same procedures/ measuring devices at diff times & test correlation between the 2 versions
78
wts inter rater reliability
tests external reliability
- 2 or more observers record behaviours during same observation using same behavioural categories
then they test the correlation between each tally of behaviour 2 identify if behavioural categories r appropriately operationalised
then test w test of relation
79
how 2 improve reliability
1. observation: improving training given 2 observers can improve their accuracy in assigning particular observed behaviour 2 correct behavioural categories. operationationlisation= process of clearly defining behavioural category .. pilot studies can help identify
2. using structured interviews rather than unstructured interviews. then all ppt will have same exp & can compare rsults
3. use closed questions 2 reduce range of possible responses & if theres a set test for wt u need 2 measure use that
4. use standardised experiments
80
wt is internal valid
- questions cause & effect relationship between change researcher made 2 independent variable & observed change in dependent variable. if change in DV was influenced by any other factor than IV = findings lack internal validity
81
wt is q external validity
qs if study can b generalised beyond study in outside world
82
wt factors affect internal validity
1. social desirability
2. demand charactis
3. investigator effects
4. uncontrolled extraneous variables
83
wt validity can gender bias link 2
- population validity = extent to which sample used is representative of target population
84
wt is criterion validity
- if we can compare the data from a test to another measure of same variable
- & identify a correlation
- confidence in validity of a test increases
85
wt r 2 types of criterion validity & how to test
- concurrent = extent 2 which data from newly created test is similar 2 established test of same variable conducted at same time .
- tested correlation tests above, high concurrent validity= if strength of correlation is 0.8+ or higher
predictive=
- extent to which performance on test can predict future performance e.g gcse predict Alevel performance
86
how 2 improve validity
- random allocation- control ppt variables
- standard procedures- controls extraneous vari
- counterbalcnin- order effects
- singe & double blind
- peer review
- external validity improved by demonstrating that findings r generalisable
replication= cause & effect relationship observers in original study isnt jus limited to 1 experimental set ul
87
wts empirical method is psych
- process of collecting data from direct exp in psychological research--> data we gather from direct observation of ppts
- includes observation but also experimentation self report case studies & content analysis
88
objectivity meaning
- data shld b collected & interpreted in ways that avoid bias, meaning data isnt influenced by researcher opinion / expectation
research that's been affected by bias produces subjective conclusions
89
4 ways of improving objectivity
- systematic data collection= data gathering carefully planned and designed / researcher should use established qs or tests
- double blind= researchers who dont know researcher aims collect data
- peer review= 2 identify biased research e.g researchers making conclusions that r not supported by data & stopping that work from being published
- control variables 2 compare
90
wt is falisfiability in psych theories
- for a psych theory to be scientific needs 2 be constructed in a way where it can b empirically tested= means theory can be tested in a way that demonstrates its not true
- think ab black swan theory someone said all swans were white but second black swan observed = theory had been falsified
- Freuds theories r criticised coz they cant be empirically tested therefore they cant be falsified
91
wts a paradigm shift
- phiopsher suggested scientific fields develop in a series of "scientific revolutions" known as paradigm shifts
92
wt is a paradigm
when scientists from each scientific field make assumptions these r called paradigms & they gather evidence 2 support their shared views
93
wts a paradigm shift
when new contradictory evidence to initial paradigm (set of assumptions made from researchers and diff scientists) is initially rejected coz it goes against initial beliefs
but once enough scientific evidence created everyone changes belief
old paradigm rejected new one accepted
= called paradigm shift
94
wt r some paradigm shifts in psych
- early psychs used introspection 2 develop theories of the mind freud used case studies Wundt used scientific experimentation. this was paradigm shift away from earlier religious beliefs where explanations were coz of "sin"
- movement from psychoanalytic approaches 2 behaviourism. coz researchers rejected study of internal mental processes seeing them as unscientific coz they weren't observable. behaviorusist instead focused on fully observable stimulus-response mechanisms. behaviourists used large-scale & highly controlled studies 2 provide strong enough evidence 2 theories
- neuroscience was one fMRI scans allowed u to directly observe functioning of brain mapping mental processes onto brain regions
95
wt r stages of scientific theory construction
1. observation = psych start observing naturalistic behaviour in real world
2. construct testable operationalised hypothesis
3. conduct experiment & gain experimental data. if data signif using stat tests, researcher can claim cause x effect relationship between variables
4. propose theory that explains results: researcher constructs theory that can explain sign results gained in exp
96
wt is top down method
1. researchers start w an established theory & develop hypotheses that test 1 of theory's assumptions. result of this= support theory our discredit
97
wts an abstract
1/2 paras at start of scientific report acts as summary or report.
- includes aim & tested research hypothesis
- give details ab how researcher conducted study (method)
who took part (sample)
and major results & conclusions
- speed up researcher process so can quickly find info
98
how 2 turn qualitative data in2 quantitative data
- content analysis
99
wts primary data
1. researcher = responsible 4 generating data
created 2 ans research q
common way 2 collect primary = observations interview questionnaires vase studies
100
wts secondary data
- using second hand data
- data all collected by 3rd party
-e.g old government/ business statistics
101
AO3 PRIMAry and secondary data
PRO secondary= easy 2 analyse coz all collected reduces time
CON SECONDARY= decreeased validity data not answering research q directly. researcher has no role in collecting therefore cannot control bias/ensure taken directly
PRO= primary increased validity a data collected ans question DIRECTLY. increased validity as researcher can control data collectionnprocess
CON PRIMARY= time consuming expensive
102
wt is meta analysis
- process of collecting & combing results of range of previously published studies asking similar research q.
- data collected is compared & reviewed together
- part of it is 2 statistically combine all data to produce overall effect size & conclusion
103
AO3 meta analysis
PROS:
-1 large sample size of meta analysis produces results that r more statistically powerful than studies w smaller number of ppt
- 2 as it looks at large n. of patterns small studies w biases wont contribute to overall pattern of results making meta analysis more trust worthy than initial study
-3.. studies testing same variable in various contexts e.g across cultures which can b compared revealing unexpected relationships
CONS:
1. all weaknesses of secondary data. researchers has no control over data collected so we dk if there were any biases
2. studies showing statistically signif result r more likely 2 b published in meta analysis than studies w non signif results as they'll make study look better (FILE DRAW PROBLEM)
3. choice of which studies included / excluded can be biased
104
wts the file draw problem
1. studies that only show signif result r more likely 2 be published than studies that dint
105
wt do descriptive statistics include
- measures of central tendency
- measures of dispersion (how spread out datas r e.g range/standard deviation)
106
AO3 mode
- not distorted by extreme outliers
- mode helpful 4 discrete numbers
- giving modal group is only way of giving average of data in categories
CONS
- can b no modes if every value is diff / multiple modes = likely in small data sets
- doesnt include all values in calculation so it is not as sensitive as mean measure of central tendency
107
AO3 median
pro
median = central value so not affected by outliers
ez to calculate
con
doesnt use all scores in calc so not as sensitive as median
if there r even n values then we use avg of 2 which is not an actual recorded value
108
AO3 range
pro=
-ez 2 calc esp compared 2 standard deviation
con=
- extreme scores can easily distort
- range doesnt show if scores r clustered around mean or more evenly spread
109
