psychology revision - research methods topic (1)
Post on 01-Jun-2018
224 Views
Preview:
TRANSCRIPT
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
1/17
A2 Psychology Revision
Research Methods
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
2/17
Key Words
Empirical Data Data which can be physically seen.
Objectivity Data and research which is not subjective or bias.
Peer Review A third-party/external body within the sae !ield" validates research.
Abstract A suary o! a study/research.
Experiments #arrying out a test under controlled conditions/environents.
Self-Report et!ods Participants are as$ed to report their own attitudes" abilities or
!eelings.
Pilot St"dies A sall-scale trail o! a study to test all aspects o! the design.
#nter-rater Reliability %wo or ore observers are present throughout the analysis o! data.
$est Re-test Reliability A test is repeated a second tie using the sae saple and test.
Split-%alf et!od A test split between two di!!erent situations to see i! the sae
results are !ound.
Stratified Samplin& A group o! participants selected according to their !re&uency within
a population.
#nferential Statistics
$est
Procedures !or a$ing in!erences about the population !ro which
saples are drawn.
Probability A nuerical easure o! the li$elihood/chance that certain events
will occur.
Si&nificance A set o! !indings are su!!iciently strong enough !or us to accept.
'!ance %he extent to which soething occurs randoly.
Reflexivity %he process o! sel!-exaining &ualitative research.
'ontent Analysis An indirect observation o! soebody through their wor$ 'this is
usually presented through trends.
("ll %ypot!esis A hypothesis stating ()o !indings*.
Alternative %ypot!esis A hypothesis stating what you except to !ind !ro a success!ul test.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
3/17
Observations
Proposed Theory
Testable Hypothesis
Conduct A Study
Draw Conclusion
Refne
Observations
Testable Hypothesis
Conduct A Study
Draw Conclusion
Propose Theory
Refne
Science
APP+,#A%,) ) $%E A)OR *EA$+RES O* S'#E('E
0cience is a eans o! !inding out the truth and !acts about our world1 however in order !or this to
be achieved various !actors ust !irst be deonstrated throughout the scienti!ic research1
• Empiricism , ,n!oration is gained through direct observation or experient.
• Objectivity , Personal expectations should not a!!ect what !indings are recorded.
• Replicability , Previously recorded ethods and procedures are re-tested to see i! the
outcoe is the sae.
• 'ontrol , Research attepts to !ind relationships through experiental ethods which
re&uire a degree o! control 'e.g. we vary the independent variable and observe its e!!ects on
a dependent variable. 3e ust ensure all other conditions are the sae 'controlled.
• $!eory 'onstr"ct , A theory is a collection o! general principles that explain observations
and !acts.
APP+,#A%,) %3 $%E S'#E($#*#' PRO'ESS
#nd"ction
%his involves reasoning 'creating a
theory !ro the particular to the general.
An exaple o! this is )ewton4s +aws.
Ded"ction
%his involves reasoning !ro the general
to the particular 'starting with a theory and
loo$ing !or instances that con!ir this. An
exaple o! this is Darwin4s %heory.
Popper ./012 proposed thehypothetico-deductive odel which
suggests a theory should coe !irst as
this can be used to generate hypotheses"
which can be !alsi!ied.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
4/17
Eval"ation of Psyc!olo&y #s A Science3
Stren&t!s 4imitations
• ,n the nineteenth century" psychologists
wanted psychology to becoe a science
because it would enable the to produce
veri!iable $nowledge !or their research.
• Psychologists clai that huan behaviourcan be easured as objectively as any
other scienti!ic variable. #lais that studies
are 5experienter bias4 and experience
5deand characteristics4 are just the sae
as an experienter easuring a atos
behaviour 'as it will change its !orat when
observed.
iller ./502 suggests that psychologists
siply use the tools o! a scientists 'e.g.
content analysis and &uantative analysis"
but the essence o! science has eluded
the.
K"!n ./672 suggests psychology is not a
science because it doesn4t have any single
paradig 'a shared set o! assuptions
e.g. the behavioural" cognitive and
biological approaches.
4ain& ./682 suggests that psychology and
science have very di!!erent goals.
Psychology uses an ,diographic approach 6ta$ing into account individual cases 'he
suggested that the only way to treat
individuals was to view their own case"
whereas science generalises using a
noothetic approach.
7uhn suggested the scienti!ic ethod
9alidatin& (ew Knowled&e:Researc!
8P+A)A%,) ) $%E RO4E O* PEER RE9#EW
Peer review is the assessent o! scienti!ic wor$ and research" by others who are experts within
the sae !ield. %he intention o! peer reviewing is to ensure that any research conducted and
published is o! a high and valid &uality.
Peer review serves three ain purposes1
• Allocation of Researc! *"ndin& , Research is paid !or by various governent andcharitable bodies" there!ore public bodies re&uire reviews to enable the to decide which
research is li$ely to be accurate" valid and worthwhile.
• P"blication of Researc! #n )o"rnals - Peer review has been used when publishing
research into journals to help identi!y and prevent any incorrect or !aulty data entering the
public doain.
• Assessin& t!e Researc! Ratin& of +niversities , All universities are expected to conduct
research which is assessed in ters o! its &uality" !uture !unding depends upon the rating
given.
Eval"ation of Peer Review
Stren&t!s 4imitations
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
5/17
• Peer review gives another way o! validating
data" and provides another sa!ety
precaution again potentially daageable
research being lea$ed into society.
• 9y using researchers within the sae !ield
as the research" it ay produce ore
reliable and accurate results.
Smit! .///2 claied peer review is (slow"
expensive" pro!ligate o! acadeic tie"
highly subjective" prone to bias" easily
abused" poor at detecting gross de!ects
and useless at detecting !raud*.
,t isn4t always possible to !ind an
appropriate expert to review proposed
research" this eans poor research ay be
passed because the review doesn4t
understand.
Reviews do not have to disclose their
identity" which could lead to rivalry.
Synoptic; Peer review !avours publication-
bias as it wants to proote 5valid4 andpositive research.
8P+A)A%,) %3 REPOR$#(< PS='%O4OPER#E($S
All experients involved an ,: ',ndependent :ariable and a D: 'Dependent :ariable. %he ,: is
varied in order to see how this a!!ects the D:" thus deonstrating a relationship.
3hen designing an experient psychologists ust ta$e into account the various easures
'experiental designs which can be used" !or exaple1
• Repeated eas"res - ach participant is tested on all o! the ,:s.
• #ndependent
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
6/17
• atc!ed Pairs - Participants in an independent group are atched with participants in
another group on $ey variables.
4aboratory Experiment , An experient carried out within a controlled environent.
Advanta&es;
;igh internal validity
because extraneous variables are
controlled.
#ontrol increases
replicability.
Disadvanta&es;
Reduced external/ecological validity and
undane realis because o! the controlled
environent.
xperienter e!!ects 'experienter-bias anddeand characteristics.
*ield Experiment , An experient carried out within a ore natural environent" although soe
!actors ay still be controlled.
Advanta&es;
xperienter e!!ects are
inial.
,ncreasedexternal/ecological validity and undane
realis.
Disadvanta&es;
Deand characteristics are unli$ely but could
still occur due to the controlled !actors ',:.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
7/17
(at"ral Experiment , An experient which a$es use o! existing ,:s 'no control.
Advanta&es;
)o experienter e!!ects.
,ncreased
external/ecological validity and undane
realis.
Disadvanta&es;
Participants are not allocated to conditions
which could reduce the experients validity.
)atural studies are hard to replicate.
8P+A)A%,) %3 RESEAR'% E$%ODS
Self-Report et!ods , Psychologists use &uestionnaires and interviews to !ind out what people
thin$ and !eel. %hey can be structured or un-structured 'although structured allows !or better
replication" and !eature open or closed &uestions 'open &uestions producing ore un-expected"
&ualitative data. 0el!-reporting ethods do however !eature deand characteristics as
participants are ore li$ely to portray theselves in a positive light.
Observational St"dies , Psychologists use behavioural categories to record particular instances
o! behaviours" or various sapling ethods" such as1 tie sapling 'recording every
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
8/17
%hey do however !ail to collate generalised data" and o!ten use retrospective data 'previously
collated data which ay a$e the results unreliable.
eta-Analysis , %his is an analysis o! a range o! studies within the sae topic/!ield" and can be
used to draw overall conclusions and theories. %his can also be used to assess the reliability and
validity o! general theories" odels and approaches.
8P+A)A%,) %;R SAP4#(< $E'%(#?+ES
Opport"nity Samplin& , Participants are selected by using people who are ost easily available"
this is an uch easier and &uic$er way o! gathering participants" but is not representative and
provides no eans o! drawing generalised conclusions.
9ol"nteer Samplin& , Participants are selected by as$ing !or volunteers" which can provide
access to a variety o! participants" a$ing the study ore representative. ;owever by using
volunteers" the saple is ore li$ely to be highly otivated with !ree tie" ipacting again on
generalising the research.
Random Samplin& , Participants are gathered using a rando nuber techni&ue" the target
audience is !irst identi!ied and then each proposed participants given a nuber" this nuber is
then randoly selected by either a lottery-techni&ue '5drawn !ro a hat4 or !ro a rando nuber
generator. %his type o! sapling produced an un-bias saple because everybody has an e&ual
chance o! being selected.
Stratified and ?"ota Samplin& , 0ub-groups 'strata within a population are identi!ied" and a
predeterined nuber o! participants are then selected !ro these groups 'in proportion to their
representation within the population. ,n strati!ied sapling this is done using a rando-nuber
techni&ue" in &uota sapling this is done using opportunity sapling. %his ethod is ore
representative than the others" and is ore li$ely to produce generalised data.
Snowball Samplin& , Participants are !irst identi!ied" and then as$ed to !ind other suitable
participants theselves1 this process then continues until enough participants are gathered. %his
ethod o! gathering participants ay produce biased saples" because the individuals involved
ay only target a liited variety o! population.
Reliability and 9alidity
8P+A)A%,) ) RE4#A@#4#$=
%he reliability o! any ethod o! easureent 'e.g. interview" &uestionnaire re!ers to how
consistent it easures soething. A reliable test will give siilar results" in siilar circustances.
%here are two di!!erent types o! reliability1
#nternal Reliability , %his re!ers to whether a test-ethod is consistently easuring what it
should be. ,t can be iproved using a split-hal! test 'see below.
External Reliability , %his re!ers to how consistent the test-ethod is over tie" whenrepeated in di!!erent situations. ,t can be iproved using a test-retest ethod 'see below.
As entioned above" we can assess the reliability o! a test-ethod using the !ollowing1
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
9/17
Split-%alf et!od A ethod o! assessing internal reliability by coparing two halves o! a
test-ethod 'e.g. a &uestionnaire to see i! they produce the sae score.
$est-Retest et!od %his assesses external reliability" by coparing the results o! a test given
in two di!!erent situations. ,! a positive correlation is !ound" it has strong
external reliability.
#nter-Rater Reliability %wo or ore observers4 record individual scores on the test/participants"
a!ter coparing the results" i! a positive correlation is !ound there is good
reliability.#ntra-Rater Reliability Re!ers to consistency o! researcher4s behaviour1 should produce siilar
results or a$e siilar observations or carry out interviews in the sae
way in di!!erent occasions.
8P+A)A%,) %3 9A4#D#$=
%he validity o! any ethod o! easureent re!ers to how truly/realistically it easures soething.
A valid test will easure what it is supposed to easure 6 nothing else>
+i$e reliability" there are two di!!erent types o! validity1
• #nternal 9alidity , %he ability o! a study to test the hypothesis that it was designed to test
just by anipulation o! ,:. '?or exaple" i! a anipulated ,: a!!ects an D:" the investigation
has good internal validity.
− *actors affectin& internal validity deand characteristics 'resolved using single-
blind techni&ue" social desirability '!ixed through anonyous &uestionnaire"
experienter bias '!ixed through double-blind techni&ue
•External 9alidity , ;ow uch the !indings o! a test-ethod can be generalised to di!!erentsettings" e.g. the environent 'ecological validity and people 'population validity.
3ays o! controlling validity
Sin&le-blind
tec!niB"e
%he experienter $nows what group a participant is in but the participant
hi/hersel! does not.
Do"ble blind tec!niC %he experienter nor the subjects $now which condition they are in
As entioned above" we can assess the validity o! a test-ethod using the !ollowing1
*ace 9alidity %he test-ethod loo$s li$e it is easuring what it is supposed to.
'onstr"ct 9alidity %he test-ethod tests the theory about the variable it is easuring.'onc"rrent 9alidity %he test-ethod has a strong correlation with an already established
test.
'ontent 9alidity %he test-ethod tests everything 're&uired.
$emporal 9alidity %he test-ethod can be endured over tie and are not era-dependant.
Pop"lation 9alidity %he test-ethod can be generalised to the whole population.
'ontext-Ecolo&ical
9alidity
%he test-ethod can be generalised to the general/natural environent.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
10/17
Et!ical #ss"es #n Psyc!olo&ical Researc!
,00@ ) E$%#'A4 #SS+ES W#$% %+A( PAR$#'#PA($S
$!e @PS 78862
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
11/17
#nte&rity Psychologists ust be accurate and honest.
%he code o! conduct 'above outlines ethical guidelines which psychologists should !ollow1 in
addition to this ethical coittees assess research !or its ethical principles. %hey ay punish
psychologists !or not coplying" and norally o!!er education to psychologists about their duties
as researchers.
Socially Sensitive Researc! , Sieber and Stanley ./552 de!ined socially sensitive research as(studies in which there are potential social conse&uences or iplications" either directly !or the
participants in the research or the class o! individuals represented*.
,00@ %3 E$%#'A4 #SS+ES W#$% (O(-%+A( A(#A4S
W!y +se (on-%"man Animals in Researc!3 , %here are !our ain reasons why researchers
would use non-huan anials in research" they include1
•
0oe researchers want to study anials because they !ind the !ascinating" and generallywant to help the succeed in li!e.
• Anials o!!er the opportunity !or greater control and objectivity in research procedures.
• 3e ay use anials when we can4t use huans" !or exaple ;arlow4s Mon$ey
xperient.
• ;uans and anials have enough o! their physiological and evolutionary past in coon"
that we ay justi!y conclusions drawn !ro experients involving one" to the other.
Although this is still under-debate.
oral )"stification of +sin& Animals in Researc! , %here are three ain arguents to the
justi!ication o! using anials in research" they include1
• Sentient @ein&s , %here is evidence that anials respond to pain" however this is not the
sae as conscious awareness. 0oe priates also show evidence o! conscious
awareness.
• Speciesism , Sin&er .//82 argued that discriination on the basis o! species in no
di!!erent !ro racial or gender discriination" and thus suggests the use o! anials is an
exaple o! speciesis.
• Animal Ri&!ts , Sin&ers view is utilitarian" eaning whatever produces the greater !ood
!or the greater nuber is ethically acceptable. 0o i! anial research can stop pain and
su!!ering it is justi!iable.
'onstraints '"rrently in Place , %he Animal Scientific Proced"res2 Act of ./56" re&uires
that anial research should only be carried out in licensed laboratories" with licensed researchers"
on licensed projects. +icenses are only granted under the !ollowing conditions1
• Potential !indings are iportant enough to justi!y the use o! anials.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
12/17
• %he research cannot be done using non-anial ethods.
• %he iniu nuber o! anials is used.
• Any disco!ort or su!!ering is $ept to a iniu.
R"ssel and @irc! ./1/2 also created the three R4s1
® Red"ction 6 @se o! !ewer anials.
® Replacement 6 3here possible use o! alternative ethods should be used.® Refinement 6 @se o! iproved techni&ues to reduce stress.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
13/17
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
14/17
#nferential Analysis Probability and Si&nificance
8P+A)A%,) ) DE*#(#$#O(S
Probability - ,n!erential statistics allow psychologists to loo$ at patterns within results to see i!
they have arisen by chance. ,! it could not have arisen by chance" the pattern is described as
signi!icant.
Si&nificance , ,n order to carry out an in!erential test" you need a null hypothesis ';= which will be
used when results show nothing is !ound and an alternative hypothesis '; which will be used
when results show a correlation/relationship. Results are said to be signi!icant when a large
'signi!icant di!!erence i! !ound between variables and results.
'!ance , ,n!erential tests allow us to deterine whether a pattern in the data" could have arisen
by chance or whether the e!!ect occurred because there was a real di!!erence/correlation. %o do
this a signi!icance level is used 'o!ten pC=.= 6 eaning a probability o! EF that the results are
not due to chance. ,! the research wants to be ore certain" they us a probability level o! pC=.=
'EEF or pC=.== 'EE.EF.
8P+A)A%,) %3 +S#(< #(*ERE($#A4 $ES$S
$ypes of Data , 3hen deciding which test to use you ay need to identi!y the level o!
easureent that was used" these are explained below1
(ominal
Data which can be
categoriGed.
Ordinal
Data which can be ordered and
their position noted'di!!erences ay not be e&ual.
#nterval
Data easured in e&ual
intervals 'coparativenubers/scores are ta$en.
'alc"latin& t!e 'ritical 9al"e - A test statistic calculated !or any set o! data is called the
observed value 'because it is based on the observations/results !ound. %o decide i! this observed
value is signi!icant" it is copared to a critical value '!ound within a critical values table 6 di!!erent
!or every in!erential test.
%o !ind a critical value" you need three pieces o! in!oration1
De&ree of *reedom df2 ,n ost cases this is the nuber o! participants ')" !or
independent group studies there are two values !or )" one !or
each group ) and )2.
One-$ailed or $wo-$ailed $est A one-tailed 'directional hypothesis predicts the direction o!
the results" whereas a two-tailed 'non-directional hypothesis
does not.
Si&nificance 4evel )orally pC=.=.
0oe tests are signi!icant i! the observed value is greater than the critical value" while soe testsare the reverse. %o wor$ out which one we should use" we say1
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
15/17
(,! there is an R" then the observed value should be gR eateR than the critical value 'e.g.
speaRan4s Rho and chi-s&uaRe. ,! there is no R 'e.g. ann-whitney and wilcoxon then the
observed value should be less than the critical value*.
8P+A)A%,) %;R ERRORS
%here are two ain errors" a type one error and a type two error.
• $ype One Errors *alse-Positive2 , %his occurs when we wrongly accept the alternative
hypothesis 'and thus wrongly reject the null hypothesis" the results are due to chance
there!ore we ust accept the null hypothesis.
%his type o! error is caused by a high signi!icance level 'e.g. =F.
• $ype
$wo
Errors *alse-(e&ative2 , %his occurs when we wrongly accept the null hypothesis 'and
thus wrongly reject the alternative hypothesis" the results are accurate.
%his type o! error is caused by a low signi!icance level 'e.g. F.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
16/17
Noinal Data! Chi"S#uared
Correlation!
$ndependent %roups!
Spearan&s Rho
'ann"(hitney )
(ilco*on T
+es
+es
+es
No
No
No
#nferential $ests
8P+A)A%,) ) '%OOS#(< A $ES$
'!i-SB"are
• %he hypothesis predicts a di!!erence
between two conditions or an association
between co-variables.
• Data is independent 'recorded only once.
• Data is noinal 'e.g. in categories not a
percentage.
Spearmans Rho
%he hypothesis predicts a correlation
between two co-variables.
%he two sets o! data are pairs o! scores !ro
one person or thing 'they are related.
%he data are ordinal or interval not noinal.
ann-W!itney U
• %he hypothesis predicts a di!!erence
between two sets o! data.
• %he two sets o! data are !ro separate
groups o! participants 'independent groups.
• %he data are ordinal or interval not noinal.
Wilcoxon T
%he hypothesis predicts a di!!erence
between two sets o! data.
%he two sets o! data are pairs o! scores !ro
one person or a atched-pair 'e.g. related.
%he data are ordinal or interval not noinal.
-
8/9/2019 Psychology Revision - Research Methods Topic (1)
17/17
?"alitative Data
8P+A)A%,) ) DE*#(#$#O(S
?"antitative Data , %his is data which represents how uch" how long or how any there are o!
soething. ,t is presented in a nuerical !or.
?"alitative Data , %his is any data which is not presented in a nuerical !or. ,t is norallygathered !ro un-structured interviews/&uestionnaires" unstructured observations and case
studies.
8P+A)A%,) %3 ?+A4#$A$#9E DA$A A(A4=S#S
%here are any ways o! analysing &ualitative data" including1 discourse analysis 'studying written
or oral discourses" ethnography 'observing people in their natural environents and ,PA
'understanding how people a$e sense o! their own experiences.
'odin& , %his is a way o! identi!ying categories" thees" phrases or $eywords that ay be !oundin any set o! data. %his is done to help re!ine the researcher4s !indings" and help the be ore
objective in their approach.
%he categories" thees" phrases and $eywords are selected by one o! two approaches1
• $!ematic Analysis $op-Down Approac!2 , #odes represent ideas and concepts !ro an
existing theory/explanation 'e.g. clinical characteristics.
• #9#$=
%he traditional approach in psychology clais that there is one real world" and &uantitative
research see$s to discover that reality 6 validity is a easure o! the extent to which that has beenachieved.
%he &ualitative approach denies the existence o! any one world" claiing each individuals
perspective is reality. Hualitative researchers ac$nowledge the need !or re!lexivity 'the recognition
that a researcher4s attitudes have an in!luence on the research they conduct. :alidity o! this
research ay be deonstrated using triangulation 'coparing the results !ro a variety o! studies
o! the sae thing or person.
top related