document1

54
7/18/2019 1 http://slidepdf.com/reader/full/1563db865550346aa9a934eda 1/54 PowerPoint ® Presentation by Jim Foley © 2013 Worth Publishers Chapte r 1 Thinking Critically with Psychologi cal Science

Upload: colleen-shannon-mcdonough

Post on 03-Mar-2016

215 views

Category:

Documents


0 download

DESCRIPTION

power point

TRANSCRIPT

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 1/54

PowerPoint®

Presentationby Jim Foley

© 2013 Worth Publishers

Chapter 1ThinkingCritically

withPsychological

Science

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 2/54

Critical thinking refersto a more careful style offorming and evaluatingknowledge than simplyusing intuition.

In addition to the scientifc

method, critical thinking willhelp us develop moreeective and accurate waysto gure out what makespeople do, think, and feel the

things they do.

 psychological science…does this mean “criticize”?

Why do I needto work on mythinking? Can’tyou just tell mefacts about psychology?

 The brain isdesigned forsurviving andreproducing, butit is not the besttool for seeing

reality! clearly.• To improve ourthinking, we willlearn to catchourselves in somecritical thinking

errors.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 3/54

en our natura t n ngstyle ails!

"indsight#ias!

"I knew it all

along.#

$%erconfdence error!"I am sure I

am correct.# The coincidenceerror, or

mistakenlypercei%ing

order inrandom e%ents$

"The dice mustbe %ed becauseyou rolled threesi%es in a row.#

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 4/54

"indsight“&ias”

 The mind builds itscurrent wisdom aroundwhat we have alreadybeen told. &e are"biased# in favor of oldinformation.

For e%ample, we maystay in a badrelationship because it

has lasted this far andthus was "meant to be.#

Why call it“bias”?

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 5/54

$%erconfdence 'rror 1!

erformance &e are much too certainin our 'udgments.

&e overestimate ourperformance, our rate ofwork, our skills, and ourdegree of self(control.

$%erconfdence 'rror (!

 !ccuracy 

&e overestimate theaccuracy of ourknowledge. )eople

are much morecertain than theyare accurate.

*vercondence is aproblem ineyewitness

testimony. *vercondence is

also a problem ontests. If you feelcondent that youknow a concept, try

e%plaining it tosomeone else.

 Test for this$ "how longdo you think it takesyou to+# e.g. "'ustnish this one thing I!m

doing on the computerbefore I get to work#-

/nd your unscramblingspeed

"#$"%!&"%'()

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 6/54

 and o%eruse our intuition)

*rom an e%olutionary

perspecti%e!"indsight#ias

might be anoshoot ofour useful

habit ofanaly0ing anevent andtrying togure out

why itoccurred.

$%erconfdence errormight help

us lead otherpeople1

certaintybuildscondence in

followersmore thanaccuracy

does.

Percei%ingorder helps

us make

predictions1we 'ust needto test these

and notoverdo it.

$ometimes ourintuition gi*es

the rightanswer+ which

makes us trustit e*en more,

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 7/54

+aking our ideas moreaccurate #y

#eing scientifc&hat did "/ma0ing2andi# do about theclaim of seeing auras3e developed a testableprediction, which would

support the theory if itsucceeded.Which it didnot,

 The aura(readers were

unable to locate theaura around 2andi!sbody without seeing2andi!s body itself, sotheir claim was not

supported.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 8/54

4ut to guide you, you!llneed a scientic

/TTIT567.

ay, owdo - goa#out#eing

scientifc)

Is there

math?-esttubes?

 8ou!ll need tobe systematic.

S i if i d

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 9/54

Scientifc .ttitude Part 1!Curiosity

"ypothesis! 

9uriosity, ifnot guided bycaution, canlead to thedeath of

felines and

perhapshumans.

/efnition!always asking new:uestions

“-hat beha*ior I’m noticing in that

guy. is that common to all people?'r is it more common when understress? 'r only common for males?”

S i if . i d P (

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 10/54

Scientifc .ttitude Part (!Skepticism

/efnition! not accepting a fact! as truewithout challenging it1 seeing

if facts! can withstandattempts to disprove them

$kepticism+ like curiosity+ generates/uestions0 “Is there another e1planation

for the beha*ior I am seeing? Is there a problem with how I measured it+ or how Iset up my e1periment? 2o I need tochange my theory to 3t the e*idence?”

S i if . i d P 0

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 11/54

Scientifc .ttitude Part 0!"umility

"umility reersto

seeking the truth

rather thantrying to beright; a scientist

needs to be

able to acceptbeing wrong.

“Whatmatters is

not myopinion oryours+ butthe truthnature

re*eals inresponse to

our/uestioning,”

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 12/54

Criticalthinking!

analy0inginformation todecide if it

makes sense,rather than

simply

accepting it.oal! gettingat the truth,

even if it meansputting aside

your own ideas.

;ook forhidden

assumptions anddecide if

you agree.

;ook for

hidden#ias,politics,

values, orpersonalconnectio

ns.

)ut aside

your ownassumptions andbiases,

and lookat the

evidence.

<ee ifthere wasa =aw inhow the

information was

collected.

9onsider if

there areotherpossible

e%planations for thefacts orresults.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 13/54

%etting to the truth0

The Scientifc +ethod

 The scientic method is theprocess of testing our ideasabout the world by$

If the data doesn!t t our ideas, then wemodify our ideas, and test again.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 14/54

Scientifc+ethod! Toolsand oals

 re%ealed #y the scientifc

method!

 The brain can recoverfrom massive earlychildhood braindamage.

<leepwalkers are notacting out dreams.

*ur brains do nothave accurate

memories lockedinside like video les.  There is no "hidden

and unused >?percent# of our brain.

)eople often change

The #asics!  Theory

3ypothesis *perational6enitions

2eplication

2esearchgoals3types!

6escription 9orrelation )rediction 9ausation

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 15/54

Theory! the #ig picture

7%ample of atheory$ "/ll/636 symptomsare a reaction toeating sugar.#

/ theory, inthe language ofscience, is a set

of principles,

built onobservationsand other

veriable facts,that explains

some phenomenonand predicts itsfuture behavior.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 16/54

"ypotheses! inormedpredictions

"Testable# means thatthe hypothesis isstated in a way thatwe could makeobservations to ndout if it is true.

/ hypothesis is a testable prediction

consistent withour theory . What would be a prediction fromthe “!ll !2&2 isabout sugar”theory?

*ne hypothesis$ "If a kid gets sugar, the kid willact more distracted, impulsive, and hyper.#

 To test the "/ll# part of the theory$ "/636symptoms will continue for some kids evenafter sugar is removed from the diet.#

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 17/54

 whypotheses!

theories can #ias our o#ser%ations

&e might select onlythe data, or theinterpretations of thedata, that supportwhat we already

believe. There aresafeguards againstthis$3ypothesesdesigned to

disconrm$perationaldefnitions

uide or makinguseulo#ser%ations!3ow can wemeasure "/636symptoms# in the

previous e%ample inobservable terms Impulsivity @ A

of timesBhourcalling outwithout raising

hand. 3yperactivity @A of timesBhourout of seat

Inattention @ Aminutes

continuously ontask before

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 18/54

The ne4t3fnal step inthe scientifc

method!replication

 8ou could introduce a small change in thestudy, e.g. trying the /636Bsugar test oncollege students instead of elementarystudents.

2eplicating research means

trying it again usingthe same operational

denitions of theconcepts andprocedures.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 19/54

2esearchProcess! thedepression

e4ample

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 20/54

5ow that we6%e co%eredthis

We can mo%e on to this

2 h l d

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 21/54

2esearch goal andstrategy!

 description<trategies for gatheringthis information$Case Study!observing andgathering informationto compile an in(depthstudy of one individual5aturalistic$#ser%ation!gathering data aboutbehavior1 watching butnot interveningSur%eys and-nter%iews! havingother people report ontheir own attitudes and

/escripti%e

researchis a

systematic,objective

observation of people. The goal is

to provide aclear,

accuratepicture ofpeople!s

behaviors,thoughts,

andattributes.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 22/54

Case Study

"1amining one

indi*idual in depth 4ene3t0  can be a

source of ideas abouthuman nature in general

"1ample0  cases of

brain damage havesuggested the functionof dierent parts ofthe brain e.g. )hineasCage-

2anger0 

overgenerali0ation fromone e%ample1 "he gotbetter after tapping hishead so tapping mustbe the key to healthD#

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 23/54

$#ser%ing“natural”#eha%ior means

 just watching(and takingnotes), and nottrying to changeanything.

 This method can

be used to studymore than oneindividual, andto nd truthsthat apply to abroaderpopulation.

5aturalistic $#ser%ation

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 24/54

The Sur%ey

/efnition!  /method of gatheringinformation about

many people!sthoughts orbehaviors throughself(report rather

than observation. Eeys to getting useful

information$

4e careful aboutthe wordin  of

Wordinge7ectsthe results you

get from asurvey can bechanged by yourword selection.

'4ample!

50 2o youha*emoti*ation tostudy hard forthis course?

50 2o you feela desire tostudy hard forthis course?

a

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 25/54

apsychologysciencemistake wasmade here)

"int 81! "arry Truman won9

"int 8(!TheChicago-ribuneinter%iewed peoplea#outwhomtheywould%ote or9

"int 80!in 1:;<9

"int8;!#yphon

e9

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 26/54

Why take a sample)

If you want to nd outsomething about men, youcan!t interview every singleman on earth.

• <ampling saves time. 8oucan nd the ratio of colors inthis 'ar by making sure theyare well mi%ed randomi0ed-and then taking a sample.

 population

sampl e2andom

sampling is atechniue for

making sure thatevery individual in

a population has aneual chance of

being in your

sample.

"2andom#

means thatyour selectionof participantsis driven onlyby chance, notby any

characteristic.

i#l

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 27/54

Correlation

eneral/efnition! an

observation thattwo traits orattributes arerelated to eachother  thus, theyare "co#(related-

Scientifcdefnition! a measure of howclosely twofactors *arytogether , or how

well you can predicta change in one

-n a case study! The fewer hours

the boy wasallowed to sleep,

the moreepisodes ofaggression he

displayed.

. possi#leresult o manydescripti%estudies!disco*ering acorrelation -n a

naturalistic

o#ser%ation!9hildren in aclassroom whowere dressed inheavier clotheswere more likely

to fall asleepthan thosewearing lighter

clothes.-n a sur%ey! The greater the

number ofFacebook friends,

the less time wasspent studying.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 28/54

*inding Correlations!Scatterplots

)lace a dot on thegraph for each person,corresponding to the

numbers for theirheight and shoe si0e. In this imaginary

e%ample, heightcorrelates with shoe

si0e1 as height goesup, shoe si0e goes up.

     "    e      i    g      h

     t

Shoesi=e

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 29/54

 Correlation! *ace#ook and

Studying

 These are two factorswhich correlate1 theyvary together .

 This is a negati%e

correlation@ as onenumber goes up, theother number goesdown.

C l ti C A i t

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 30/54

Correlation CoeAcient•  The correlation coecient is a num#er representing the

strength and direction of correlation.•

The strength of the relationship refers to how close the dotsare to a straight line, which means one variable changese%actly as the other one does1 this number varies from ?.?? toGB( H.??.

•  The direction of the correlation can be positi%e bothvariables increase together- or negati%e as one goes up, theother goes down-.

  G H.?? ( H.??

)erfectpositivecorrelati

on

)erfectnegativecorrelatio

n

orelationship,

nocorrelation

%uess the Correlation Coe6cients

When scatterplots re%eal

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 31/54

When scatterplots re%ealcorrelations!3eight relates to shoe si0e, but does it

also correlate to "temperamentalreactivity score# / table doesn!t showthis, but the scatterplot does.

If d

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 32/54

If we nd acorrelation, what

conclusions can wedraw from it

7et’s say we 3nd thefollowing result0

there is a positi*ecorrelation betweentwo *ariables+ice cream sales+ andrates of *iolent crime&ow do we e1plainthis?

Correlation is not

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 33/54

Correlation is notCausationB

")eople who=oss more

regularly haveless risk of heartdisease.#

")eople withbigger feet tendto be taller.#

If this data isfrom a survey,can we concludethat =ossingmight preventheart disease *rthat people withheart(healthyhabits also =ossregularly

6oes that meanhaving biggerfeet causes height

Thi ki iti ll # t

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 34/54

Thinking critically a#outthe te4t!

If a low self(esteem test score"predicts# a high depressionscore, what have we conrmedthat low self(esteem causes orworsens depressionthat depression is bad for self(esteemthat low self(esteem may be part

of the denition of depression,and that we!re not reallyconnecting two dierent variablesat all

se es eem corre a es

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 35/54

 se es eem corre a eswith depression,

there are still numerous possible

causal links0

So how do we fnd out

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 36/54

So how do we fnd outa#out causation) &y

e4perimentation9 '4ample!

removing sugarfrom the diet ofchildren with/636 to see if itmakes adierence

In thedepressionBself(esteem e%ample$tryinginterventions thatimprove self(esteem to see if

they cause a

'4perimenta

tion!manipulatingone factor in a

situation todetermine its

eect

Dust to clariy two

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 37/54

 Dust to clariy twosimilarsounding

terms…

8irst you sample+ thenyou sort 9assign:,

2andomassignment

ofparticipantsto control ore%perimentalgroups is how

you controlall variablese%cept the

one you!remanipulating.

2andomsampling ishow you get

a pool ofresearch

participantsthatrepresents

thepopulation

you!re tryingto learnabout.

l # 7

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 38/54

Place#o e7ect

Place#o e7ect!experimental

e!ects that arecaused byexpectations

about theintervention

3ow do we make sure that thee%perimental group doesn!te%perience an eect becausethey e%pect to e%perience it

7%ample$ /n e%perimental groupgets a new drug while the controlgroup gets nothing, yet bothgroups improve.

%uess why,

Working with the

place#o e7ect!9ontrol groups maybe given a place#o an inactive substanceor other faketreatment in place of

the experimentaltreatment .

 The control group isideally "blind# towhether they aregetting real or fake

treatment.Kany studies aredou#le#lind neither participantsnor research staknows whichparticipants are in the

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 39/54

The Control roup

If we manipulate a variable in an e4perimental groupof people, and then we see an eect, how do we knowthe change wouldn!t have happened anyway

• &e solve this problem by comparing this group to acontrol group, a group that is the same in every wayexcept the one variable we are changing.

"1ample0 two groups of children ha*e !2&2+but only one group stops eating re3ned sugar,

4y usingrandom

assignment$randomly

selecting some

study participants tobe assigned to

the controlgroup or theexperimental

group.

3ow do makesure thecontrol groupis reallyidentical inevery way tothee%perimentalgroup

5aming the %aria#les

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 40/54

 The variable we are able to manipulateindependently of what the other variables

are doing is called the independent%aria#le E-FG.

L If we test the /636Bsugar hypothesis$L <ugar @ 9ause @ Independent MariableL /636 @ 7ect @ 6ependent Mariable

 The variable we e%pect to e%perience achange which depends on themanipulation we!re doing is called thedependent %aria#le E/FG.

L6id ice cream sales cause a rise in violence, orvice versa There might be a confounding

variable$ temperature.

"he other variables that might have ane!ect on the dependent variable areconounding %aria#les.

5aming the %aria#les

*illing in our defnition o

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 41/54

/n e4periment is a typeof research in which the

researcher carefullymanipulates a limited

number of factors (#$s)and measures the impact

on other factors (%$s).

Nin psychology,

you would belooking at theeect of thee%perimental

change IM- on a#eha%ior or

mental process 

6M-.

*illing in our defnition oe 1perimentation

Correlation %s causation!

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 42/54

Correlation %s9 causation!the #reasteeding3intelligence

Huestion• <tudies have found that children

who were breastfed score higheron intelligence tests, on average,than those who were bottle(fed.

• 9an we conclude that breastfeeding 9/5<7< higherintelligence

• ot necessarily. There is at leastone conounding %aria#le$genes. The intelligence testscores of the mothers might behigher in those who choosebreastfeeding.

• <o how do we deal with thisconfounding variable 3int$

e%periment.

2uling out conounding

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 43/54

2uling out conounding%aria#les!

e4periment with random

assignment !n actual study in the te1t0 women were randomlyselected to be in a group in which breastfeeding was

 promoted 

GOpoints

Critical Thinking

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 44/54

Critical Thinking

Watch out!descripti%e,naturalistic,

retrospecti%eresearchresults are

otenpresented as i

they show

causation9

.naly=ethisfctionalresult!

“Peoplewho attendpsychotherapy tend to#e more

depressedthan thea%erageperson9”

6oes this

Summary o the types o

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 45/54

Comparing 2esearch +ethods

2esearch+ethod &asicPurpose "owConducted

What is+anipulated

Weaknesses

Summary o the types o2esearch

6escriptive

 To observe andrecord behavior

)erformcasestudies,surveys, ornaturalistic

observations

othing o control ofvariables1single casesmay bemisleading

9orrelational  To detectnaturallyoccurringrelationships1to assess howwell onevariablepredicts

another

9omputestatisticalassociation,sometimesamongsurveyresponses

othing 6oes notspecify cause(eect1 onevariablepredictsanother butthis does notmean one

causes theother7%perimental  To e%plorecause(eect Kanipulateone ormorefactors1randomlyassignsome tocontrol

group

 Theindependentvariables-

<ometimes notpossible forpractical orethicalreasons1results maynot generali0eto other

conte%ts

*rom data to insight!

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 46/54

*rom data to insight!statistics

&e!ve done ourresearch and gathereddata.)ow what?

&e can use statistics,which are tools for

organi&ing, presenting,analy&ing, andinterpreting data.

The 5eed or Statistical

2easoning/ rst glance at ourobservations might give amisleading picture.

'4ample! Kany peoplehave a misleading pictureof what income distributionin /merica is ideal, actual,or even possible.

Malue of statistics$

H.to present a moreaccurate picture of our datae.g. the scatterplot- thanwe would see otherwise.

P.to help us reach validconclusions from our data1

statistics are a crucialcritical thinking tool.

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 47/54

Tools or /escri#ing /ata-he bar graph is one simple

display method but e*en thistool can be manipulated,

'urbrandoftruckisbetter;

'urbrandoftruck isnot sodi<erent.

Why is there a di<erence in the apparent result?

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 48/54

+ode

Lthe mostcommonlevelBnumberB

score

+ean

arithmetic"average#-

Lthe sum ofthe scores,divided by the

number ofscores

+edianmiddle

person!s score,or Q?th percentile-

LthenumberBlevelthat half of

people scoredabove andhalf of thembelow

+easures o centraltendency !re you looking for just ')"

)(=4"# to describe a population’s income+ height+ orage?

  'ptions0

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 49/54

+easures o centraltendency3ere is the mode, median, and mean

of a family income distribution. otethat this is a skewed distribution1 afew families greatly raise the mean

score.

&hy does this seesawbalance otice these

gaps

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 50/54

. di7erent %iew, showingwhy the seesaw

#alances!

 The income is so high for some familieson the right that 'ust a few familiescan balance the income of all thefamilies to the left of the mean.

+easures o %ariation!

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 51/54

Kean

Small standardde%iation

Iarge standardde%iation

2ange$ the dierence between thehighest and lowest scores in adistribution

Standard de%iation$ a calculation ofthe average distance of scores fromthe mean

+easures o %ariation!how spread out are the scores?

k d l

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 52/54

Skewed %s9 5ormal/istri#ution

Income distribution is skewed by the very rich.

Intelligence test distribution tends to form asymmetric "bell# shape that is so typical that itis called the normal cur%e9

Skewed distri#ution

5ormal cur%e

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 53/54

.pplying the concepts Try, with the help of this rough drawingbelow, to describe intelligence test scores ata high school and at a college using theconcepts of range and standard deviation.

Intelligence testscores at a highschool

Intelligence testscores at acollege 1J

J

/rawing conclusions rom

7/18/2019 1

http://slidepdf.com/reader/full/1563db865550346aa9a934eda 54/54

/rawing conclusions romdata!are the results useful?/fter nding a

pattern in our datathat shows adierence betweenone group andanother, we can askmore :uestions.

Is the dierencerelia#le! can weuse this result togenerali&e or topredict the futurebehavior of thebroader population

Is the dierencesignifcant! couldthe result havebeen caused byrandom' chancevariation between

3ow to achieve reliability$5on#iased sampling$ Kake surethe sample that you studied is agood representation of thepopulation you are trying to learnabout.

Consistency$ 9heck that thedata responses, observations- isnot too widely %aried to show aclear pattern.+any data points$ 6on!t try togenerali0e from 'ust a few cases,

instances, or responses.&hen have you found statisticallysignifcant dierence e.g.between e%perimental and controlgroups-&hen your data is reliable /6&hen the dierence between the