Module 10B - Design of research instruments YM.ppt

8/16/2019 Module 10B - Design of research instruments YM.ppt

1/27

Design of research

instruments

Module 10B


2/27

Researchinstruments?

• Questionnaire

• Interview guides

• Check-lists• etc


3/27

designingResearch

instruments?


4/27

Considerations !efore designing"uestionnaire?

• #hat do we want to know?

• #hat is the right data collectiontechni"ue?

• #ho will we ask "uestions?• Do we understand the to$ic su%cient to

design a "uestionnaire?

• &re our informants mainl' literate or

illiterate?• (ow large is the stud' sam$le


5/27

)$en-ended vsclosed "uestions?


6/27

*+am$les for o$en-ended

• #hat is 'our o$inion a!out the ,Bservices $rovided !' the hos$ital?

•#hat do 'ou think are reasons fornot taking ,B medication regularl'?


7/27

Closed Questions

& num!er of t'$es

1 .traightforward res$onse – What is your sex (gender)? M /

– Do you have cough > 2 weeks?


8/27

Closed Questions

2 Checklist

– Which of the following symptoms did you have?• Cough 3 2 wks

• #eight loss

• 4ersistent /ever

• Blood' cough

• Chest $ain


9/27

Closed Questions

5 Ranking scales

Please rank these activities in order of theamount of pain they cause you (! would"e most pain and # would "e least)

.itting

Running

6'ing in !ed

#alking


10/27

Closed Questions

7 6ikert scale

$%& testing in ' services would "e useful

o8 I strongl' disagree

o8 I disagree "uite a lot

o8 I disagree 9ust a little

I:m not sure a!out this

es8 I agree 9ust a little

es8 I agree "uite a lot

es8 I strongl' agree


11/27

Closed Questions

;


12/27

)$en-ended vsclosed "uestions?


13/27

)$en-ended "uestions

4ermits free res$onses which should !erecorded in the res$ondents: own words

seful for o!taining information a!out• /acts with which the researcher is not

ver' familiar

• )$inions8 attitudes8 and suggestions ofinformants

• .ensitive issues


14/27

Issues to consider

• &llows 'ou to $ro!e more dee$l'into issues of interest !eingraised

• &nal'sis is time consuming andre"uires e+$erience

• ,horough training ofinterviewers is necessar'


15/27

Closed "uestions

Advantages• It saves time• &nal'sis is eas'

Risks• 4rovide less information com$ared to

o$en-ended "uestions

• B' reading the list of $ossi!le answers ina given se"uence the interviewer ma'in@uence the choice of the res$ondent


16/27

4artiall' categoriAed "uestions

)$en ended "uestion !ut some answersare $recategoriAed

*+am$le

#ho told 'ou to seek health care?1 .elf

2 Mother

5 /ather

7 Relative; eigh!or )ther

)ther


17/27

4artiall' categoriAed "uestions

&dvantages• &nswers can !e recorded "uickl'• &nal'sis is easier

Risks• Interesting information ma' get lost if

categoriAation is too earl'• Interviewers ma' force information in

categor'• Interviewers ma' sto$ after o!taining 1

answer• Interviewer ma' !e tem$ted to $resent

answer o$tions


18/27

(ow to design a"uestionnaire?


19/27

Designing a "uestionnaire

1 Concentrate on content

2 /ormulate "uestions

5 Decide on se"uencing of "uestions

7 *valuate the format of the"uestionnaire

; ,est the "uestionnaire on colleague

&d9ust "uestionnaire ,est "uestionnaire on stud' $o$ulation

E &d9ust "uestionnaire


20/27

Content

• se 'our o!9ectives and varia!les asstarting $oint

• Decide what "uestions will !e needed tomeasure 'our varia!les and reach 'ouo!9ectives

• &ll "uestions necessar' to fulFll the aims of

the stud' should !e included

• Questions not directl' relevant to aims mustnot !e included


21/27

/ormulating "uestions =1>

Simple G Questions should !e short and askonl' one $iece of information $er "uestion

Intelligible G se words that are known !'

the stud' $artici$ants

Unambiguous G Questions should not havemore than one meaning =eg are 'ou ondrugs?>

Unbiased - o leading "uestions =Do 'outhink that smoking is !ad for 'ou?>


22/27

/ormulating "uestions =2>

1 &sk the "uestion to a friend

2 Check

5 Do 'ou get the t'$e of res$onse

'ou want?


23/27

.e"uence of "uestions

• ,he se"uence order should !e logical H Informantfriendl'

• .u!9ect must feel eas'8 like natural conversation

• .tart with eas' "uestions =age8 $lace of residence>

• ,hen interesting =$refera!l' o$en> non controversial

"uestion

• &sk more sensitive "uestions later =se+ual !ehavior8

drug use>


24/27

/ormatting the "uestionnaire

• Introductor' $age e+$laining

– the $ur$ose of the stud'8

– re"uesting informed consent

– assuring conFdentialit'• .$ace for num!er8 date8 location of interview

and name interviewer

• 4ut together "uestions that !elong together8

su!headings ma' !e used• .u%cient s$ace to write answer to o$en

"uestions

• Bo+es for ticking $re-categoriAed answers


25/27

.elf administered "uestionnaire

• Re"uires clear guidelines on how the

answers should !e Flled in

• )ften low res$onse rate

• seful for

– studies on sensitive to$ics

– large scale surve's


26/27

Checklist

• sed to s'stematicall' observe !ehavioror the condition of e"ui$ment etc

• Contains all items that must !e

considered during the o!servation

Observations of Pit Latrines

yes little no

Evidence of use

Toilet smells

Flies in and around toilet

Defecation in the area around


27/27

,raining of interviewers

• Reading

• 6isteningo!serving

• Role $la'

• &ctual $ractice in the Feld

Module 10B - Design of research instruments YM.ppt

Documents