Question Design (1)

download Question Design (1)

of 58

Transcript of Question Design (1)

  • 8/16/2019 Question Design (1)

    1/58

    WHAT IS A QUESTIONNAIRE?

     An instrument (form) tocollect answers to questions

    collect factual data

    gathers information or measures A series of written questions/items in a fixed,

    rational order

  • 8/16/2019 Question Design (1)

    2/58

    WHY USING A QUESTIONNAIRE?

     A well designed questionnaire:

    Givesaccurate andrelevant information to your

    research question

    Minimizes potential sources of biasWill more likely be completed

    ⇒ As simple and focused as possible

  • 8/16/2019 Question Design (1)

    3/58

    Statistical Terms

    • Population: Discrete group whose members can be defined by a set

    of at least one shared characteristic

    • Sample: A selection of that target population (can also be a

    subsample)

    • Random Sample: Sample in which every member of a given

    population has an equal chance of being selected

    • Margin of Sampling Error (MOSE): Every random sample has a

    margin of error (statistical price you pay for not interviewingeveryone)

    • Response Rate: ate at which people agreed to participate in

    survey! "f those who refuse are different from those who agree#

    results may be s$ewed (higher response rate % less chance for s$ew)

  • 8/16/2019 Question Design (1)

    4/58

    Statistical Terms

    • Population: Discrete group whose members can be defined by a set

    of at least one shared characteristic

    • Sample: A selection of that target population (can also be a

    subsample)

    • Random Sample: Sample in which every member of a given

    population has an equal chance of being selected

    • Margin of Sampling Error (MOSE): Every random sample has a

    margin of error (statistical price you pay for not interviewingeveryone)

    • Response Rate: ate at which people agreed to participate in

    survey! "f those who refuse are different from those who agree#

    results may be s$ewed (higher response rate % less chance for s$ew)

  • 8/16/2019 Question Design (1)

    5/58

    ADVANTAGES OF QUESTIONNAIRES

    Can reach a large number of people

    relatively easily and economically

    Provide quantifiable answers

    Relatively easy to analyse

  • 8/16/2019 Question Design (1)

    6/58

    DISADVANTAGES OF QUESTIONNAIRES

    Provides only limited insight into

    problemLimited response allowed by questions

    Maybe not the right questions are asked

     Varying response

    Misunderstanding/misinterpretationNeed to get it right first timeHard to chase after missing data

  • 8/16/2019 Question Design (1)

    7/58

    TYPES OF QUESTIONNAIRE

    Face to face

    Telephone

    By mail

    E-mail/InternetSelf-administrated

    Interviewer

    -administrated

  • 8/16/2019 Question Design (1)

    8/58

    SELF-ADMINISTERED QUESTIONNAIRE

     Advantages:Cheap and easy to administer

    Preserves confidentiality

    Completed at respondent's convenienceNo influence by interviewer

  • 8/16/2019 Question Design (1)

    9/58

    SELF-ADMINISTERED QUESTIONNAIRE

    Disadvantages:Low response rate

    Questions can be misunderstood

    No control by interviewerTime and resouces loss

  • 8/16/2019 Question Design (1)

    10/58

    INTERVIEW-ADMINISTERED

    QUESTIONNAIRE

     Advantages:Participation by illiterate people

    Clarification of ambiguity

    Quick answers

  • 8/16/2019 Question Design (1)

    11/58

    INTERVIEW-ADMINISTERED

    QUESTIONNAIRE

    Disadvantages:Interviewer bias

    Needs more resources

    Only short questionnaires possibleEspecially on telephone

    Difficult for sensitive issues

  • 8/16/2019 Question Design (1)

    12/58

  • 8/16/2019 Question Design (1)

    13/58

    Basic Survey Process

    Steps:

    1 !eciding on a mec"anism

    # !ra$ing a sample

    % !esigning t"e &uestionnaire

    &! 'ielding a survey and collecting the data

    ! Analysis

    ! eporting

  • 8/16/2019 Question Design (1)

    14/58

    STAGES IN DESIGNING A QUESTIONNAIRE (1)

    Planning the study:Decide on goalsIdentify risk factors for getting bitten by dogs

    Know the subjectLiterature, experts on dog bites

    Formulate a hypothesisPostmen more likely to get bitten by dogs than thenormal population

    Define information needed to test hypothesisOccupation, owning dog, outdoor activities, attitudetowards dogs

  • 8/16/2019 Question Design (1)

    15/58

    STAGES IN DESIGNING A QUESTIONNAIRE

    (2)

    Determine study population:

    Know the respondentsOccupation

    Special sensitivitiesEducation

    Ethnic

    Language

    ⇒Questionnaire needs to be adapted to yourpopulation, not the opposite!

  • 8/16/2019 Question Design (1)

    16/58

    STAGES IN DESIGNING A

    QUESTIONNAIRE (3)

    Design questions:

    Content of the questions

    Format of the questions

    Presentation and layout

    Coding schedule (if appropriate)

    Pilot and refine questionnaire

  • 8/16/2019 Question Design (1)

    17/58

    WHAT DO YOU PREFER?

  • 8/16/2019 Question Design (1)

    18/58

    WHAT DO YOU PREFER?

  • 8/16/2019 Question Design (1)

    19/58

    WHAT MAKES A WELL DESIGNED

    QUESTIONNAIRE?

    Good appearance

    → easy on the eye

    Short and simple

    Relevant and logical

    ⇒ High response rate

    ⇒ Easy data summarisation and analysis

  • 8/16/2019 Question Design (1)

    20/58

    QUESTION ORDER

    Decide onorder of items/questions Easy → difficult

    General→ particular

    Factual → abstract

    Where to placesensitive questions?

    Be aware of ordering effects!

  • 8/16/2019 Question Design (1)

    21/58

    QUESTION ORDER

    Questions should be ordered so as to seem

    logical to the respondent

    First questions should be relevant and easy

    Questions are effectively ordered from mostsalient to least salient

    Demographic questions should not be

    covered at the beginning

    Potentially objectionable questions areplaced near the end

  • 8/16/2019 Question Design (1)

    22/58

    QUESTION ORDER

    Group questions by topic/ response options

    Starting questionsSimple

    With closed format

    Relevant to main subject

    Non-offending

    Neither demographic nor personal questionsDon’t put most important item last

  • 8/16/2019 Question Design (1)

    23/58

    QUESTIONNAIRE INTRODUCTION

    Survey/interview introductionWho you are/ what is your ILP/thesis topic

    Why you are investigating

    Where you obtained the respondent’s nameHow and where you can be contacted

    Guarantee of confidentiality

    Length of interview (be honest)

    Usefulness of study should be clearto all respondents

  • 8/16/2019 Question Design (1)

    24/58

    CONTENT OF QUESTIONS

    Clearfocus on research question 

     Avoid sidetracking

     Avoid unnecessary information

    Demographic informationContact information (if non-anonymised)

  • 8/16/2019 Question Design (1)

    25/58

    DO:

    Use simple wording

    Be brief

    Be specific

  • 8/16/2019 Question Design (1)

    26/58

    DO NOT:

    Be vague

    Be condescending or talk down to

    respondent

    Use biased wordingUse abbreviations or scientific jargon

    Use objectionable questions

    Be redundant

  • 8/16/2019 Question Design (1)

    27/58

    FORMAT OF QUESTIONS

     Adjust toresponding audience

    Professionals vs. public

    Keep sentences simple and short

    Define key wordsRemember option “don’t know”

  • 8/16/2019 Question Design (1)

    28/58

    FORMAT OF QUESTIONS

     Ask for one information at a time

    Do you own a dog or have frequent contacts with dogs?

     Yes  

    No

    Use mutually exclusive and exhaustive answeroptions

     Vertical order of answer options

  • 8/16/2019 Question Design (1)

    29/58

    BE ACCURATE

    Do you often touch dogs? 

     Yes  

    No

    vs.How often did you touch a dog during thepast 3 months?

    Once

    Twice Three times or more Not at all Don´t now

  • 8/16/2019 Question Design (1)

    30/58

    BE APPROPRIATE

     !re you a drun?

     Yes  

    No  

    vs.

    How often have you consumedalcoholic "everages during the past #months?

    Daily $%# times&wee

    Once a wee

    'ess than once a wee

    Don´t now

  • 8/16/2019 Question Design (1)

    31/58

    BE OBJECTIVE

    Did you drin the strange "rownish drin in(rague?

     YesNo

     Vs.

     )hich "everage did you consume?

     )ater

    *eer  )ine +arad, None of them Don´t now

  • 8/16/2019 Question Design (1)

    32/58

    BE SIMPLE

    Did you smoe not less than a meanamountof - cigarettes&$ days from ./// onwards?

     Yes  

    No

    vs.

    Did you smoe an average of $ pac ofcigarettes&wee for the last 0 years?

     Yes   No

    Don´t now   

  • 8/16/2019 Question Design (1)

    33/58

    BIAS

    Bias = systematic differences in the

    measurement of a response

  • 8/16/2019 Question Design (1)

    34/58

    INFORMATION BIAS

    Recall biasCases more likely to remember than controls

    Observer bias Different interviewer – different interpretations

    Different interpretation of similar questions

    Reduce by structured questionnaire

  • 8/16/2019 Question Design (1)

    35/58

    NON-RESPONSE BIAS

    Those who respond are different from those who

    do notTelephone interviews: more females, elderly

    ReduceEnsure high response rate

    Random choice of interview partners

    Correct during analysis (eg age, sex)

  • 8/16/2019 Question Design (1)

    36/58

     FORMAT OF QUESTIONS

    Two main question formats

    Closed format→ forced choice

     Yes 

     !lways

    No 1ometimesDon2t now Never

    Open format→ free text

     )hat is your most distressing symptom?(lease descri"e444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444

     

  • 8/16/2019 Question Design (1)

    37/58

    OPEN OR CLOSED?

    Closed

     Advantages:Simple and quick

    Reduces discrimination against less literateEasy to code, record, analyze

    Easy to compare

    Easy to report results

  • 8/16/2019 Question Design (1)

    38/58

    CLOSED QUESTIONS

    Disadvantages: Restricted number of possible answers

    Loss of information

    Possible compromise: Insert field „others“

  • 8/16/2019 Question Design (1)

    39/58

    OPEN QUESTIONS

     Advantages:Not directive

     Allowsexploration of issues 

    to generate hypothesisqualitative research, focus groups, trawling questionnaires

    Used even if no comprehensive range of alternative choices

    Good for exploring knowledge and attitudes

    Detailed and unexpected answers possible

  • 8/16/2019 Question Design (1)

    40/58

    OPEN QUESTIONS

    Disadvantages:Interviewer bias

    Time-consuming

    Coding problemsDifficult to analyze!

    Difficult to compare groups

  • 8/16/2019 Question Design (1)

    41/58

    CLOSED QUESTIONS

    1. Straightforward response

     )hat is your age in years? 444 years

    How long have you owned a dog? 444

     years )hat is your se5 6gender7?

    8ale

    9emale 

    Did you stay in Hotel : on $3&-&;0? YesNo

    Don2t now

  • 8/16/2019 Question Design (1)

    42/58

    CLOSED QUESTIONS

    2. Checklist

     )hich of the following outdooractivitiesdid you do last wee?

  • 8/16/2019 Question Design (1)

    43/58

    CLOSED QUESTIONS

    3. Rating scale

    Did you do use sunscreen during the followingoutdoor activities during the past si5 months?

     !lways 1ometimes 1eldomly Never  

  • 8/16/2019 Question Design (1)

    44/58

    CLOSED QUESTIONS

    4. Rating scaleNumerical

    How useful would you thin that informationonthe ris of "iting from stray dogs would "e?(please circle)  . $ 3 > 0 # -

      Not at all useful ery useful

     !nalogueHow much is your pain severe 6put the ticon the line7

    ;.;

  • 8/16/2019 Question Design (1)

    45/58

    CLOSED QUESTIONS

    5. Scales for measuring attitude (Lickert)

     

    1tray dogs carry a higher ris of ra"ies

    No@ A strongly disagreeNo@ A disagree quite a lot

    No@ A disagree Bust a little

    A2m not sure a"out this

     Yes@ A agree Bust a little Yes@ A agree quite a lot

     Yes@ A strongly agree

  • 8/16/2019 Question Design (1)

    46/58

    PROBLEMS AND PITFALLS

     Avoid questions that asktwo things at once -

    you won’t know which ‘bit’ people are answering:

    Have you ever had stomach ache anddiarrhoea?

     Ambiguity.....

     Do you go to the woods a lot?

  • 8/16/2019 Question Design (1)

    47/58

    PROBLEMS AND PITFALLS

     Avoid jargon/abbreviations/slang

    How often do you get up at night to (C?6pass urine7

    1hould ADCs "e treated in the community?

     Avoidnot mutually exclusiveoptions )hat age are you?

    .#%$; 

    $;%$0 

    $0%3; 

    30%>; 

  • 8/16/2019 Question Design (1)

    48/58

    PROBLEMS AND PITFALLS

     Avoidleading questions

    Do you thin that the food in the hotelmade you sic?

    Did the hotel sta seem unhygenic to you?

    Do you agree that the hospital sta wereclose to e5haustion?

     Avoid making questionnairetoo long

    Typographical / spelling errors

  • 8/16/2019 Question Design (1)

    49/58

    PILOTING AND EVALUATION

    Pilot with asimilar group of people to your

    intended subjects

    Highlights problems before startingEffects of alternative wording

    Overall impression on respondents and interviewers

    Final polishing after several amendments

  • 8/16/2019 Question Design (1)

    50/58

    PRESENTATION AND LAYOUT

    Clear consistentlayout  Adequate space to answer

    Large font size Appropriate page breaks

     Avoidexperimental layouts

    fancy logos

    printed on recycled paper/is an equal opportunity employer etc

    S O O

  • 8/16/2019 Question Design (1)

    51/58

    PRESENTATION AND LAYOUT

     Usingcolour or printing questionnaire on coloured

    paper may help

    Usefilter questions, if necessary

    Giveclear instructions about how to answer the

    questions

  • 8/16/2019 Question Design (1)

    52/58

    WAYS TO INCREASE PERCEIVED

    REWARD

    Show positive regard

    Say thank you Ask them for “advice”Give social validationGive a tangible reward

    Make the questionnaire interestingState an upcoming deadline

  • 8/16/2019 Question Design (1)

    53/58

    WAYS TO DECREASE PERCEIVED

    COST

     Avoid subordinating language

    Do not embarrass the respondentDo not inconvenience the respondentMake questions appear short and easy Avoid asking personal information (if it must

    be asked, use ‘soft’ wording and inform aboutconfidentiality)Keep subsequent requests similar

  • 8/16/2019 Question Design (1)

    54/58

    CREATE A NAVIGATIONAL PATH

    Instructions should be placed right where

    they are needed

    Matrices are confusing: it is best to order

    questions from top to bottom, and left to

    right

    Uselarger font to attract attentionUse color shading to attract attention or

    show groupingsUse spacing and similarities to show

    groupings

  • 8/16/2019 Question Design (1)

    55/58

    WAYS TO ESTABLISHTRUST

    Provide a token of appreciation

    Provide a sense of legitimate authorityMake completing the questionnaire seem

    important

    Remind respondent of previous relationship

    with sponsor (if applicable)

  • 8/16/2019 Question Design (1)

    56/58

    SUMMARY

     A well designed questionnaire:

    Will give appropriate data which allow toanswer 

    your research question

    Will minimise potential sources ofbias, thusincreasing the validity of the questionnaire

    Will much more likely becompleted

  • 8/16/2019 Question Design (1)

    57/58

    FINALLY, KEEP YOURQUESTIONNAIRESHORT 

    AND THE QUESTIONSSIMPLE,FOCUSED AND

    APPROPRIATE

  • 8/16/2019 Question Design (1)

    58/58