Data

 collec)on  methods:  
Ques)onnaire    

Ali  BELAIDI  
 Objec)ves    
• Be  able  to  design  ques5onnaire  to  tap  different  
variables;  
• Be  able  to  evaluate  ques5onnaires,  dis5nguishing  the  
“good”  and  “bad”  ques5ons  therein.  
• Understand  the  issues  related  to  cross  cultural  
research.  
• Be  able  to  make  logical  decisions  as  to  appropriate  
data  collec5on  methods  in  survey  research.  
• Know  the  advantage  and  disadvantage  of  each  
method.  
• Be  able  to  make  logical  decisions  as  to  appropriate  
data  collec5on  methods  for  specific  studies.      
 Types  of  ques5onnaire  
• A  ques5onnaire  is  preformulated  wriJen  set  of  
ques5ons  to  which  respondents  record  their  
answers,  usually  within  rather  closely  defined  
alterna5ve.  
• They  are  efficient  data  collec5on  mechanism  
when  a  study  is  descrip5ve  or  explanatory  in  
nature.  
• Ques5onnaire  are  generally  designed  to  collect  
large  numbers  of  quan5ta5ve  data.  They  can  be  
administered  personally,  mailed  to  the  
respondents,  or  electronically  distributed.  
 Personally  administered  
ques)onnaires  
• When  the  survey  is  confined  to  a  local  area  a  
good  way  to  collect  data  is  to  personally  
administered  the  ques5onnaire.  
• The  researcher  can  collect  all  the  completed  
response  within  short  5me.  
• Any  doubts  that  the  respondents  might  have  on  
any  ques5on  can  be  clarified  on  the  spot.  
• The  researcher  is  offered  the  opportunity  to  
introduce  the  topic  and  mo5vate  the  
respondents  to  offer  their  frank  answers.    
 Disadvantage  of  personally  
administered  ques5onnaire    
• The  researcher  may  introduce  a  bias  by  
explaining  ques5ons  differently  to  different  
people.    
• Par5cipants  may  be  answering  different  
ques5ons  as  compared  to  those  to  whom  the  
ques5onnaire  was  mailed.  
• It  takes  5me  and  a  lot  effort.  
 Mail  and  electronic  ques)onnaire  
• Wide  geographical  can  be  covered  in  the  
survey.  
• They  are  sent  to  the  respondents,  who  can  
complete  them  at  their  convenience.  
• The  return  rates  of  such  ques5onnaires  
typically  low.  
• A  30%  response  is  considered  acceptable.  
• Any  doubts  the  respondents  might  have  
cannot  be  clarified.    
• With  very  low  return  rates  it  is  difficult  to  
establish  the  representa5veness  of  the  
sample.  
• Some  effec5ve  techniques  can  be  employed  
for  improving  the  rates  of  response  to  mail  
and  electronic  ques5onnaire.  
• Electronic  ques5onnaire  surveys  are        
 Guidelines  for  ques5onnaire  design  
QD  
• QD  should  focus  on  three  areas.  
– 1st  relates  to  the  wording  of  the  ques5ons.  
– 2nd  refers  to  the  planning  of  issues  with  regard  to  
how  variables  will  be  categorized,  scaled,  and  
coded  aXer  receipt  of  the  response.  
– 3rd  pertains  to  the  general  appearance  of  the  
ques5onnaire.  
• All  three  are  important  issues  in  QD  because  
they  can  minimize  bias  in  research.            
 Principles  of  wording  
The  principles  of  wording  refer  to  such  factors  
as:  
• The  appropriateness  of  the  content  of  the  
ques5ons.  
• How  ques5ons  are  worded  and  the  level  of  
sophis5cated    
• The  type  and  form  of  ques5on  asked;  
• The  sequencing  of  ques5ons;  
• The  personal  data  sought  from  the  
respondents.  
 Content  and  purpose  of  the  ques)on  
• The  nature  of  the  variable  tapped  –subjec5ve  
feeling  or  objec5ves  facts-­‐  will  determine  
what  kinds  of  ques5ons  are  asked.  If  the  
variables  tapped  are  the  subjec5ve  nature  
(e.g.  Sa5sfac5on,  involvement),  where  
respo0ndents’  beliefs,  percep5ons,  and  
aatudes  are  to  be  measured,  the  ques5ons  
should  tap  the  dimensions  and  elements  of  
the  concepts.    
• Where  objec5ve  variables,  such  as  age  and  
educa5on  level  of  respondents,  are  tapped,  a  
single  direct  ques5on  –  preferably  one  that  
has  an  ordinal  scaled  set  of  categories  –  is  
appropriate.  
• Thus,  the  purpose  of  each  ques5on  should  be  
carefully  so  that  the  variables  are  adequately  
measured  and  yet  no  superfluous  ques5on  are  
asked.    
Language  and  wording  of  the  ques5on  
• The  language  of  the  ques5onnaire  should  
approximate  the  level  of  understanding  of  the  
respondent.  The  choice  of  words  will  depend  
on  their  educa5onal  level,  the  usage  of  terms  
and  idiom  in  the  culture,  and  the  frames  of  
reference  of  the  respondents.    
• It  is  important  to  word  the  ques5on  in  way  
that  can  be  understood  by  the  repondent.  
• If  some  ques5ons  are  either  not  understood  
or  are  interpreted  differently  by  the  
respondent,  the  researcher  will  obtain  the  
wrong  answers  to  the  ques5ons,  and  
responses  will  thus  be  biased.  
•  hence  the  ques5ons  asked,  the  language  
used,  and  wording  should  be  appropriate  to  
tap  respondents’  aatudes,  percep5ons,  and  
feelings.      
 Types  and  form  of  ques5on  
• The  type  of  ques5on  refers  to  whether  the  
ques5on  is  open-­‐ended  or  closed.  The  form  of  
the  ques5on  refers  to  whether  it  is  posi5vely  
or  nega5vely  worded.  
•    
 Open-­‐ended  versus  closed  ques)ons  
• Open-­‐ended  ques5ons  allow  respondents  to  
answer  them  in  any  way  they  chose.  
• A  Closed  ques5on,  in  contrast,  asks  the  
respondents  to  make  choices  among  a  set  of  
alterna5ves  given  by  the  researcher.  
• Closed  ques5ons  helps  the  respondents  to  make  
quick  decision  to  choose  among  the  several  
alterna5ve  before  them.  
• They  also  help  the  researcher  to  code  the  
informa5on  easily  for  subsequent  analysis.        
Posi)vely  and  nega)vely  worded  ques)ons  

• Instead  of  phrasing  all  ques5ons  posi5vely,  it  

is  advisable  to  include  some  nega5vely  
worded  ques5on  as  well,  so  the  tendency  in  
respondents  to  mechanically  circle  the  point  
toward  one  end  of  the  scale  is  minimized.  
•    
 Double-­‐barreled  ques5ons  
• A  ques5on  that  lends  itself  to  different  
possible  responses  to  its  subparts  is  called  a  
double-­‐barreled  ques5on.  Such  ques5on  
should  be  avoided  and  two  or  more  separate  
ques5on  asked  instead.  
• Double-­‐barreled  ques5on  can  confuse  the  
respondents  and  obtain  ambiguous  
responses;  hence  double-­‐barreled  ques5ons  
should  be  eliminated.    
 Ambiguous  ques)ons  
• Even  ques5ons  that  are  not  double-­‐barreled  
might  be  ambiguously  worded  and  the  
respondent  may  not  be  sure  what  exactly  they  
mean.  Thus  responses  to  ambiguous  
ques5ons  have  built-­‐in  bias  inasmuch  as  
different  respondent  might  interpreted  such  
items  in  the  ques5onnaire  differently.  The  
result  is    mixed  bag  of  ambiguous  responses  
that  do  not  accurately  provide  the  correct  
answer  to  the  ques5on  
 Recall-­‐dependent  ques)ons  
• Some  ques5ons  that  are  not  double  require  
respondents  to  recall  experiences.    
 Leading  ques)ons  
• Ques5ons  should  not  be  phrased  in  such  a  
way  that  they  lead  the  respondents  to  give  
the  responses  that  the  researcher  would  like  
them  to  give.    
 Loaded  ques)ons  
• Another  type  of  bias  in  ques5ons  occurs  when  
they  are  phrased  in  am  emo5onally  charged  
manner.  
 Social  desirability  
• Ques5ons  should  not  be  worded  such  that  
they  elicit  socially  desirable  response.  
• Some5mes  certain  items  that  tap  social  
desirability  are  deliberately  introduced  at  
various  points  in  the  ques5onnaire  and  an  
index  of  each  individual’s    social  desirability  
tendency  is  calculated  therefrom.    
 Length  of  ques)ons  
• Finally,  simple,  short  ques5ons  are  preferable  
to  long  ones.  As  a  rule  of  thumb,  a  ques5on  or  
a  statement  in  the  ques5onnaire  should  not  
exceed  20  words,  or  exceed  one  full  line  in  
print.  
 Sequencing  of  ques)ons  
• The  sequence  of  ques5ons  in  the  
ques5onnaire  should  be  such  that  the  
respondents  is  led  from  ques5ons  of  general  
nature  to  those  that  are  more  specific,  and  
from  ques5ons  that  are  rela5vely  easy  to  
answer  to  those  are  progressively  more  
difficult.  This  funnel  approach,  as  it  is  called,  
facilitated  the  easy  and  smooth  progress  of  
the  respondent  through  the  items  of  
ques5onnaire.  
 Classifica)on  data  or  personal  
informa)on  
• Classifica5on  data,  also  known  as  personal  
informa5on  or  demographic  ques5ons,  elicit  such  
informa5on  as  age,  educa5onal  level,  marital  
status,  and  income.  Unless  absolutely  necessary,  
it  is  best  to  ask  for  the  name  of  the  respondent.  
If,  however,  the  ques5onnaire  has  to  be  
iden5fied  with  the  respondent  for  any  reason,  
then  the  ques5onnaire  can  be  numbered.  
• The  reason  for  using  numerical  system  in  
ques5onnaire  is  to  ensure  the  anonymity  of  the  
respondent.  
 Principles  of  measurement    
• There  are  some  principles  of  measurement  to  
be  followed  to  ensure  that  the  data  collected  
are  appropriate  to  test  our  hypotheses.  These  
refer  to  the  scale  and  scaling  techniques  used  
in  measuring  concepts,  as  well  as  the  
assessment  of  reliability  and  validity  of  the  
measures  used.  
General  appearance  or  “getup”  of  the  
ques)onnaire    
• Not  only  it  is  important  to  address  issues  of  
wording  and  measurement  in  ques5onnaire  
design,  but  it  is  also  necessary  to  pay  aJen5on  
to  how  the  ques5onnaire  looks.    
• An  aJrac5ve  and  neat  ques5onnaire  with  
appropraite  introduc5on,  instruc5ons  and  
well-­‐arrayed  set  of  ques5ons  and  response  
alterna5ves  will  make  it  easier  for  the  
respondents  to  answer  them.    
 A  good  introduc)on  
• A  proper  introduc5on  that  clearly  discloses  
the  iden5ty  of  the  researcher  and  conveys  the  
purpose  of  the  survey  is  absolutely  necessary.  
It  is  also  essen5al  to  establish  some  rapport  
with  the  respondents  and  mo5vate  them  to  
respond  to  ques5ons  in  the  ques5onnaires  
wholeheartedly  and  entusias5cally.        
 Organizing  ques)ons,  giving  instruc)ons  and  
guidance,  and  good  alignment  
• Organizing  the  ques5ons  logically  and  neatly  
in  appropriate  sec5ons  and  providing  
instruc5ons  on  how  to  complete  the  items  in  
each  sec5on  will  help  the  respondents  to  
answer  them  without  difficulty.  Ques5ons  
should  be  neatly  aligned  in  a  way  that  allows  
the  respondent  to  complete  the  task  of  
reading  and  answering  the  ques5onnaire  by  
expending  the  least  5me  and  effort  and  
without  straining  the  eyes.