Questionnaire and Interview

8 Schedule

The purpose of theresearch determines whether thesurvey procedure

hould be structured or unstructured. Generally, the .structured ap
oroach is chosen when hypotheses are to be tested while .unstructured
approach is used when an exploratory study is to be conduçted. The
structured procedure improves the quality of the data by minimising
the measurement error. In this procedure, the dat are collected either
through self-admninistered questionnaires or through face-to-fce inter
views or by the combinatËon of these two methods. We will,explore,
in this chapter, some fundamental issues pertaining to the natiure,
structure, content, design and construction of both. We will concen
rateprimarily on questioning rather than strictly on questionnaires or
nterview schedules. For instance, type of questions, content of ques
relevant
tions, phrasing ot questions and sequence of questions' are as
schedules.
fo questionnaires as to

WHAT IS A QUESTIONNAIRE?
sent by mail,
uestionnaire is the structured set of questions usually hand
also. The delivery
hough sometimes it is delivered by hahd organisation, and so on.
ould be at homne, school/college, 'office, contans a setotques
Duestionna1re 1S described as a document that
providéd personally by the
ons, the answers to which are to be'
Respondents". to the respondents
The importance of the survey is explaincd envelop
rough acovering letter. Usually, a self-addressed stamped
questionnaire to reduce, their
sent to therespondents along with the
returning the qustionna1re is
penses. The tollow-up request for,
ade through repcated letters. large samples are de
Questionnaire is used as a tool when: (i)very
 194 Questionnaire and Interview Schedule

Sired, (11) costs have to be kept low, (iii) the target groupsofwho are
ease admini.
likely to have high response rates are specialised, (iv)
Stration is necessarv, and (v) moderate response rate is considered
satisfactory. for data
In deciding whether questionnaire is an appropriate tool
(Black and
OuectIOn, tollowing four aspects must be born in mind
Ghampion, 1976:3 729): (1) Identify situations for which questions are
best suited. (2) Discuss advantages and disadvantages of a questIOn
naire as a research tool of data collection. (3) Delincate dimensions to
be associated with questionnaire construction. (4) Differentiate be
tween several types of questionnaires.
For analytical purposes, the following five types of questionnaiFes
may be identified:
(1) Topic: Whether the questionnire deals with one specific topic or
several topics.
(ii) Size: Whether the questionnaire is small (printed on post card) or
middle-sized (5-6 pages) or large-sized (9-10 pages), i. e., we may.
classify them as short and long questionnaires.
(iii) Target: Whether the questionnaire is addressed to specific group of
people or to general people.
iv) Type of response required: Whether, the questions are closed, open,
ènded or combination of the two types.
(v) Method of administration: Whether the questionnaire is mailed or
it is personally administered to subjects to complete in the pres
ence of the researcher or his assistant.

WHAT IS. A INTERVIEW SCHEDULE?

The set of structured questions in which answers, are. recorded by thes
1ntervIewer himself 1s called 1ntervView schedule; or simply the sched
ule. It is distinguished from the, questionnaire, in the sense that in the
latter (gçestionnaire) the answers are filled in by the respondent him
self. Though the questionnaire is used when the respondents are
educated, theschedule can be used both for the illiterate and the edu
cated respondens. The questionnaire is used whenthe respondents
are scattered in alarge geographical area but the schedule is used wheni
the tespondents are located in à small' ärea sÍ that they can be person
ally contacted. In the questionnaire, the size, physical appearance and
attractiveness are more imnportant than in the schedule in securing in
formatioa from the respondents. The wordings of the questions in thej
 Questionnaire and Interview Schedule 195

questionnaire have to be simple since the interviewer is not present

to

explain the meaning to the respondent. In the schedule, the investiga

terms.
tor gets the opportunity to explain difficult
The questions in the questionnaire/schedule seek three types of
information: ().demographic information which identifies the inter
viewee, (ii) substantive information which is focused on the subject
under study, and (i) additiona!informatign phich may supporhe
substantive information. However, the con_tiuçtion of the schédule
will, there
or the questionnaire involves the same considerations, We
fore, discuss theirformation together.
FORMAT OF THE QUESTIONNAIRE/SCHEDULE:
SOME PRACTIAL CÖNCERNS
Questionnaire/schedule ormat refers to the general model which pro
vides guidelines on how the questions should be placed in a sequenc
and in a logical order of relationship to each other, what type of ques
questionnaire/schedule
tions should be considered, how long the
should be, and how theuestionnaire/schedule should be presented so
f that itis clear and easy to understand.
The fo'mat of the questionnaire/schedule should be carefully con
cerned with the following aspects:

Length
depend upon: () what
How long the questionnaire/schedule be would items are necessary so
the researcher wants to know and how nany
type of study (since self-ad
that the data will be crediblé; (i) on the thanface- to-face interviews;
ministêred questionnaires may be shörter
for the study; (i)
(iii) on the time which the rescarcher. has ayailable on researcher's
on the time the respondnts can and will take; and (v)
eso1rces.
For obtainingnecessary and adequate data and credible answrs, it
questionnaire should be given im
is necessary that the length of the
long. t is eually important that
portance, i.., it should be reasonablyresponding
filling up questionnaires or to interview schedule
time for
generally limited to 3-40 minutes in comparisonto face-to-façe it
1S minutes: Another consideration
terview which can continue for 45-60, available? Willthey take in
is the respondents. How long can they: be
answering questions, seripusly?Xoung., people may:. be
terest in people.
available lor less Liluc than the middle-aged and theold
 196 Questionnaire and Interview Schedule

Clearly typed
Questions should not be hard to read. They should be typed/printed
clearly.
Adequate space for answers
recording answers so that the re,
P'roper space may be provided formargins
spondent has not'to write in and on the back. Som
number for an answer (e.glot
open-ended questions ma÷ require a single
blank (-) may be left for this r
age, income, caste, etc.). Only a
sponse category.
One question should be put in one
line, leaving space for the re
sponse. Here is one example:
A. 1. Yes
2 No
3 Don't know

B 1. Yes
2 No
3 Don't know
) Don't know
Yes No
C.

he

format but B and C are righa

In the above example, A is wrong
formats.

(u Avoid abbreviations
Questions should nòtbe abbreviated.
duc

Proper instructions
writing answers, the instructionsshould be explicit. For nomina
For respondent should be asked either to
questions (with oneanswer), the
( ) or a box may be provided for
tick or cross or circle, and a blank
it. For exarnple:
 Questionnaire and Interview Schedule 197
1. Scx: Male
Female
Male X
Female
1ted Male X] Female
Male 1 Female

The boxes ( ) are preferable to blanks ().

mé It is always desirable to place response
categoris below one an
-8other because it will reduce confusion.
Q. Describ your educational level:
Illiterate

DO00000D0
re

Can read and write

Primary (5th)

Secondary (8th)

Higher secondary (12th)

Graduate

Postgraduate

Professional degrce

Professional diploma

The problem in this format is that it will require more space and
papersand åppear longer whict
he questionnaire will include more or more responses can be
response rate. As such, two
ghtnay aftect the
on the same line.
buL horizontally side by side
; primary secondary
iterate can read and write

response format can also be prepared as under:

The
ducation:
Graduaie
Primary Higher secondary
Can rcad and write
Illiterate
inal

for
 198 Questiconnaire and lntervicw Schedule

reframed as: "How many years did you ing

Ihe question may also be vears will enable rcsearcher to
Put in schoolcollege?" The number of
determine the educational level.
The
for
Branching of questions
Branching of some questions is necessary, Suppose a.question 1s askea: any
Do. you participate, in sports, mu_ic, debates?" Now some respon none
may participatë in one actívity, some in two,and some in in thoQu
aents relevant to everyone
This means that aIl activities may not be Son
This requires a separate designing as illustrated in the follow whi
Survey.
ing examnple:
Yes/No
ties
1. Do you participate in sports?
If yes, which sports do you perform? No
Yes
1
Football Lor
2
Cricket
1 be :
2
Hockey 2
1
Other (specify)
Yes/No
2. Do you participate in music?
If yes, what type of music? The
classical/ligh/instrumental/any other (specify) mis
Yes/No stat
3 Doyou participate in debates?
abl
ans

Determining number and response categories

In ordinal questions, the.number of response categories are often sub
jective, and the researcher. fails. to .decide the number of categorie_
Pre
between the highest and the lowest. Generally, the number is three of tary
four or five, as shown below: COU
() Regularly/occasionally/hardly ever/never lar
(i1) Excellent/good/poor/undecided
(ii) Strongly agree/agree/neutral/strongly disagree/disagree Re
(iv) Very important/important/somewhat important/not important/
Th
don't know has
(v) Always/sometimes/rarely/never rese
The number of categories should be decided whether the findings
will be affected by more or less numbers.
 Questionnaire and Interview Schedule 199

Babbic (1998: 147-150) has explained following guidelines for fram

ing and aslking questions:

Ouestions should be clear and unambiguous

The question like, "What do you think about the proposed peace plan
for Kashmir?" may not be c.ear to respondent who does not know
anything 'about the peace plän! ,: i

Ouestions sbonld be relevaint

Sometimes the respondents are asked' to give opinions on issues on
which they have never given any thought, e.g., What is your opinion
on the economic policies of the BJP, the Congress, and the CPI par )
ies?" Such questions are bourdto be disregarded by the respondents.
)
Qhúestions sbould be sbort
Long and complicated items are to. be avoided. The respondent should
be able to read an item quickly, understand its meaning and think of
an answer without difficulty.

cgative questions sbould be avoided

The appearance of a negatiòn inthe question paves the wvay tor easy
misinterpretation. For. example, asking to agree or disagree with the
Staterment, "Indiashould not recognise the military rule in Fiji,
a size
word .'nor and
able portion of the respondents will not read the
answer on that basis.

hased terms. sbould be avoided

Prejudice affects the answers. For example, the question, "ave mili
hampered our
tary rulers in the neighbouring country always particu
country'sprogress?" may encouragesome respondents to give
questions do.
lar response more than other

Re[pondents niiust be competent to dnsWEr

Theresearcher should always ask himself whether the respondents he
toanswer questions on the issue of
lhas cliosen are competeni enough give their views
to
research. For example, asking daily wage labourers
 2U0 Questionnaire and nterview Schedule

on 'communal violence' may not be rational. Similarly, asking stu.

denis to indicate the manner in which university's total income ough Thou
to be spent will be wrong because students may not have faurly good sidera
involved in them.
iknowledge of the nature of activities and the cosLs
quest

Respondents must be willing to answer

Many a ume people are unwilling to share opinions with others, e.g.;
Muslims in India.
asling Muslims about Pakistan's attitude towards 1

Sarantakos (1998:226-27) has referred to five qåestionnaire fo S

matS:

) Funnel format: where questioning moves from general to

irom impersonal to personal and trom non-sensitive to
specific,
sensitive:
General: What are different methods of controlling the size 2
of family?
Specific: What family planning methods are generally used
by the lower-caste people?
Impersonal: Are Muslims generally in favour of or against fam 3.
ily planning measures?
Personal : As a Muslim, are you in tavour of using family
planning measures?
Non In your opinion, what types of contraceptives are
sensitive: mostly used by villagers?
Sensitive: What contraceptives do you use?
(1) Inverted funnel format: where the questioning progresses trom
specific to general, from personal to impersonal, and from sensi-;
tIVe to non-scnsitive.

(iii) Diamondformat: is a combination of the funnel and inverted fun

ncl formats, where questions progress trom specitic to general and
back to specific, from personal to impersonal and back to personal
and sO On.
(iv) Box format: where questions are uniform throughout the ques
tionnaire, with all questions being kept at the same level (i.e.,
using box for tick-narking answers in each question).
(v) Mixed format: which contains sections, each adopting one of the
above formats; for example, the first section may employ the fun
nel format, the second the box format, and the last the inverted
funnel format.
 201
Questionnaire aid Interview Schedule

ANRANGING SEQUENCE OF THE QUESTIONS

hough the ordering of the questions would depend on several con
iderat1ons but some points should be given importance in ordering
uestions (Fink and Kosecoff, 1989:43-45).
1 The first set of guestions shonld be related to the topic under study.
For exmple, suppose, the topic is "Deficiencies in the Existing
Educational System'". One of the first guestions in the earlicr set
of questions could be: Hlow far are you,satisfied .with the regular
ity of teachers in taking classes? (fully satisfied/somewhat
satisíied/dissatisfied/extremely dissatisfid). People respond best
when the first questions ask for objective facts. Once they. become
certain of study's purposes, they will usually provide answers to
relatively subjective questions.
2 Questions should precede from the most familiarto the least. In asur
ed
vey of deficiencies in the existing educational system, items can first
be asked about the respondent's own feelings, then of the feelings
of other students, teachers, parents, administrators and so on.
Avoid very general questions. Question like "when did you start
reading newspaper'"? is very general question. The proper ques
ly tion would be: "Did you have any interest in reading newspaper
when you were in class tenth?"
-tei
4. Place relatively easytazansuer questions frst When difficuit ques
tions are asked. in the beginning, the respondent, may soon teel
tired; he may not answer questions seriously. The respondent can
m take more time in answering difficult questions put in the end.
s1-5. The demographic questionsshould be putatat the
the eend. These questions
marital
pertain to age, income, .occupation, caste, education,
n status, residence, background, etc., and can be answered easily.
nd 6. Place the sensitive questions in the middle: Questions pertaining to
educational
al. attitudes towards political corruption, government's
of professional educa
policy, incentives for controlling quality must be placed, in the
es tion, reviewing reservation policy, etc., and is not
e., middle sc that respondent is willing to pay attntion
too fatigued io answer properly.
putting look-alike questions at one place. For example, 8 to 10
he . Avoid agree or disagree
n questions--all of which ask the respondent to respondent may
and, the
ed with statements---may lead to bordom format willmiuimise.che
a-iferent
give up. PuttingquesIions in questions should be grouped
loss of interest. Alternatively,
202 Questionnaire and Interview Schedule

such a way that they describe the same format, e.g., The next ber o
question will ask you if you agree or disagree with 10 different classi
statèments." This is called 'providing transitions'. rately
8.
Put questions in a logical order. Questions should be put in such a
logical order that it may not appear that the respondent has to
switch over suddenly from abstract to dirsstquestions or irom On t
one topic to.another, eg., asking questions on family, followed by classi
burning problems in the country, respondent's c£rcer aspiratións, inforr
communal riots in the state, fçnctioning of political'elite, and so Vides
on, is not the logical order of questions. Oppenheim (1966:38-39) deterr
Phillips (1971:14 1) and Bailey (1982:14 1) have'suggested applying domi
funnel technique in ordering questidns. By this; they mcan that famil=
general broad and open-ended qustions be asked first, followed whict
by more.specificquestians, Theeasy questions put the respondent: Ondar
at ease, A filter' question is one that determines whether future respo
questions are applicable 'to therespondent. For example, first ask vho
whether the respondent smokes änd then 'ask how many ciga tive"
rettes in aday.
mar1
Sometimes answer to one question wil affect answer cf another.
This seriously diminishes the value of the questionnaire. There ther p
frame
fore, proper ordering of questions is very important. For
example, here are two questions A and B: matio
A. are O0
Do you consider ýour class teacher an ideal teacher? The f
B. What are the qualities of an ideal teacher? after
Here question Bshould come before qüestion A. Here is 'another
tiOn g
example of ordering questions:
A. How satisfied are you with the economic policy of the
sent Prime Minister? pre-élos
B. How do you rate the performance of the present The c
i[ter? Prime Min the re
Question B should precede A' because a person who is' searc
with the Prime Minister'scconomic policy (and perhapsdissatisfied teach
else) might rate Prim Minister's nothing to St
9. leadrship lower than
Ouestions of recall should also be organised açcording to otherwise. denTs
sequence. their natural stude
tiviti
TYPES OF QUESTIONS also t
Ouestions in the
questionna1re/schedule vary with respect to a num quire
Who
 Questionnaire and Interview Schedule 203

ber of criteria Diagraan 1 on the next page describes four bases for
dassifying questions We will bricfly describe each one of them sepa:
rately.

primary, sccondary and tertiary questions

On the basis of the nature of information clicited, questions may be
classified as primary, secondary and tertiaty. Primary questions elicit
information dixcctly related to the research topic. Each questíon pro
Vides information about a specific aspect of the topic. or example, for
determining the type of femily (whether it is husband-dominant, wife
at
dominant, equalitarian), the questien "who takes decisions in your
d íamily" is a primary question Secondary questions elicit informaticn
1t
which do not relate directly to the topic, i.e., the intormation i5 of sec
re
ondary importance They only guard the truthfulness of the
k respondents, e.g., in the abovetopic (on type of family), the question
who decides the nature of gift to be given in marriage to family rela
tive" or "who finally selects the boywith whom the daughter is to be
married" are the secondary questions. The tertiary questions are of nei
ther primnary nor of secondary importance. These only establish a
e
framework that allows convenient data collectionand sufficient infor
Or
mation vithout exhausting or biasing the respondent.These questions
are of two sub-types: (1) padding questions, and (b) probe questions.
The former questions act as a breather and are usually placed before or
after thesensitive questions; the latter questions only expand informa
er tion given by the respondent.

e-élosed-cnded and open-ended questions

the closed-ended guestions are the fixed-choice questions They require
a-the respondent to choose a response from those provided by the re
searchcr. Here is one example: "Whom do you consider an ideal
dreacher?" (a) who takes teaching seriously; (b) who is always available
g tostudents for discussions and guidance; (c) whose approach to sru
e. dents' problems is flexible; (d) who does not believe in punishing
al students; (e) who takes interest in co-curricular.and extra-curricular ac
tivities; and () who believes in teaching not only through lectures but
alsothrough life-situations.
|The open-ended questions are free-response questions which re
n-guire respondents Lo answer in theit own words..For example: (1)
Whom do you consider an ideal teacher?(2) Ilow would you rate the
 Questionnaire and Inteview Schedule 205

performance of the last government? (3) What do ;ou feel is the most
important issue facing India today?
The following questions illustrate the
open-ended and the closed-ended forms: diference between the
(closed): After the introduction of profit-fharing scheme in
factory, would you say that the annual productionyour
has
increased or decreased orhas rermaned the såine?

(open):
Increased/Decreased/Same
How would you describe the poduction in your factory
this year compared to last years
(closed): Do you have harmonious/ yormal/ conflicting relations
with your spouse?
(open): How would you describé the relations with your
Spouse?
(open): How would you describe the government's scheme of
training and financial help to scavengers in weaning
them away from the traditional?
(closed): Do you think that government's scheme of training and
financial help to scavengers has been fully successful/par
tially successful/failur?
Since open-ended questions entail more work both for the re
searcher and the respondents, these äre used sparingly in the
questionnaires. Some scholars use a middle path in using closed and
interviews or
open questions. They use open questions in preliminary
spontaneously; this in
in pre-tests to determine what respondents sayquestions for the final
formation then is used to construct closed
many re
questionnaire. However, this method is not followed by
searchers.
The advantages of open-ended questions are:
the response categories, he
1. Since the researcher does not know. all
categories from the respondents.
finds out the appropriate answer understanding.
2. The researcher gets insight in respondent's
are very large (say, 50 or more),
When the total answer categories if
3.
would be awkward to list altof them on a questionnaire; but
it
were omitted, then there would not be appropriate answers
some
available for all respondents. the researcher
respondent gets freedom in answering, logic
4. Since the information based on the respondent's
gets more and varied Sometimes, the information and responses
and thought processes.
 206 Qu ionnaire and Interview Schedule

receivC are so unexpected that the rescarcher's ideas are com. 3.

pletely nanged.
5. They art preferable for complex issues that cannot be condensed
into a fe small categories. 4.
The disadiantages of open-ended type of questions are:
1. Somètims responses received are irrelevant.
2. It is difficult to classify and code all
responses.
3. Since the data are not standardised, statistical analysis and compu
5

tation of percerntages bcome difficult.

4. Sometimes,'thé responses given are very lengthy and analysiñg side
them becomes_time-consuming. clos
5. Semilirerate . respondents find it difficult to answer open ques. 1.
tions since they require better ability toexpress one's feelings.
6. Open-ended questions can require much of the respondent's time
and effort and may engender a high refusal rate
On the other hand, the advaFtages of closed-ended questions are:
1. They provide a greater unitormiy of responses.
2. Itis easy to code, score and procegsatándard answers which saves 2
time andmeey.
3. The respondent has not to usemuch brain as he is often clearer
about the meaning of question.
4. Little time is taken to complete questionnaire.
5. Answers can be compared Irom person to person. 3.
6. Irrelevant responsesare not received and the answers are rela
tively complete, e.g., an open-ended question how often do you:
smoke" may receive an answer "whenever:I feel like smoking
but a closed-ended question may receive an answer, "one packet a
day, or two packets in a day, or four cigarettes in a day", and so;
On.

7. Response rate is high, particularly in sensitive questions like in

Come, .age, etc.. l1 the. answer in closed-ended question: is
category, the respondent may easily identify himself with th 4
r¡nge in which his income/age falls.
The disadvantages of closed-ended questions are:
1..The respondent may not get all alternative
responses as some im
portant responses might have beern omitted by the researcher.
2. The respondent does not think
and does not
giving, free information, He tickscven wrong involve himself id 5
answer.
 Qnestionnaire and Interview Schedule 207

the respondents do not find those answers in tL

Dm. Many a time feelings or atti
questions which correspond to their true
3.
closed
sed tudes.
response_ guesses and
4. The respondent who does not know the answer ran
chooses oneof the conven1ent responses or giVeSAN
domlv.
Detecting the mistake whether the reapöndent hs tickedhe
5.
rightanswer is 'not possiBle.
sçggested that five COn
K.L Kahn and C.F. C£nnell (1957) have chöosing open-ended or
1n siderations shoiuld be viewed asimportant in
closed-ended questions:
1es 1 The objectives of the study. and the pur
I the objectËves are limitedatitudes
of some and
pose is only to classify respondents in terms :

efficient.
me behaviour, closed questions would be appropriate and
information is sought
But if the survey objectives are broader anddepth, of respondent's
on the basis of opiniona expressed and the
knowledge, open questions would be preferable.
2 The level of information of the respondents on the topic under study. )
ves
of the respondents wilI have
If it is believed that the vas: majority
sufficient information,' opeDqqestions will be appropriate; but if
the level of information possessed by the respondents is uncertain,
rer

closed-ended questions may be preferred.

3.
opinions, struc
Respondents' interest. How well, are respondents' interested in the
tured? If it is fclt that the respondents are equally.
ela problem under study and nust have thought about the problem
attitude,
beforc and they can take a position or expressa definite
the closed-ended questions may, be satisfactory. For example, the
closed
probleni of students' transport tß schools is such that
eta
SO ended questions may provide sufficient information. If the issue
involved 1s on reserved representation to some specific castes:or to
Women, open-ended questions would be better.
Motivations of respondents to communicate their experien ces and
a
4
the open-ended
thougbts. If the'respondents are highly motivated, motivated, closed
questions will be successful, but if they are less SOmetimes

m.
ended questions would be preferable. Closea quespeople prefer to

dampen respondents' motivation because some

Cxpress their opinion in their own words.
of respondents" charac
1be cxten! of the researcher's knowledge respondents vocabulary
teristics. lf the researcher undcrstands the
 208 Questionnaire and Interview Schedule
va
2
and knows something about the information and knowledge po; ab
sessed by them and also their level of motivation closed question: 3. re
would besufficient. qu
According to Lindzey Gardner (1968: 565-66), the following five 5 th
conditions are important in determining the choice between closed.
ended and open-ended questions: here:
1. The objectives of intervieiù: Open-ended question is more appropr 1 qu
ate when the research objective is to discover respondent pl
opinions and attitudes or to determine the level of their inform 2 res
tion or to find out the intensity of their feelings, or to assess thñ In
3.
basis on which they have formed the opinion. Closed question i; ac
useful when the objective is limited to the classification of the re
spondents with respect to some attitudes or opinions. 5. rat
2. Respondents' information level: The less-informed respondent may Sc.
DOrtan
prefer closed question as he can choose among alternative answers
while an educated one may like open-question. The open-ended 10t
opende:
question would be more appropriate if the topic is outside the ex che ot.
perience of many respondents. closed
3. Respondents' intelligence: If the respondents are expected to have
and P
clear-cut attitude on the question or if they are believcd to have
L1Ons,
given enough thought to the question, the closed form is more ap maunt:
propriate. But if the respondents' thoughts are less structured on SUes a
the topic in question, open-ended question would be appropriate.
The high level of respondent's intelligence will enable him to
think of various alternative responses and choose one amongsIrec
them. These
4. Motivation to communicate: The closed question demands less re spons
spondent effort, less revelation and is less threatening to him questu
When the interviewer knows in advance the likely level of re spond
sponses, a closed question may be desirable. questi

5. Insight in respondents' characteristics: Researcher's advance know! that p

edge of respondents' characteristics, depth of knowledge
expertness in area and inotivation tO Comnmunicate c¡n focus on
open questions, but if the respondent knows little, closed ques: D
tions would be better for him.
B: In
According to Bailey (1982:126-127) closed-ended questions slhoul! Im
be used where:
D
1. the answercategories are clar, distinct, discrete and relaively few
in number;
 Questionnaire and Inteuie Schedule 209

2 variables to be measured are nominal or ordinal

ledge pos. ables cannot be measured
through thcse
and interval vari
questions response categories are exhaustive and quest[ons;
wing five 4. questions are sclt-contained and requiremutua/ly exclusive;
few nstructions; and
en closed. 5. the sample has a lower educational level.
On the other hand, the
where: open-ended questions should be used
approp
ondent 1. questiOns are complex and responses cannot be put in a few sim
inform ple categories;
ssess respondents' unique views are to be elicited;
1estion 3.
investigatioDs are preliminary;
5f the re accuracy, details and exhaustiveness are mnore
5 ratio and internally scaled variables.are to be important; and
dent nnay Schuman and Presser (1979:709) conducted ameasured.
study on most in
answers portant problems facing the country (inflation, crime,
cn-ended open-ended and closed-ended questions and concludedetc.) comparing
that they could
e the ex not definitely say whether one form would always be more valid than
the other. However, they suggested that
researchers who wish to use
to have closed questions can begin with open questions. Although Schuman
and Presser fail to demonstrate the
to have superiority of
nore ap tions, Bradburn and Sudman (1979:19), as alreadyopen-ended ques
pointed out, do
ured on maintain that open-ended questions are superior when threatening is
opriatc, sues are being studied.
him t3irect and indirect
mongs questions
These questions rfer to relationship not between question and re
less response but between question and the object of the response. Direct
o him questionsare personal questions which elicit information aboutrhe re
I of relspondent himself/hersel[ e.g., "Do you believe in God?" Indirect
questions seek information about other people, e.g., "Do yoy think
knowl that people of your status and age believe in God now-a-days?"Other
CXamples are:
wledge Ii: Indirect question: Do college teachers these days read more English
DCuS On:
3ques or Hindi books?
Direct question: Do you read English books?
should : Indirect question: How would you describè the relations among
members in your family?
elyfew Direct question: Do you quarrèl with your spouse frequently/oc
casionally/ rarely/never?
 210 Questionnaire and Interview Schedtule

preferab
Nominal, ordinal and interval questions Q. Do
lominal question is one in' which its response falls in twoor more
(a)
categories, e.g., male/female; rich/pöor; married/unmarried; rural/ur.
ban; illiterate/educated; Shia/Suni; Hindu/Muslim) lowever, the (b) I
categories haveno rank order. Nominal question is also called classif;. If y
months:
cation scale.Ordinal question is one in which the responses are placed
in rank order of categories The categories may be ranked from high.
est.to.lOwest, greatest to least, cr. first to last. Thete is no implicatißn (b) Filte
that there is an equal distance between succeeding categoris. Here arc
hese q
some exxamples: of the re.
1. Smoking: regularly/occasioaally/never tion, e.g
2. Reserving 33 per cent seats for women in Parliament: Agree/dis to wards
agree/don't know already
3. Relations with colleagues in ofice: excellent/satisfactory/dissatis. | smoke)"
factory/can't say
Ordinal scales are sometimes referred to as ranking scales. Interual P
question is one in which the distance between two numbers is equa.
For éxample: The cont
1. Present age:10 or below/i1-20/21-30/31-40/41 and above : tion of
2. Income per annum: Below Rs: 13,000/18,o00-36,000/36,000 accessibil
54,000/54,000-72,000/Above 72,000 type of is
3. Age at marriage: Below 18/18-22/22-26/26-30/Above 30 and Sara
must.be
Other types of questions
(a) Contingency questions Double
A contingency question is one 'whose relevance to the' respondent is One que
determined by his response to an eailier screening question, e.g., ole, Do
Q.1 Are you in favour of using some method in cóntrolling birth? SCs, ST:
question
Q2. Doyou prefer vasectomy/concom/pii/safe period?
Th¹, second question is. a contingency. question. Other exampl OBCs ar
could be: Df ethni
Qi1. Do youuse this produçt? such, th
Q.2. Since when are you using this product? make it
Here Q.2 is a contingency.question. The dËrection to this type of should
question will be: It yes, to. Q.1 answer Q.2;. if no, skip to Q.3, seats for
The need for the contingency question arises
tíon need not be relevant to all respondents. The use because every ques? )
of contingernc
questions can be reduced by draing a homogencous sample. The (üi)
 Questionnaire and Interview Scledule 211

oreferable format for contingency question would be as follows:

vátching movies?
0. Do you go to cinema houses for
re (a) Ycs
(b) No
If yes, how often do you go? (a) once in a month; (b)
he Once in few
fi months; (c) once or twice in a year.
ed
(b) Filter questions
hese questions aim at eliciting information related-to ageneral aspect
of the research itopics and are usually followed, by more specitic ques
tion, e.g., Do,you smoke? :;7 guestions aregeared
towards eliciting additional and especific information'on an 1ssue
already addressd by a filter question, e.g., "Do you (being a girl)
tis smoke?»

al PITFALLS IN QUESTION CONSTRUCTÍØN OR

QUESTION CONTENT
The content of questions is most important element of:the onstruc
tion of a questionnaire.While order ofiquestions : may influeñc
0- accessibility to information; the content of:questions 'willAead to the
ype of information sought in the study. Becker{1989), Mahr (1995)
and Sarantakos(1998:237)-havheld.that the.content. of qustions:
must be organisd as to avoid the folloing types of questions:
Double-barreled questions
19
One question should not contain twO Qrmore questions. For.exam
ple, Docs your.office have a recruitment policy, of reserving seats for
SCs, STs, OBCs and women" This guestion, in fact includës tour
questions. An office may have reservation fÍr SÇs andsTS but not for
OBCS andwomen, or it may have reservation for allthree categoties
of ethnic minorities (SCs, STs and oBs) but not för WOmen.

such, the worcing in the question confuses the respondent and can
nake it difficult for him to say yeso nß THiSThe questitiôn
should be: Does your office'have a recruitment policy. ofreserving
seats for:
SCs/ST
OBCs Yes
(iii) Wcmen Yes
 212 Questiornnaire and Interview Schedule

Since qucstions-uitharoLAnd'An be double-barrcled ques.

tions, one ansuo eannet-beexpcctcd. For example, the question, "Do
you contribute for newspapers and magazines?" should have been
worded as: Do you contribute for the following:
Newspapers Yes No

() Magazines Yes No

Ambiguous questions
Sometimes the words/termsused are-vague-andambiguous, e-g-, p0
litical elite, joint family, social development, women's empowerment,
and so on. Instead of term political elite', if words used are those top
political leaders who are the decision-imakers in politics like central
ministers, chief ministers, party presidents and secretaries, etc.", the
respondents can casily reply the question. Similarly, in the term joint
family', the respondent may include his son living separately with his
wife and children, while the researcher may view it as two separate
households of father and son. Therefore, the question should be:
Who arc the members of your family who live together under the
Commnon rÍof, have a' common kitchen and function under one
authority" Later ön, the researcher can interprete this family accord
ing to his own perspective. Take the ward classAsking a question,:
"do youbelong-to-Bower;-iddBeoruppcca"2wouldbeambiguous
for the respondent. The correct thing wòuldbe-teask, "what is your
monthly tam1y.JncQmeG
ometimes, the.question may not be easy,
Suppose, the question is: Were you educated in private or public
school?" It is possible that the respondent might' have studied for few
years in a private school' ánd for the remaihing period in a public
school. In such case, how will he answer the above question? Since a
Single question can get two answers, such questions should be prop
erly constructed. Slangs should also be avoided in qucstion wording.

Difficult wording questions

Sometimes, in an effortOuse-one wOrdin place of many wods, the
researcher uses difficul words in-the questions For example, asking
"what factors prevent social mobility in a society based on ascr1ptive
oriented stratification?" should have been asked in simple words
society?" Simi
what prevents changing social status in a caste-based
dipsomaniac?" coild
larly, the question, "is anyone in your family a
 Schedule 213
Questionaire and Interview

here
anybody in your family a drunkard?"" But
"is
lhave bcen asked as,dipsoManiac may not be understood by all respon
respondents as it
lso, just as term term 'drunkard' might resent the How fre
lents, sinilarly Lhe question could therefore be:
The alcohol who is used to
Ihas a derogatory label. your family consumes
quently the member of understand difficult words depends up9n the
(aking it?" The ability torespondentsSince all respondents.who receye ways
educationallevel of the to be highlyeduçated, it is al
presumed
questionnaires cannot be easy-to-understand words. For,exanple,;a
Jesirable to use simple and role performance lead to hyperten
question, does multiplicity
of simsltaneous participation
simple terms: "does
sion?" can be asked in crcate strains and stresses?
in varied types of activities
having some knowledge about the sub
Sometimes, the researcher uses such words which the group
under study value it as symbol of
culture of the group outsider to use because they
in a
crimninals
does not want any the researcher is studying protec
authority. For example, their oWn argot for interaction and realise
use to
prison. The criminals these terms, the criminals fail
tion. Wlhen the
researcher uses
using their words only for the pur
that he is a social
scientist who is
therefore, desirable for the
their organisation. It is,
poses of studyinguse because of which the respondenits re
researcher not to such words required informn¡tion.
to give the
him and fail the.researcher
fuse to cooperatewithresent the use of such words. by teasing girls,
The youtl1 may also vords about their drinking habits,
when heuses specific on.
so
dress and speech, and
Abstract questions and specific
answers. The
Questions should have specific education"
meaning
confüses the respondent
family education' It is
question, "describe your with the 'concept fämily educational level but
fámiliar
becausc he is not and siblings'
describe father's, mother's study of family in Mahuva
casy to his
family education. I.P. Desai, in education by studyingandthe
not family
level of member in school/college
(Gujarat), determined the by each family total number of members who
:umber of yearsspent
sum total of years by the correct nethodot assess1ng
dividing the or'may not be a
may cannot be ex
were educated. This family but respondents surelyquestions about
cducational level of a methods. Similarlydiffic°lt to answer.
unfamiliar
pected to use such hanpincss, iustice, etc., are
ibstract concepts like
 214 Questio72naire and Interview Schedule

Answers to a question like How happy you are with the pattern of
education in your university" in the form of---very happy, moder. b

ately haPpy, less happy, very unhappy--may have low rcliability.

fa
(v
Leading questions clo
Sometimes,the question is so framed that.the.respondentfinds answer ga
to it in the.question itself. For example, the question: "It is said tha les
thè BÊP. reverses in seats in Uttar Pradesh in 1999 parliamentary ele Op
tions were the result of in-fighting. Do you agree?" Or the question
The miserable performance of the Akali Dal in Punjab in 1999 elec
tion was due to promotion of family hegemony.coupled with the
Badal-Tohra split. Do you agree?" The structure of the question mini
mises orimaximises-theprobability ofbias in respondent's answer. In go
ste
the above two questions, the form.of the questions artificially in
1.
creases the probability of a particular response. It is, therefore,
necessary that the question should be asked in its neutral form.

Sensitive questions
2.
Sometimes sensitive qucstions on sex, method üsed forpreventing
childbirth, money açcepted'as illegal gratification by the colleagues it
the department, homosexuality in hostels and prisons, 'etc., fail to get
3.
answers. Bailey (1982:121) calls it 'sociai desirability' bias. The prc
SCriptive norms (telling what one should do) as well ás
the proscriptive
norms (telling what one should not do) prevent a'person from giving 4.
true responses. A criminologist working on developing atheory on
the causes of crime (say, muurder onlygenerally fails to get correct in-: 5.
formation, from the persons aççused of committing murders as the
respondents think thathe soIeCt response may be used tor prosecut 6.
ing them. Since the respondents have nothing to gain but more to losç
by. revealing true answers,. they hardly take
with the'researcher. interest in cooperating 7.

Sundman and Bradburn (1974:50) studied the role of

structure in. the questionnaire, i.e., question length, questionquestio, 8.

culty,.open-ended or closed questions, position of the diffi

questionnaire, and so on. They, found that these question the
in 9.
the response in the factors did not affect
non-threatening
sponses in the threatening questions. questions but these, did affect re
For example, they found that ()
words) are less threatening than long short questions (of 12 or less
questions (of 30 or morewords); Ho
 Questionnaire and Interview Schedule 215

( ) more difficult questions receive interior responses; (11i) threatening

hehavioural questions yield worse responses if placed early in the
auestionnaire than if placed later; (iv) for a simple 'yes' or 'no' answer,
factors like question format and question length make littledifference:
) open-ended qucstions obtained higher levels of response than
closed questions; and (vi) questions on behaviour (ike sex, drinking,
gambling) which posed some degree of threat to respondents yielded
less answers than questions which did not pose such threat (sec Bailey.
op.cit.: 123).
IN QUESTIONNAIRE CONSTRUCTIÖN
Questionnaires are corstructed in a systematic manner,The process
goes through a number of interrelated steps. Th» most commnonly
steps are (Sarantakos, 1998:239-40):
1 Preparatioz: The researcher thinks of various items to be covered
in the questionnaire, arrangement of these items in relation to one
another, and taking into consideration questions prepared and
used in other similar studies.
2 Constructing the first lraf: The researcher formulates a number of
questions including direct/indirect, closed/open-ended and pri
mary/secondäry/tertiary questions. )
3. Selfevaluation: The researcher thinks abour relevance, symmetry,
clarity in language, etc.
4. Exiernal evaluation: The first draft is given to one or two ex
perts/colleagues for scrutiny and suggestions for changes.
5. Reyision: After receiving suggestions, some questions are elimi
nated, some changed and some new questions added.
Pre-test or pilot study: A pre-test or a pilot study is undertaken to
check the suitability of the questionnaire as a whole.
7. Revision: The minor and major changes may be made on the bas1s
of experience gained in pre-testing.
8. Second pre-testing: The revised questionnaire is then subjected to a
second test and arnended, if necessary.
9. Preparing final draft: After editing, checking spelling,space for re
sponse, pre-coding, the final draft is prepared.
PRE-TESTING OF QUESTIONNAIRE
Howsoever carcfully the questionnaire might have been prepared,
 216 Questionnaire and Interview Schedule
size rmay bc vcry
some confusing, missing, inap.
there could be some ambiguities, and unanswerable naires may be li
questions. There
propriate, redundant, inadequate or open-cnded questions. It
interviews. Thnus
could also be insufficient space for answering and remove such most ofthe repl
pre-test the questionnaire
is, therefore, necessary toqucstionnaire may be administered to a fey to complete. In
questions. For, this tihc ultimately. Pre.
who are similar to those who are to be studicd "don'tAccessibilit:
persons
the actual respondents. Many
testing should not be done on poorly-worded qucstion which nced When the resp
kncw' answers may indicate a in the same manner rcached by corr
should be conducted
to tedeleted. The pre-test questionnaire, the prc-test should als
the final study. If it is a mailed
schedule, the pre-test should be through ín No intervi
be mailed. If it is interview
terview. Since the inter
researcher has first to deal with unanswered he cannot infl
After the pre-test, the which get similar responses
questions, followed by those questions be deleted. He should then oWn opinion o
therefore
from all respondents and may opinions given by the respondents
suggestions, comments and 8. Greater ar
take up
words or deleting some offending questions. How-|
for.adding some
researcher nced not accept all suggestions of the respondents. The absence c
ever, the spondent to
LVANTAGES OF QUESTIONNAIRE undesirable c
cO the respon
have strengths and weak
Questionnaíres, as a tool of data collection, events 2nd 1nc
disadvantages. Some advantages of
nesses and thus advantages and Singleton and Straits (1999:259),
questionnaires, as pointed out by 6. Responde
(1998:224) are Jisted below:
Selltiz et al (1976) and Sarantakos
Ihe respond
1enceHe is
Lower COst
are less cxpensive than other inethods. Even the staff 1lls up Lie q
Questionnaires tirst and take
not much as either the researcher himself may mail or one
requircd is the ques
appointed for hand-distributing
or two investigators may be TA/DA to the investigators and thejStandar
tionnaires. Giving salary and survey. In questionnaire (be
research officers incrcases the cost of the money on
Each respon
sides the cost of printing) the researcher has, only to spend difference i!
get
questionnaires and stamped envelopes for The thus facilitau
postage for sending the
or, on follow-up letters.
ting back the filled-in questionnaires
mailed questionnaires thus cost less. 8. No vari
Questionna
2. Time saving variation.
the respondents may be geographically dispersed and sample
Since
 Question naire and Interview Schedule 2.17

size may be very large, the time required for getting back the question
naires may be little greater but usually less than that for face-to-face
interviews. Thus, since all questionnaires are sent simultaneously and
h most of he replies are received in 10-15 days, schedules take months
toComplete. In simple terms, questionnaires produce quick results.
3. Accessibility to widespread respondents
When the respondents are separated geographically, they Can be
rcached by corrcspondencewhich saves travel cost.
4. No interviewer's bias
Since the interviewer is not physically present at interviewee's place,
he cannot influence his answers, either by prompting or by giving h1s
oWn opinion or by misreading the question.

8. Greater anonymity
The absence of the interviewer assures anonymity which enables re
spondent to express free opinions and answers en ta sacial
undesirable questions. The absence of the interviewefassures privacy
to the respondents because of which they willingly give details of all
otherwise.
events and incidents they would have not revealed

6. Respondent's convenience
at his conven
The respondent can fill-in the questionnaire leisurely time. Since he
to complete all questions at one
ience He is not forced to
he can answer easy questions
1ills up the quesionnaire 1n spare time,
e first and take time for difficult questions.

z. Standardised wordings
to same words and therefore there is little
Each respondent is exposed questions. The comparison of answers 1s
difference in understanding
thus facilitated.

8. No variation without
QuestiOnnaures are a stabl, consistent and uniform mneasure,
variation.
 218 Questioyaire and Interview Schedule
Ouestionnaire and Interview Schedule 219

LIMITATIONS OF QUESTIONNARE Interview

1. The mailed questionnaires can be used only. fof Facors Questionnaire
Schedule (Mailed)
This restricts the number of educated people.
2. The return cate.af
respondents. Resporise rate High Low

is 30 to 40
questionnaires is bowThe common return rate Cost High Low
per cent.
3 The mailing address may. not be correct which may Control on situation High Low

gible respondents. Thus, the sample sclected many omit some eli:plicability to geographicaly cispersed population Moderate High
described as biased. a time i Applicability to heterogeneous population High Low

Sometimes different respondents interprete questions differently Obtaining detailed anc additional in-ormation
4. High Moderate
Low
The misunderstanding cannot be corrected. Speed (tinme) Low

5..There. may be bias in the response selcctivity because the responProbing, prompting ànd classification of questions High Low

dent having no interest in the topiç may not.giyeresponse to all

questions. Since the researcher is not present to explain the mean THE COVER LETTER
ing of certain concepts, the respondent may leave the question The cover letter aims at explaining the research topicintroducing the
blank. objectives of the study, describing how the respondent was selected,
6. Questionnaires do not provide an opportAnityte--sollectaddi-giving some important instructions for answering the questions, moti
tional information-uhile-they-are-beig-tormpleted. vating respondents to participate in the study and assuring anonymity
7. Researchers are not sure whether the person to whom the ques and confidentiality and reoving their doubts or mistrust. Here is an
tionnaire was mailed has hinself answered the .questions or example:
somebody else has filled up the questionnaire:
Many questions remain unanswered. The partial response affects We are suveying students and teachers in selected colleges and univér
the analysis. sity departments for a UGC sponsored programme to find out ro
what extent the teaching is considered satisfactory, and to ascertain
9 The respondent can consult orher persons before filling in thè the deficiencies in the existing educational system. Your name was se
questionnaire. The responses, theretore, cannot be viewed as his lected at random from the list of studenLs/teachers supplied by the
opinions. colleges/departments. Our question naire will take no more than 20
10. The reliability of respondent's background information cannot be minutes. Kindly answer all questions.
verified. A míddle-class person can identify himself as rich person
or a person of intermediate caste can describe himself as upper The cover letter is thus to contain the following main points:
Caste person. Identifying the researcher and the research sponsor.
11, Since the size of the questionnaire has to be kept small, full infor Explaining the social significance of the study.
mation cannot besecurçd from the respondents.. Pointing out the main objectives of the study.
12. There is lack of depth or probing for a more specific answer. Specifying requirements for completion of questionnaires.
Nachmias and Nachmias (Researcb Methods in Sõcial Sciences, through brief instructions.
1981:202) have compared advantages and disadvantages of question Giving reasons for respondent's cooperation.
naireand interview schedule in termns of eight factors as follows: Assuring anonymity and confidentiality.
Giving approximate time the questionnaire is likely to take in till
ing up.