Chapter 4

Questionnaire Design

Learning Objectives

Upon completion of this chapter, you will be able to:

• Understand the meaning and importance of questionnaire
• Understand three stages of the questionnaire design process
• Learn about pre-construction phase of the questionnaire
• Learn about construction phase of the questionnaire
• Learn about post-construction phase of the questionnaire

4.1 Introduction

To conduct surveys, researchers always need a good questionnaire. There are some ­problems in designing good questionnaires because there is no universally accepted rule to be taken as the guideline while preparing the questionnaire. Theories will always enhance the ­construction of the questionnaire, but the experience of a researcher has a decisive role in the construction of a questionnaire. In the field of business research, many interviewers use the same questionnaire to conduct any survey. They cannot use different questionnaires to measure the same characteristics of interest. There must be a common questionnaire for the use of all the interviewers to avoid the bias in the research. A good questionnaire should be designed such that it will be able to fulfil research objective. The construction of the questionnaire has many constraints such as number of questions to be included, order of questions, physical form, and so on. Furthermore, increasing scope of business research has given an option to surveying firms to conduct continuous interviewing. Excessive interviewing is also a serious problem to the researchers. In his research paper titled “Threats to Survey Research: Excessive Interviewing and ‘Sugging’”, Mike Brennan has warned the researchers about the changing public attitude towards survey due to excessive interviewing and “sugging” (selling under the guise of research) in New Zealand.¹ Although the study was conducted in New Zealand, the findings are true for researchers worldwide. Designing does not involve picking questions randomly and abruptly incorporating them into the questionnaire. It is a process based on logic and systematic effort, which is rather scientific in nature.

4.2 What Is A Questionnaire?

A questionnaire consists of formalized and pre-specified set of questions designed to obtain responses from potential respondents. Questions in the questionnaire reflect the research objective under investigation. For example, suppose a researcher wants to measure ­consumer attitude for a product and is in the process of designing a questionnaire. To ­measure the consumer attitude, the researcher has to first quantify the consumer attitude, which inherently is a feeling. The researcher has to collect relevant statements from the literature and then convert them in the form of questions to quantify a feeling such as the consumer’s attitude. Finally, he or she prepares a set of questions that ultimately reflect various dimensions of consumer attitude rated on a rating scale. The summated score obtained from each consumer is the consumer attitude score for the concerned respondent. Similarly, any other feeling can be measured using a well-structured questionnaire. Questionnaires are generally situation and culture specific. So, the same questionnaire for measuring attitude cannot be used in all the cultures and situations. Questionnaire design process requires a careful attention to each step as the questionnaire or research instrument should be adapted to the specific cultural environment and should not be biased in terms of any one culture.² Following section specifically focuses on the questionnaire design process.

4.3 Questionnaire Design Process

Designing of the questionnaire is a systematic process. This section explores the systematic process of questionnaire design in three phases: pre-construction phase, construction phase, and post-construction phase. Figure 4.1 shows all these steps.

Figure 4.1 Steps in questionnaire design process

4.3.1 Phase I: Pre-Construction Phase

Phase I is the pre-construction phase of the questionnaire design process. It consists of three steps: specific required information in the light of research objective, an overview of respondent’s characteristics, and decision regarding selecting an appropriate survey technique.

4.3.1.1 Specific Required Information in the Light of Research Objective

This is the first and probably most important step of questionnaire design process. A poor conceptualization at this stage will lead to collection of a lot of irrelevant information and, more than that, this will also lead to loss of some relevant information. This will unnecessarily waste researcher’s time and energy that is always precious and must be conserved. This also will lead to the increased cost of project in hand. So, the first and the foremost important decision relates to finding the required information through a questionnaire. To generate specific information, the researcher should clearly define the objective and the other research components such as theoretical model, research questions, and hypotheses. Clarity in all these components is very important as these only will provide the base of specific information to be collected through survey. In addition, the exploratory research can be launched to identify relevant research variables.

4.3.1.2 An Overview of Respondent’s Characteristics

While collecting the information, an overview of the respondent’s characteristics is a vital consideration. A researcher must construct the questionnaire in the light of the respondent’s profile. For example, “purchase intention” is directly related to purchase capacity of the individuals. A questionnaire discovering purchase intention for a particular product, administered to various respondents without keeping their earning in consideration, will always lead to vague data. Hence, respondent’s financial status and accordingly preparation and administration of the questionnaire are the key to generate relevant information. Similarly, while assessing the impact of “new tax policy,” the researcher has to contact tax payers or potential tax payers. Simply a random selection from the population and then administration of questionnaire is not going to solve the purpose. Even while constructing a questionnaire, a researcher has to keep the profile of a probable respondent in mind. Various factors such as the respondent’s qualification, age, experience, income, marital status, occupation, and so on have a decisive and impactful role in conducting a research. All the respondents may not be alike in all these factors and this is the reason a different profiling and different consideration with respect to this profiling are very important to construct a questionnaire.

4.3.1.3 Decision Regarding Selecting an Appropriate Survey Technique

While constructing a questionnaire, a deep thinking process is required to select an appropriate survey technique. On the basis of the mode of administration, personal interview, telephone interview, mail interview, and electronic interview are commonly used survey techniques. Each of the techniques has its relative advantages and disadvantages. Chapter 7 provides a detailed discussion of various survey and observation techniques and its relative advantages and disadvantages. The questions in the questionnaire must be constructed, sequenced, and placed according to the mode of the survey. For example, a lengthy and difficult-to-answer question must be avoided in a telephone interview method but may be appropriate in personal interview technique.

4.3.2 Phase II: Construction Phase

Phase II is the real construction phase of the questionnaire design process. It consists of six steps: decision regarding question format, structured questions versus unstructured questions; decision regarding question relevance and wording; decision regarding question sequencing; decision regarding question response choice; decision regarding the questionnaire layout; and producing first draft of the questionnaire.

4.3.2.1 Decision Regarding Question Format: Unstructured Versus Structured Questions

Questionnaires use two types of question formats. These are open-ended questions and closed-ended questions. In the case of a “closed-ended” question, the respondent has to format the judgment to fit the response categories and when “open-ended” questions are used, the judgment has to be verbalized into a preliminary answer.³ The closed-ended question format can be further divided into dichotomous, multiple-choice questions, and scales. The following sections focus on open-ended questions and closed-ended questions.

Open-ended Questions

Open-ended questions are unstructured questions. The open-ended questions provide a free-to-answer opportunity to the respondents instead of fixed-response choices. In an open-ended question, a respondent remains free to provide his or her opinion about any topic in his or her own words. While narrating an opinion or attitude, the respondent provides a deep insight about the phenomenon under study. This really helps a researcher to design a questionnaire in a more structured manner. Hence, the open-ended questions are extremely useful for an exploratory research. In addition, these questions provide the respondent an opportunity to freely express his or her feelings and provide an insight into the problem in hand. Here, the researcher must be very cautious that the respondent should not be overburdened with more than what is required. To meet the objective of getting the required range of answer, the researcher can deliberately provide a pre-decided space for writing the answer. This can check the excessive and unimportant writing without mentioning it to the respondent. Following are some examples of the open-ended questions:

The open-ended questions also have some serious limitations. One of the major limitations is to handle the interviewer and the interpretation bias. While writing the answers, the interviewers generally use their own way of writing instead of a verbatim answer. This sometimes distorts the main issue and unnecessarily opens the different dimensions that may not be very important for the research. To get the required answer, the researchers are supposed to use tape recorders. Answers to the open-ended questions require a careful interpretation. In any case, it is very difficult to have a subjective interpretation without human bias. Open-ended questions have the tendency to provide an extra importance to the respondents who are more expressive.

In some cases, the respondents do not feel comfortable with the open-ended questions as it requires more effort and time to fill. In general, the respondents do not welcome any survey. The researchers try to motivate the respondents to participate in the survey through various means such as providing incentives. Long and descriptive type of answers really frustrates the respondents and, as a result, they try to avoid it. Even when the respondent has an idea about the matter under investigation, he or she finds a great deal of difficulty in expressing it because there is no response choice available to him or her. In case of a telephone interview, providing the open-ended question may not be a good choice. The open-ended questions are also not good for self-administering questionnaires. The open-ended questions are difficult to code and require a lot of time.

Closed-ended Questions

Closed-ended questions are structured questions. The question structure does matter.⁴ The closed-ended questions provide response alternative to the respondents instead of giving them a free-to-express response option. The choice offered to the respondents can be either in the form of a rating system or a set of response alternatives. These response alternatives are presented to the respondents and they select the most appropriate one. The closed-ended questionnaires are generally cheaper, more reliable, and faster to code, and analyse the collected data.⁵

The closed-ended questions have several advantages. Administration of the closed-ended questionnaire is relatively easy as the need for explaining question dimensions is minimal. The closed-ended questions reduce the respondent’s burden as the respondent is provided with the response alternatives and the burden of expressing a feeling or an opinion in his or her own thinking is the least. In this manner, the closed-ended questions motivate the respondents to complete the survey. A structured questionnaire also reduces the interviewer bias as the respondent has to select from a fixed alternative list. As the question is already structured, the interpretation bias is also reduced, which ultimately saves time and cost. Coding and tabulation of data also become very easy as the response alternatives are fixed and the researcher has no burden to create categories from a wide range of different narratives provided by the different respondents.

The questionnaire with closed-ended questions has several disadvantages as well. These questions are very difficult to construct and require some expertise. The researcher has a limitation in terms of providing all the exhaustive alternatives to a respondent and cannot provide a list of 30 possible answers to the respondents. He or she has to go for an extensive study of the subject under study and has to obtain an expert input to provide the most possible alternatives to the questions. This is of course a time- and money-consuming exercise and requires a lot of effort. To cover all the alternatives to the questions, the researchers specify an alternative as “Please specify if any other.” When many respondents select this option, the essence of closed-ended question diminishes as the obtained answer must again be categorized by the researcher. As discussed earlier, the closed-ended question format can be further divided into dichotomous, multiple-choice questions, and scales.

Dichotomous Questions

Dichotomous questions have only two response alternatives usually presenting the two extremes “yes” or “no.” To make the alternatives balanced, the researchers often present a third neutral alternative “don’t know.” For example, to boost the sales, a motorbike company has reduced the price and provided “0% interest scheme” to potential purchasers, as a special scheme for 1 year. The company now wants to uncover the purchase intention of the potential customers and asked a dichotomous question with a set of other questions such as:

The researchers also ask dichotomous questions to understand the demographic profile of the respondents. For example, an insurance company has launched a new policy with special features for the children. To understand the response, the company can ask the following dichotomous questions:

The main advantage of dichotomous questions is that they are simple to construct and administer. This does not require expertise and requires less time to construct. On the one hand, these questions are easy to code and analyse. Interpretation bias is also less as the responses are less (two or three) and are clear. On the other hand, the disadvantage is that these questions sometimes generate biased responses if the researcher phrases the question in a particular way. Leading questions (discussed later) generally generate biased responses. Another problem with the dichotomous questions is the forced choice available to the respondents in terms of answering yes or no when the researcher has not included a neutral response. To avoid the forced-choice option, if the researcher includes a neutral option, the respondents will conveniently prefer this neutral option and avoid preferring yes or no choices. Naresh K. Malhotra in his book Marketing Research has presented the guidelines to include neutral option in the multiple choice questions. He says: “If a substantial portion of the respondents can be expected to be neutral, include a neutral alternative otherwise do not include the neutral alternative.”⁶

Multiple-Choice Questions

While asking multiple-choice questions, the researcher presents various answer choices to a respondent and the respondent is supposed to select any one from the options. The multiple-choice questions are also referred as multichotomous questions. A private bank assessing the customer intention to have a relationship with the bank can ask the following multiple choice questions:

The advantages of multiple choice questions are same as that of the dichotomous questions. Similar to the dichotomous questions, the multiple choice questions are easy to administer and to get the responses. It also reduces the interpretation bias and leaves little scope for the interviewer bias. These questions are also easy to code and analyse. The number of alternatives to be provided to the respondents is a key consideration while framing the multiple-choice questions. A researcher should present all the possible mutually exclusive alternatives to the respondents. As discussed already, to make the list exhaustive, an alternative, “if others please specify,” must also be included. The order of alternative presented to the respondent is another problem with the multiple-choice questions. The respondents have the tendency of electing the first and the last alternative, specifically the first one. To avoid this problem, a researcher can prepare several sets of questionnaire by changing the order of alternative in each. In this manner, all the alternatives will be having a chance to secure the first-order position as well as all the positions in the order. This helps to overcome the bias in selecting a particular alternative by the respondents from the list of alternatives.

Scales

Scales are also closed-ended questions, where multiple choices are offered to the respondents. Scales are discussed in detail in Chapter 3.

4.3.2.2 Decision Regarding Question Wording

Question wording is a typical and an important aspect of the development of a questionnaire. Negative wording of questions discourages the respondents and decreases the response rate, and even if the respondent responds, it may be a biased answer. The negative wordings impact the process of interpreting the questions, leading at least some respondents to misinterpret how to respond and, thus, reducing or destroying the usefulness of the questions or the series of questions.⁷ In some cases of conducting personal interviews, the researchers feel that they will be able to overcome the wrong wording during the discussion, but they suggest a different understanding. The effects of question wording on response accuracy did not seem to be moderated by the respondent’s knowledge gained via discussion.⁸ A small change in wording a question can either encourage or discourage the responses. The guideline to evaluate the questions that are to be incorporated in the questionnaire is shown in Figure 4.2.

Figure 4.2 Decision parameters regarding question wording

Question Wordings Must be Simple and Easy to Understand

While preparing a questionnaire, the researcher must always be careful as the respondent may be a common person who will not be able to understand the technical language. He or she should use simple, easy to understand, and direct words. What is very simple for the researcher may be very difficult for a common person. So, while assessing the difficulties in understanding the word, a common person’s understanding must be taken into consideration instead of that of a group of highly educated persons. For example, instead of asking:

What is your perception about the new advertisement campaign?

A researcher should frame the same question as:

What is your view about the new advertisement campaign?

A respondent may not always be familiar with the word “perception.” This word can rather perplex him or her leading to non-response or a misleading answer. Both the situations are not healthy for the research. Instead of using a relatively heavy word such as “perception,” a researcher can use a light and easy-to-understand word “view.”

Vague or Ambiguous Words Must be Avoided

Some words such as “often,” “occasionally” and “usually,” “how long,” “how much,” and “reasonably well” may be confusing for a respondent because these words specify a specific time frame. In addition, different words may have different meanings for a respondent. For example, the below question to determine consumer intention to spend on shopping in a year generates confusion for a consumer as he or she is not able to decide what the researcher’s intention is.

In a year, how much will you spend on shopping?

In this case, the respondent will be using his or her opinion about the question and answer alternatives, and accordingly he or she will be answering the question. A much better way of asking the same question may be as follows:

In a year, how much do you plan to spend on shopping?

This presents clear options to the respondent, and his or her assessment options are closed as the researcher has given five clear options to get the respondent’s intention to spend.

Double-Barrelled Questions Must be Avoided

Double-barrelled questions are those with wordings such as “and” or “or.” In a double-barrelled question, a respondent may agree to one part of the question but not to the other part. These questions also pose a situation where the respondent cannot provide his or her answer to the second part of the question without accepting the first part. For example, instead of asking a question:

In this type of question, the second part assumes that he or she prefers outing with the family in the weekend without even asking the respondent. Furthermore, the respondent is perplexed if his or her answer is “no” to the first part of the question. Instead of asking this type of double-barrelled question, the researcher can ask this question in two parts:

Dividing the question into two parts avoids confusion for the respondent. Only if the answer to the first question is “yes,” the respondent will be attempting the second question. If the answer to the first question is “no,” then the respondent will be ignoring the second question and, hence, there will be no confusion in answering the questions.

Avoid Leading and Loaded Questions

A leading question is the one which clearly reveals the researcher’s opinion about the answer to the question. This is a biased way of presenting a question to get the desired answer. For example, a leading question can be presented as follows:

The question is presented in such a manner that the researcher gets a desired answer. A slight emphasis on the word “Aren’t you” will change the respondent’s answer. More specifically, asking the question such as “Are you satisfied with the new feature of the product?” will be more unbiased way as it gives freedom to the respondent to provide either “yes” or “no” answer.

Identifying the loaded question bias in a question requires more judgment because the wording elements in a loaded question allude to the universal belief or rules of ­behaviour.⁹ The loaded questions are posed to address the inner feeling of the respondent and the response is almost predestinated. For example, a loaded question can be given as follows:

This question is a specific loading question and probably every respondent will like to answer it as “yes” because the question indirectly addresses their feeling of patriotism. Elimination of the words “patriotic Indian” from this question will definitely yield a different answer. The more accurate way of asking the same question is “Every Indian will prefer an Indian brand of soap.”

A leading question generally emphasizes either the positive or negative aspect of the question. Sometimes, it becomes necessary to ask a question with either the positive or negative aspect. In this situation, a split-ballot technique is used to avoid bias due to positive or negative aspect of the question. This technique involves the construction of a single question in two alternative phrases, and the question based on one phrase is administered to half of the respondents and the question based on the other phrase is administered to the other half of the respondents. For example, in a 1 to 5 rating scale, where 1 represents strongly disagree and 5 represents strongly agree, a question measuring the increased use of fully automatic washing machine can be administered to half of the respondents as follows:

Reverse phrasing of the same question can be provided to the other group as follows:

The split-ballot technique is used with the expectation that two alternative of the same question will be able to generate more accurate summated score as compared to the situation when a single-phrasing question is used.

Avoid Using Overstated Words

It is always better to pose a question in a natural way rather than in a positive or a negative way. Overstatement of words will always bias the individual response. For example, a question related to the purchase of a water purifier with overstatement of words is as follows:

The answer will always be overblown due to the first part of the question, which generates a worry in the mind of the respondent and results in a positive answer, which is not possible otherwise. A more poised way of asking the same question is shown below.

Implied Assumptions Must be Avoided

While constructing a questionnaire, implied alternatives must always be avoided to avoid unnecessary confusions. For example, the question:

has an implicit assumption that the discount policy on bulk purchase offered by Company “A” is working excellent and by answering “yes,” the company will continue its policy. The respondent assumes that giving a no answer will encourage the company to stop the scheme. Posing this type of question creates a binding on the respondent to provide his answer in a certain pre-determined manner. A more poised approach of asking the same question may be as follows:

This way of asking a question emphasizes the discount scheme and seeks the respondent’s answer in an indirect manner with reference to Company A’s offer.

Respondent’s Memory Should Not be Overtaxed

In some situations, a researcher’s questions are based on the respondent’s memory. The respondents tend to forget the incidents that are not important for them. These incidents may be important for the researcher, and he or she has framed question on these incidents. The following example is a difficult question to answer:

Is is difficult for a respondent to keep a record of eating ice creams in the past 2 months. This is an unnecessary testing of the respondent’s memory. Even if he or she provides the answer, it may not be accurate, rather it would be an approximate answer. A more accurate and informal way of asking the above question may be as follows:

Generalization and Estimation Must be Avoided

A researcher should not pose questions that have a generalized answer. Generalization means respondent’s belief, “what must happen” or “what should happen.” For example, a question related to a new small car purchase in a generalized manner can be as follows:

This question has a generalized answer as every respondent will be seeking for the warranty while purchasing a car whether it is a small or a big one. The more accurate way of asking this question may be:

The questionnaire designing process must always consider that the respondent should not be left with the estimation or computation. Most of the respondents are either unwilling to compute or incapable to compute. For example, if a researcher is worried about the fact that the increasing pulse prices will lead to the low consumption of pulse in India in the year 2009–2010, he or she would ask a question:

For a common respondent, this type of question is difficult to answer. Hence, it leads to non-response or a faulty response. To answer this question, a respondent has to first assess the consumption of pulses by his family for past 6 months. Furthermore, to find the average per person consumption, this average must be divided by the number of family members. For response simplicity, this question must be divided into two components as shown below:

Average per person consumption must be computed by the researcher, and the respondent is left with the provision of answering two simple questions.

Respondent’s Ability to Answer Must be Considered

Questions must be designed in the light of the respondent’s ability and experience. For example, a question targeted to officers older than 55 years to assess the importance of Internet banking is as follows:

Most of the officers older than 55 years are not comfortable with a new concept like “internet banking.” Thus, targeting this question to this respondent segment may not be an appropriate exercise.

Another example is related to a research question concerning the difference in ­consumer attitude in pre- and post-liberalization period in India. The question may be posed to the ­concerned respondents as follows:

Targeting this question to young respondents may not be an appropriate choice. The first question is that which period will be considered pre-liberalization period and which will be the post-liberalization period. Even if the researcher takes a pre-decided cut-off as before 1998 (when the impact of reforms noticed) as the pre-liberalization period and after 1999 till date as the post-liberalization period, answering for the probable respondents may be a difficult exercise. This is because young respondents were small kids during the pre-liberalization period, and they used to shop with their parents. Hence, seeking their opinion about the purchase behaviour may not be a right step. Targeting this question to the respondents younger than 22 years will provide a misleading answer. Targeting this question to the respondents older than 35 years may be an appropriate way as the respondents of this age group would have been approximately 17–25 years in the pre-liberalization period. In this age group, they may be treated as independent consumers and are able to provide independent answers.

4.3.2.3 Decision Regarding Question Sequencing

Question sequence also plays a key role in generating the respondent’s interest and motivation to answer the question. Questions should have a logical sequencing in the questionnaire and should not be placed abruptly. The effects of question order are of two types: context and sequence. Context effects are those of consistency (when responses to the later questions become more congruous due to an earlier stimulus) or contrast (when there is a greater difference due to ordering).¹⁰ To facilitate the responses, a researcher has to follow some logical steps in sequencing the questions in the questionnaire. This arrangement usually requires the considerations as shown in Figure 4.3.

Figure 4.3 Decision parameters regarding question sequence

Screening Questions

Researchers generally begin with some screening questions to make sure that the target respondent is qualified for the interview. For example, if a company wants to assess the impact of “buy one get one free” offer through consumers who availed it through a company showroom in a shopping mall, then it has to conduct a mall intercept interview to know the impact of scheme. In this case, every person who is coming out of the mall cannot be interviewed as there is possibility that he or she may not have visited the showroom. Thus, the first and the basic question to ask is “have you visited the company’s showroom?” The “yes” response to this question can open further interviewing process and a “no” response terminates the interview. In some cases, when the researcher is very sure about the qualification status of the respondent, he or she does not incorporate the screening question and starts from some “opening questions.”

Opening Questions

The opening questions should be simple, encouraging, and trust building. From the research objective point of view, these questions may sometimes be little irrelevant but should be good initiators. These questions should not seek in-depth information and should be as general as possible. For example, a microwave company, trying to assess “shift in consumer attitude” from traditional way of cooking, should ask a first opening question as follows:

Traditionally, Indians do not use microwave ovens as a part of their cooking system. With the change in food habits, preparation instruments have also started changing. Thus, the first question focuses on this issue and avoids respondent’s confusion and makes him aware about this change without harming his prestige.

Transition Statements

A questionnaire contains several questions related to various predetermined variables. The researchers generally collect these variables from extensive literature review. Each variable has a different set of questions. The movement from one set of questions to another set requires transition statements. For example, a mineral water bottle manufacturing com­pany is encouraged with the expanding market. The company wants to assess the potential future market and hence conducted a survey on non-users. Its researchers have identified various variables to get the potential use, of which “awareness” and “taste” are important. It has prepared the first 11 questions, with the first five questions based on the “awareness” and the next six questions on “taste.” After asking the first set of five questions, a researcher moves to the second set of six questions to get the potential consumer feeling for mineral water taste. Thus, before asking the next set of six questions, a transition statement is required to make the respondent familiar with the coming questions. Hence, a transition statement “Now, I would like to understand your opinion about the mineral water taste” will develop respondent’s connectivity for the next set of six questions related to “taste,” and he or she will be in a comfortable state of mind to answer these questions.

Difficult to Answer Questions

Difficult to answer, sensitive, or complicated questions should be placed later in the questionnaire. Placing it first will confuse the respondent and he or she will tend to terminate the interview process. Start must be fairly simple and encouraging, and difficulties must lie at the later stage to make them comfortable with the interview process. In the process of asking simple questions, first, the respondent develops a rapport with the process and even when there comes a time to respond relatively difficult questions, he or she feels a moral responsibility to wind up the process. This makes him or her feel that he or she has answered many questions and that after attempting these difficult questions the process will be over. At this stage, in personal interview, an interviewer is supposed to motivate and help the respondent to complete the process. Asking difficult questions first in a telephone interview reduces a respondent’s interest in the interview process and he or she tends to terminate the interview. Under telephone interview conditions, substantively related questions affect the responses to the target question only when asked first.¹¹

Identification and Categorization Questions

The questionnaire consists of some identification and categorization questions. Identification questions are used to generate some basic identification information such as name, mailing address, office phone number, personal phone number, or cell phone number. A researcher must keep in his mind that if a respondent does not like to provide some information, he should be allowed to do so. For example, some respondents may not be interested in giving his or her personal telephone number or cell phone number to safeguard their privacy. This is important to enhance the credibility of the survey.

Categorization questions are mainly used to generate demographic information. For example, researchers generally want to generate the information related to age, experience, gender, and occupation of the respondents. In some specific cases, in the light of a research objective, the researchers wish to generate some information related to occupation, income, designation, and so on. Sometimes, these categorization information are of paramount importance for a research. Categorization or classification data are the general information collected about the respondents or their households, which is not immediately concerned with the subject of enquiry but is used to divide the respondents into groups for the purpose of analysis or sample validation.¹² For example, when the research objective is to determine the impact of age on changing consumer learning, age of the respondents is the key demographic information to be generated during the survey.

Logical Order of Questioning

In a questionnaire, the questions must flow in a logical sequence. There are at least three approaches to suggest the roadmap to place the questions in a logical sequence; they are funnel technique, work technique, and sections technique. Funnel technique suggests asking general questions first and then the specific questions. The general questions allow an interviewer to have knowledge about the respondent’s understanding and opinion about the subject matter to be investigated. Work technique suggests that difficult-to-answer, sensitive, or complicated questions should be placed later in the questionnaire. Researchers generally place the scaled questions in this category as the respondent has to put much effort to answer these questions compared with simple questions. The third technique is the section technique in which questions are placed in different sections with respect to some common base. In general, the research objective itself provides some common base to place the questions in different sections. Construction of questionnaire is mainly a logical discretion of a researcher, such that there is no strict guideline to follow. These three techniques provide a rational thinking platform to a researcher and there is no reason why a researcher cannot use a combination of all the three as guideline to construct a good questionnaire.

4.3.2.4 Decision Regarding Question Response Choice

After deciding about the question format as a closed-ended question, the researcher has to make an important decision about the number of response choices to be presented before the respondent. It is important to understand that too many response choices will burden the respondent and he or she will be perplexed while answering. Few response choices will not be able to cover all ranges of possible alternatives. A researcher keeps only two response choices when a question has to be answered availing two alternatives. For example, in some cases, the response alternatives are reasonable “yes” or “no” or when response alternatives are “male” or “female,” two alternatives are sensibly good.

As a general rule, the researchers present a question with five to seven response alternatives. It is important to understand that in many situations, the response alternatives can be many. For example, “which town you will like to visit in the next summer vacations.” As one can understand, the response alternatives may be plenty. In this type of situation, a researcher has to frame the alternatives keeping the research objective in mind. A researcher can include the towns in light of the research objective. As discussed earlier, for opening the options of including other alternatives, a researcher must include an alternative as “please specify if any other.”

4.3.2.5 Decision Regarding Questionnaire Layout

Questionnaire layout is important to enhance the response rate. A recent study revealed that a user-friendly format, and to some extent colour, is valuable to increase mail survey response rate.¹³ The questionnaire layout plays a key and decisive role in the pattern of answering. The appearance of a questionnaire is particularly important in mail surveys because the instrument, along with the preliminary letter and/or cover letter, must sell itself and convince the recipient to complete and return it.¹⁴ It has been observed that the respondent emphasizes the questions that are placed at the top of the questionnaire compared with that at the bottom. The first part of any questionnaire is the introduction part. This is a vital part of any questionnaire as it explains in brief the purpose of seeking responses. This also assures the respondents that their responses will be kept confidential so as to motivate and encourage them. Beginning part should also consist of a clear-cut instruction to fill the questionnaire, so that the respondent can navigate his or her way through the questionnaire. A researcher can keep the navigation instructions in boldface and in different font size, so that these will not mingle with the other questions.

The questionnaire should not be very long and should not also be ambiguous. To enhance comfort of the respondent, first few questions can be placed to generate some demographic information. The questionnaire may be divided into several parts, where each part of the questionnaire must have a logical base of incorporating questions. Questions of each part should be properly numbered. Wherever required, coding of questions must also be done. This is equally important to print the questionnaire on a good quality paper. If it has to be photocopied, there must be enough attention to get the photocopies on good quality papers. Bad quality papers are perceived as the diluted seriousness of the research objective by the respondents.

4.3.2.6 Producing First Draft of the Questionnaire

At this stage, a researcher decides about the appearance of the questionnaire as it is of paramount importance to generate a good response. The questionnaire appearance is particularly important in mail surveys because the instrument, along with the preliminary letter and/or cover letter, must sell itself and convince the recipient to complete and return it.¹⁵ Printing on a poor quality paper or an unprofessional look of the questionnaire may generate a non-serious feeling among the respondents. So, the questionnaire may be printed on a good quality paper and must have a professional look. The appearance of the front cover on a mail questionnaire and the nature of first questions have been purported to have an important influence on the respondent’s decision to complete the questionnaire.¹⁶ Each question in the questionnaire must be distinctly separated. To reduce the questionnaire length, the researchers may crowd the questions, which enhance the ambiguity, which in turn is more dangerous. Hence, the researchers must be careful to not crowd the questions. Instructions to the answer should be as close as possible to the concerned questions. Font size of the questionnaire must be appropriate. Both too small and too big font sizes hinder easy reading.

4.3.3 Phase III: Post-Construction Phase

Phase III is the post-construction phase of the questionnaire design process. It consists of four steps: pre-testing of the questionnaire, revisiting the questionnaire based on the inputs obtained from the pre-testing, revising final draft of the questionnaire, and administering the questionnaire and obtaining responses.

4.3.3.1 Pre-Testing of the Questionnaire

Although the questionnaire is prepared by best experts, it may have some problems that can be identified and removed through pre-testing. Pre-testing of the questionnaire involves administering the questionnaire to a small sample of the population to identify and eliminate the potential problems of the questionnaire, if any. Testing a questionnaire can be a time-consuming process, and this stage of questionnaire has often been overlooked in the past in the researcher’s eagerness to start collecting data.¹⁷ Avoiding pre-testing may be a serious miss in the development of a good questionnaire. The pre-test is used to ensure that the questionnaire is appropriate for the survey in terms of structure and language, and it enables the researcher to check that the information obtained from the target population is actually collected through the research instrument.¹⁸

Now, there exists a question that what method should be adapted to pre-test the questionnaire. J. Blair and S. Presser have put the affectivity of four pre-testing methods in the following order: expert panels, the conventional pre-test, behavioural interaction coding, and cognitive interviews.¹⁹ Expert panel is a group of three research experts who are treated as the respondents of the survey and are supposed to identify the problems with the questionnaire. In a conventional pre-test, a researcher takes a small sample from the population and follows the same procedure of interviewing as he or she is supposed to follow for the final survey. Behavioural interaction coding involves observing the individual behaviour while filling the responses and noting deviation from the standard set of behaviour. Cognitive interviews involve discovering the problems in answering the questions. In this method of pre-testing, the interviewer tries to diagnose the problems that the respondents are facing in understanding the questions or in recalling the answer or any other related problem.

Researchers generally use two common procedures to pre-test: protocol analysis and debriefing. Using protocol analysis, a researcher asks the respondent to “think aloud” while answering the question. In this manner, the researcher is able to read the respondent’s mind. His or her response to different questions is noted and analysed. This gives an opportunity to the researcher to correct the questionnaire on the basis of the input. Debriefing is an interview conducted when a respondent has filled the questionnaire. After completing, the respondents are informed that the questionnaire they have just filled was a pre-test questionnaire and are requested to share their views about various dimensions of the questionnaire. They are also requested to find out the problems with the questionnaire, which they realized while filling it.

Pre-testing is done on a small sample collected from the target population. Although there is no magic number that should be sampled during a pre-test, most experts would agree that the sample size should be relatively small—100 or fewer people knowledgeable about the test topic.²⁰ Few researchers have a counter argument on this sample size of 100 respondents for the pre-testing. Ordinarily, the pre-test sample size is small, varying from 15 to 30 respondents for the initial testing, depending on the heterogeneity of the target population.²¹ Largely, determining the sample size is discretion of a researcher. While taking a decision about the sample size for pre-testing, as a thumb rule, a researcher should keep the nature of population diverse. More heterogeneity in population requires relatively large size of the sample compared with the situation when population is relatively homogeneous.

4.3.3.2 Revisiting the Questionnaire Based on the Inputs Obtained from Pre-Testing

To enhance the accuracy, after incorporating suggestions from the pre-testing, a researcher can go for second pre-testing. It is always possible that the second pre-testing may also reveal some of the problems of the questionnaire. Pre-testing of the questionnaire might have provided many inputs in a subjective manner. At this stage, the researcher must objectively incorporate all the inputs obtained from the pre-testing exercise. All the parameters related to the question wording (as discussed in the previous sections) must be carefully considered. Double-barrelled questions, if any, identified in the pre-testing should be either reconstructed or eliminated. Similar treatment is required for leading or loading questions. If pre-testing has identified any overstatement, it must be corrected. Any question having an implicit assumption or generalization must be reconstructed to make the respondents comfortable with the questionnaire. The questionnaire should also be able to match the respondent’s ability to provide answer. Any question overtaxing the respondent’s memory must be either simplified or reconstructed.

Similarly, question sequencing must also be re-examined on the basis of the inputs obtained from the pre-testing. Obtained inputs must be examined in the light of appropriateness of screening questions, opening questions, transition statements, difficult to answer questions, identification, and categorization questions. The input must also be examined in the light of accuracy of question format. Necessary changes, if any, must be made related to the questionnaire appearance and layout.

4.3.3.3 Revised Final Draft of the Questionnaire

At this stage, the researcher once again carefully examines the questionnaire. Insertions and deletions of the previous stage must be re-checked to provide the desired accuracy. The researcher makes the questionnaire “ready to administer” by eliminating all the minute mistakes and tries to make it error free. After careful examination of all the incorporations obtained from pre-testing, the researcher is now ready to have the final draft of the questionnaire and administers to the sample taken from a target population.

4.3.3.4 Administration of the Questionnaire and Obtaining Responses

At this stage, the researchers administer the questionnaire to the respondents and obtain the responses. These responses are coded, data are tabulated, and appropriate statistical techniques are applied to analyse the data. Chapter 9 describes the data preparation and data coding. Chapters 10–18 open the detailed discussion related to data analysis.

Endnotes

1. Brennan, M. (1992). “Threats to Survey Research: Excessive Interviewing and ‘Sugging’”, Marketing Bulletin, 3: 56–62.

2. Malhotra, N. K., J. Agarwal and M. Peterson (1996). “Methodological Issues in Cross-Cultural Marketing Research”, International Marketing Review, 13(5): 7–43.

3. DeLeeuw, E. D. (2001). “Reducing Missing Data in Surveys: An Overview of Methods”, Quality & Quantity, 35: 147–60.

4. Connolly, H. B., J. Corner and S. Bowden (2005). “An Empirical Study of the Impact of Question Structure on Recipient Attitude during Knowledge Sharing”, The Electronic Journal of Knowledge Management, 32(1): 1–10.

5. Swamy, S. (2007). Usability and Internationalization (LNCS), Part II. p. 496. Springer Berlin/Heidelberg.

6. Malhotra, N. K., (2004). Marketing Research: An Applied Orientation, 4th ed., p. 291. Pearson Education.

7. Johnson, J. M., D. N. Bristow and K. C. Schneider (2004). “Did You not Understand the Question or Not? An Investigation of Negatively Worded Questions in Survey Research”, Journal of Applied Business Research, 20(1): 75–86.

8. Bickart B., J. M. Phillips and J. Blair (2006). “The Effects of Discussion and Question Wording on Self and Proxy Reports of Behavioural Frequencies”, Marketing Letters, 17(3): 167–80.

9. Burns, A. C. and R. F. Bush (1999). Marketing Research, 3rd ed., p. 355. Upper Saddle River, NJ: Prentice-Hall.

10. Welch, J. L. and C. O. Swift (1992). “Question Order Effects in Taste Testing of Beverages”, Journal of the Academy of Marketing Science, 20(3): 265–8.

11. Schwarz, N. and H. J. Hippler (1995). “Subsequent Questions may Influence Answers to Preceding Questions in Mail Survey”, Public Opinion Quarterly, 59(1): 93–7.

12. Cauter, T. (1956). “Some Aspects of Classification Data in Market Research”, The Incorporated Statistician, 6(3/4): (133–44).

13. Lagarace, R. and J. Washburn (1995). “An Investigation Into the Effect of Questionnaire Format and Color Variations on Mail Survey Response Rates”, Journal of Technical Writing and Communication, 25(1): 57–70.

14. Boser, J. A. (1990). “Surveying Alumni by Mail: Effect of Booklet/Folder Questionnaire Format and Style of Type on Response Rate”, Research in Higher Education, 31(2): 149–59.

15. Ibid., pp. 86.

16. Frey, J. H. (1991). “The Impact of Cover Design and First Questions on Response Rates for a Mail Survey of Sky Divers”, Leisure Science, 13(1): 67–76.

17. Williams, A. (2003). “How to . . . Write and Analyse a Questionnaire”, Journal of Orthodontics, 30(3): 245–52.

18. Reynolds, N. and A. Diamantopoulos (1996). “The Effect of Pretest Method on Error Detection Rates”, European Journal of Marketing, 32(5/6): 480–98.

19. Blair, J. and S. Presser (1992). “An Experimental Comparison of Alternative Pre-Test Technique: A Note on Preliminary Findings”, Journal of Advertising Research, 32(2): 2–5.

20. Shao, A. T. (2002). Marketing Research: An Aid to Decision Making, 2nd ed., p. 279. South-Western Thomson Learning.

21. Malhotra, Op.Cit, p. 87.

Summary

To conduct any research, questionnaire is an important tool. A questionnaire consists of formalized and pre-specified set of questions designed to get responses from potential respondents.

Designing of questionnaire is a systematic process consisting of three phases: pre-construction phase, construction phase, and post-construction phase. Phase I is the pre-construction phase of the questionnaire design process. It consists of three steps: specify the required information in the light of research objective, an overview of the respondent’s characteristics, and decision regarding selecting an appropriate survey technique. Phase II is the real construction phase of the questionnaire design process. It consists of six steps: decision regarding question format (structured questions vs unstructured questions), decision regarding question relevance and wording, decision regarding question sequencing, decision regarding question response choice, decision regarding questionnaire layout, and producing first draft of the questionnaire. Phase III is the post-construction phase of the questionnaire design process. It consists of four steps: pre-testing of the questionnaire, revisiting the questionnaire based on the inputs obtained from the pre-testing, revising final draft of the questionnaire, and administering the questionnaire and obtaining responses.

Note

1. Prowess (V. 3.1): Centre for Monitoring Indian Economy Pvt. Ltd, Mumbai, accessed September 2009, reprinted with permission.

Key Terms

Categorization questions, 88

Closed-ended questions, 78

Construction phase, 77

Dichotomous questions, 79

Double-barrelled questions, 83

First draft of the questionnaire, 90

Funnel technique, 89

Identification questions, 88

Leading question, 83

Loaded questions, 84

Multiple-choice questions, 80

Open-ended questions, 77

Opening questions, 87

Post-construction phase, 90

Pre-construction phase, 75

Pre-testing of the questionnaire, 90

Questionnaire, 75

Questionnaire layout, 89

Question relevance, 77

Question response choice, 77

Question sequencing, 77

Question wording, 81

Respondent’s characteristics, 76

Scales, 81

Screening questions, 87

Section technique, 89

Split-ballot technique, 84

Structured questions, 78

Transition statements, 88

Unstructured questions, 77

Work technique, 89

Discussion Questions

1. 1. What is a questionnaire and how can we use the questionnaire as a research tool?
2. 2. What are the steps involved in the questionnaire designing process?
3. 3. What is a pre-construction phase of the questionnaire and what are the steps to be considered for launching the pre-construction phase of the questionnaire?
4. 4. What is a construction phase of the questionnaire and what are the steps to be considered to launch the construction phase of the questionnaire?
5. 5. What are structured and unstructured questions?
6. 6. Write short notes on the following terms:
   • Open-ended questions
   • Closed-ended questions
   • Dichotomous questions
   • Multiple-choice questions
   • Scales
7. 7. What precautions a researcher should consider while wording the questions?
8. 8. What are double-barrelled questions and how do they negatively impact the effectiveness of a questionnaire?
9. 9. What are leading and loaded questions? How these bias responses while administering the questionnaire?
10. 10. Under what circumstances a researcher uses split- ballot techniques?
11. 11. What are the decision parameters a researcher should consider when sequencing the questions?
12. 12. What is the importance of screening questions, opening questions, and transition statement in the questionnaire construction?
13. 13. What is the use of identification questions and categorization questions in construction of a questionnaire?
14. 14. What are the approaches to be considered when a researcher takes a decision about logical order of questioning?
15. 15. How a questionnaire layout makes an impact on the response rate. Is the questionnaire layout important to increase the response rate?
16. 16. What is a post-construction phase of the questionnaire and what are the steps to be considered to launch the post-construction phase of the questionnaire?
17. 17. Is pre-testing questionnaire essential for a questionnaire development process? Why there is a general tendency of the researchers to avoid the pre-testing questionnaire?

Case Study

Case 4: Videocon Industries Limited: Opting a Way of Consolidation for Materializing Dreams

Introduction: An Overview of the Consumer Electronics Industry in India

The consumer electronics industry has been witnessing a remarkable growth over the past few years. The fast-growing segments during the year were colour televisions, air conditioners, DVD players, and home theatre systems. Other segments of consumer electronics and home appliances have also shown a positive growth. The consumer electronics and home appliances industry broadly comprises brown goods, white goods, and small domestic appliances.

Brown goods: colour televisions, CD and DVD players, camcorders, still cameras, video game consoles, HIFI, and home cinema;

White goods: air conditioners, refrigerators, dish washers, drying cabinets, microwave ovens, washing machines, freezers, and so on;

Small domestic appliances: iron, vacuum cleaners, water purifiers, and so on.

The company is primarily into manufacturing and distribution of colour televisions, refrigerators, washing machines, air conditioners, microwave ovens, glass shells, and other components.¹

Videocon Group: A Major Player in Consumer Electronics

Shri Nandlal Madhavlal Dhoot was the founder of Videocon Group. In early 1980s, through a technical tie up with Toshiba Corporation of Japan, he produced India’s first world-class colour television: Videocon. Today, Videocon is a household name across the nation—India’s No. 1 brand of consumer electronics and home appliances, trusted by more than 50 million people to improve the quality of life. Videocon Group has a working environment that is driven by performance, strong value base, empowerment, inclusive approach, diversified talent base, and fun filled. This conducive environment is further fostered by creativity and autonomy, equal opportunities, long-term perspective training, and reward. The philosophy of Videocon Group can be well understood from their Vision and Mission statement: “To delight and deliver beyond expectation through ingenious strategy, intrepid entrepreneurship, improved technology, innovative products, insightful marketing and inspired thinking about the future.”² Table 4.01 exhibits sales and profit after tax (in million rupees) of Videocon Industries Limited from 1994–1995 to 2008–2009.

Table 4.01 Sales and profit after tax (in million rupees) of Videocon Industries Limited from 1994–1995 to 2008–2009

Source: Prowess (V. 3.1), Centre for Monitoring Indian Economy Pvt. Ltd, Mumbai, accessed September 2009, reprinted with permission.

Some Challenges Before Consumer Electronics Industry

Like any other industry, consumer electronics industry in India is also facing some problems. Some of the common problems are listed below¹:

• The sharp depreciation of rupee is exerting pressure on the cost of inputs.
• In view of the global slowdown, the consumer sentiments are getting muted as a result of which it is expected that the overall spending may go down and so is the demand for the company’s products.
• There is risk of non-adjustment of product mix in line with the market demand or keep pace with the technological changes.
• There is risk of non-adoption/availability of technology.
• There is risk of inability to keep pace with the changes in product design and features.
• The regulatory environment continues to be uncertain, and changes from time to time can delay the projects.
• Poor government spending on rural and small town electrification program and poor distribution network are major concerns.

Some problems related to the industry can be tackled, but those related to the government policies and other external environment are difficult to handle. For smooth growth, the company has to take care of all the challenges internal and external. The Videocon Group is also aware of all these challenges and ready to counter-attack by adopting some strategic changes. Consolidation of various business activities is one way to cope with the environment challenges.

Videocon Group: Changing the Way through Consolidation

K. R. Kim, the former MD of LG Electronics, was appointed as the vice chairman and CEO of Videocon Industries Limited in 2008. From 2008, Mr K. R. Kim is leading the company in the domestic and global operations. While talking about consolidation issue in 2009, Kim told The Economic Times “The Videocon Group was very fragmented with many brands and each of them run by a separate team. So, the first job was to consolidate and generate efficiencies. Just to give a perspective, we have reduced the number of branches from 160 across all the brands to just 46. Similarly, the number of warehouses is now down to 50 from close to 200.” Kim optimistically added that the key to success lies in changing the organizational culture first and then sales. While talking about the multi-brand strategy of the group, Kim said that a multi-brand strategy has its own advantageous and disadvantageous, but yes, having too many brands can be a problem as it requires marketing investments. Comparing his grand success in LG with Videocon, Kim stated that at Videocon, we are changing the way we did business earlier. As long as we are able to instil the core values of discipline, integrity, and quality, we will be able to meet the targets.³

Videocon Group has realized the problems of having a multi-brand strategy. It is willing to meet the challenges posed on the consumer electronics and home appliance ­business with the discussed consolidation strategy. Suppose Videocon is interested in assessing “brand shift” in its favour as a result of the consolidation strategy then to operate this research programme, it has to address various issues such as should the company be using comparative scaling techniques or non-comparative scaling techniques? If the company decides to use the non-comparative scaling techniques then should it be a continuous scale, Likert scale, semantic differential scale, or staple scale? How will the company decide about the nature and description of the scale? What will be the physical form of the proposed scale? How the company will be able to meet the criteria of reliability, validity, and sensitivity?

Notes

1. 1. Prowess (V. 3.1), Centre for Monitoring Indian Economy Pvt. Ltd, Mumbai, accessed September 2009, reprinted with permission.
2. 2. http://www.videoconworld.com, accessed September 2009.
3. 3. http://economictimes.indiatimes.com/Openion/ Interviews/We-are-changing-the-way-we-did…, accessed September 2009.