Collection of data through questionnaire

For construction of a questionnaire it is important to understand five basic rules of questionnaire design:

1. Understand the goals of project so that only relevant questions are included.
2. Questions should be clear and unambiguous.
3. Questions must accurately communicate what is required from the respondents.
4. Don’t assume respondents understand the questions they are asked.

Types of questions: survey can consist of two basic types of questions:

1. Open-ended
2. Closed-ended

Open-ended question: It requires respondents to generate their own answers. It gives the respondents freedom in answering questions and an opportunity to provide in-depth responses.

Example:
Why do you watch Colors channel?
What type of TV program do you prefer to watch?
What would your favourite radio station change so that you would listen more often?

Advantages:

1. The researcher gets insight in respondent’s understanding.
2. Good to construct for complex issues when the total answer categories are very large (say, 20 or more)
3. Sometimes, the information and responses are so unexpected that the researcher’s ideas are completely changed.

Disadvantages:

1. Sometimes responses are irrelevant.
2. It is difficult to classify and code all responses.
3. As the data are not standardized, statistical analysis and computation of percentages become difficult.
4. Sometimes, responses are very lengthy and analysis of them is time consuming.
5. Semi-literate respondents find it difficult to answer open questions.
6. Open-ended questions may encounter a high refusal rate.

Closed-ended questions: Are the fixed choice-questions. They require the respondents to choose a response from a set of responses provided by the researcher.

Example:

Which of the following is most persuasive medium to disseminate information on Health Communication?

Interpersonal communication 
Newspaper/magazine/pamphlets
TV 
Radio 
Internet 
Books 
Other__________________

Advantages:
1. They provide a greater uniformity of responses.
2. The responses can be easily quantified.
3. Response rate is high particularly in sensitive questions like income, age and so on. If the answer in closed-ended question is a category, the respondent may easily identify himself with the range in which his income/age falls.

Some other types of questions:

Direct
Indirect
Nominal
Ordinal
Interval
Contingency
Filter

Direct: Direct questions are personal questions which elicit information about the respondent himself/herself.

Example: Do you believe in media censorship?

Indirect: Indirect questions seek information about other people.

Example: Do you think people of your status and age believe in media censorship?

Nominal: When its response falls in two or more categories.

Example: Rich/Poor, Married/Muslim, rural/urban, Shia/Suni and so on.

Ordinal: Questions in which responses are placed in rank and order of categories.

Example:

Smoking: regularly/occasionally/never
Relations with class mates: Excellent/satisfactory/dissatisfactory/can’t say

Interval: Questions which contain a range or interval.

Example:

Income per annum: Below 18,000/ 18,000-36,000/36,000-54,000/54,000-72,000/Above 72, 000.

Contingency: A contingency question is one whose relevance to the respondent is determined by his response to an earlier screening question.

Example:

Q1. Are you in favour of using some method in controlling birth?
Q2. Do you prefer vasectomy/condom/pill/safe period?

The second question is a contingency question.

The need for the contingency question arises because every question need not be relevant to all respondents. The use of contingency question can be reduced by drawing a homogeneous sample. The preferable format for contingency question would be as follows:

Q: Do you go to cinema houses for watching movies?

(a) Yes
(b) No

If yes, how often do you go? (a) once in a month; (b) once in a few months; (c) once or twice in a year.

Filter: These questions aim at eliciting information related to a general aspect of the research topics and are usually followed by more specific question.

Example:

Do you smoke?-Filter question.
Do you (being a girl) smoke?-Contingency question.

Dos and Don’ts of writing questions:

1. Questions should be clear: Sometimes which is perfectly clear to the researcher is not clear to the respondent.

Example: After finding out which radio station a respondent has been listening to more lately, the researcher might ask:

Why have you been listening to Radio Mirchi more lately?

And many be expecting to receive the response such as “I like the music a lot more.”

But the respondent might say: It’s the only station my radio can pick up.

The question would be much clearer to a respondent if asked in this form: Which radio station you enjoy listening to more lately as compared to a few months ago?

Making questions clear also requires:

• Avoiding difficult or specialized words, acronyms and so on.
• Avoiding double or hidden meanings in words.

Avoiding difficult or specialized words, acronyms and so on: Questions should be phrased in everyday speech and, and social science jargons and technical words should be eliminated. Example:

Do you believe demassification of audience is essential in the era of cut-throat competition?

The term ‘demassification’ may not be understood by the lay man. It has to be spelt out.

Avoiding double or hidden meanings in words: Questions should be written so they are fair to all types of respondents.

Example:

How many TV shows do you think are a little too violent---Most/some/few/none.

Some respondents who feel that all TV shows are extremely violent will answer ‘none’

The appropriate question might be:

How many TV shows, if any, do you think are too violent---Most/some/few/none.

2. Questions should be kept short: Respondents who are in a hurry are unlikely to take the time to fill a questionnaire with long questions.

3. Remember the purpose of research: It is important to include in a questionnaire only items that relate directly to what is being studied. For example:

If the occupational level of the respondent is not relevant to the purpose of the survey, the questionnaire should not ask about it.

4. Do not ask double-barreled questions: A double-barrel question is that question which asks two or more questions in the same sentence. Whenever the word ‘and’ appears in the question, the sentence should be re-examined to check if more than one question is being asked.

Example: India TV has programs which are funny and misleading. Do you agree or not?

A program may be funny buy may not be misleading. The question may be split into two parts.

5. Avoid biased words or terms:

Example: Where did you hear the news on President’s new economic policy?

The question is mildly biased against newspapers as the word ’hear’ suggests ‘radio’, ‘TV’ or ‘other people’ is a more appropriate answer.


6. Avoid leading questions: A leading question is one that suggests a certain response or contains a hidden meaning.

Example: Like most Indians, do you believe in God?

The question suggtests that the respondent should answer in affirmative or run the risk of being unlike most Indians.

7. Do not use questions that ask for highly detailed information:

Example: In the past 30 days, how many hours of TV have you viewed with your family?
This is unrealistic. Few respondents could answer it. A more realistic approach is to ask:

How many hours did you spend watching TV with your family yesterday?

8. Avoid potentially embarrassing questions unless they are absolutely necessary: An overly personal question may cause embarrassment and inhibit respondents from answering honestly—asking people’s income for example.

Example: What is your annual income?
The question can be put as: Which of the categories includes your household’s total annual income?

Rs. 25,000-Rs.40, 000; Rs. 40,000-Rs. 55,000; Rs. 55,000-Rs. 70, 000; Rs. 70,000-Rs. 85,000.

Format of the questionnaire/schedule:

Questionnaire/schedule format refers to the general model which provides guidelines on how the questions should be placed in a sequence and in a logical order of relationship to each other, what type of questions should be considered, how long the questionnaire/schedule should be, and how the questionnaire/schedule should be presented so that it is clear and easy to understand. The important aspects of questionnaire are given below:

1. Instruction
2. Questionnaire length:
3. Question order
4. Layout

Instruction: All instructions necessary to complete the questionnaire should be clearly stated for respondents or interviewers.

1. Mail questionnaire require the most specific instructions because respondents are not able to ask questions about the survey.
2. Respondents must understand whether the correct response consists of circling or ticking an item or placing items in a specific order.
3. Procedural instructions for respondents are often highlighted with a different typeface, capital letters, or some graphic device, perhaps arrows or lines. Example:

Do you have a favorite radio station that you listen to most of the time?
_________ Yes ________No

If yes, please briefly explain why on the lines below:

_________________________
_________________________
_________________________

4. Some questionnaires require respondents to rank a list of items. In this case, the instructions must clearly describe which response represents the highest value.

Example: Please rate the following magazines is order or importance to you. Place a 1 next to the magazine you prefer most, a 2 next to the magazine in the second place, and so on up to 5.
__________ India Today
__________ The Outlook
__________ The Week
__________ The frontline
__________ The mainstream

5. Filter or screener questions are used to eliminate unwanted respondent or to include only respondents who have specific characteristics or who answer questions in a specific manner. Example:

In a typical week, do you listen to radio station on the AM dial?
_____Yes (Respond Q. 16)
_____No (Skip to Q. 17)

Questionnaire length: Long questionnaires cause fatigue and low completion rates. The length of a questionnaire depends on a variety of factors:
· What the researcher wants to know and how many items are necessary so that the data will be credible
· Type of study (Mailed questionnaires are shorter than face-to-face interviews)
· Type of problems or questions to be investigated
· Budget of the research
· Age of respondent (young people will be available for less time than middle-aged or old-people )

Question order:

· Researchers often include one or two preliminary or ‘warm up’ questions about the topic under investigation so respondents become accustomed to answering questions and begin thinking about the survey and.
· Preliminary questions should create interest in respondents.
· Personal questions, sensitive information, demographic data should be placed at the end of the questionnaire
· Some respondents may still refuse to answer personal items; at least the main body of data is already collected.
· Age and gender information are usually included in the first part of a questionnaire, so at least some respondent identification is possible.
· The questionnaire should be organized in a logical sequence, proceedings from general to the specific. Questions from similar topics should be grouped together, and the transitions between question sections should be clear and logical. Poor question order may bias a respondent’s answer. Example:

After several questions about the presence of violence in society, the respondent is asked to rank the major problems facing the community today from the following:
________ Naxalism
________ Corrupt Govt.
________ High prices
________Violence on TV
________War

The ‘Violence on TV’ might receive a higher rank than it would.

There is no easy solution to the problem of question ‘contaminatin’. Obviously some questions have to be asked before others. Perhaps the best approach of researchers is:

To be sensitive to the problem and check for it in the pretest
If they think question order A,B,C may have biasing effects, they should test another version using the order C,B,A.
Complete neutral ordering is not possible. In case of suspect of bias, the order of questions may be rotated.
Questionnaires with different question orders may be printed but make sure that the data are input and analyzed correctly.

Layout:

The physical design of the questionnaire is another important factor in survey research. A few points are worth noting:

1. A badly typed, poorly reproduced questionnaire is not likely to attract many responses in mail survey.
2. One should avoid cramped questionnaire say 40 questions to a page does not create a positive attitude towards it.
3. Response categories should be adequately spaced and presented in a non-confusing manner. For example:

There are too many commercials on TV.
Do you strongly agree______ Agree_____ Have no opinion_____ Disagree_____ Strongly disagree_____

The above format might lead to problems.

A more effective and less confusing method is to provide a vertical ordering of the response choices. Example:

There are too many commercials on TV.
__________ Strongly agree
__________Agree
__________ No opinion
__________ Disagree
__________ Strongly disagree

Sometimes respondents make large check marks or Xs that cover more than one blank, making interpretation difficult. If blanks are perceived as problems alternative methods (boxes, circling) should be employed. However, the response form should be consistent throughout.
Questions, specially the open-ended ones should have enough space for answers; else it is very discouraging to the respondents and will hamper the effectiveness of the survey.
Example
Why do you go to movies? __________________________________________.

The space given above is insufficient.

If the research budget limits the amount of the paper for questionnaires, respondents can be asked to add further comments on the back of the survey.