Statistics - Data collection - Questionaire Designing
A questionnaire is a form containing a set of questions, which are filled by the respondents. According to Goode Hatt,
"In general, the questionnaire refers to a device for securing answers to questions by using a form which the respondent fills in himself."
The objective of a questionnaire is twofold:
To collect information from respondent scattered in a wide area.
To achieve success in collecting reliable and dependable information in a short span of time.
Designing of questionnaire is an art rather than a science. It is an effort of compiling a set of questions by repeatedly checking out for errors, of learning what to avoid and what to include. However a basic design strategy can be e developed. The designing of questionnaire general1y goes through three phases
Developing a design strategy
Constructing the questionnaire
Drafting and refining the questionnaire.
Phase I: Developing A Design, Strategy
Specify the Information Sought - The researcher should be able to specify the list of information needs. Generally this task has already been accomplished when the research proposal or the research design was developed. The hypothesis stated earlier is the guiding light in stating the information requirement. The hypothesis establishes the relationship between the variables and the researcher can ideally develop the data that is required to he collected to prove or disapprove the hypothesis.
Determine the Communication Approach - It refers to the decision on the method used to conduct the survey i.e. personal interview, depth interview, telephone, mail. computer etc. This decision on method to be used will have a bearing on the type of questionnaire to design. The choice of communication approach is influenced by factors like to location of respondents, the time funds available, nature of study etc. The communication approach chosen results in different introductions, different instructions, layout etc. Once the communication approach has been finalized, a decision is then taken on the type of questionnaire that is to be framed.
Type of Questionnaire - In this step the researcher specifies how the data will be gathered by stating the type of questionnaire required. The questionnaire can be of four types.
Structured-Undisguised Questionnaire - The most popular type, it involves using questions with clear direct wording, having a logical order. The wording and order remains the same for all the respondents. They are very simple to administer and easy to tabulate.
Unstructured-Disguised Questionnaire - The exact opposite of the earlier type, this questionnaire hides the purpose of research and shows no clear order or tendencies. Such a questionnaire generally uses projective methods to collect data. A disguised or hidden stimulus is given to the respondent and the response is in an unstructured form.
Unstructured-Undisguised Questionnaire - In this type of questionnaire the purpose of the study is clear but the questions are generally open ended. e.g. "How do you feel about putting a ban on student union election?" The respondents are free to reply in an unstructured manner. These questionnaires are generally used in depth interviews.
Structured-Disguised Questionnaire. - The purpose of this questionnaire is to hide the motive of study but allow for ease in coding and analysis. This approach is based on the fact that direct questions may influence or bias the replies, but if the questions. are disguised than we ask the respondents What they know and not what they feel e.g. the earlier question will be framed as
What is the effect of student union election?
(a) It creates awareness
(b) It disrupts studies
Although such questionnaire offer ease of tabulation and analysis, yet because of the effort involved in framing disguised questions, this is not a every popular method.
Phase II: Constructing the Questionnaire
Determine Question Content - This step initiates the task of framing specific questions which would yield the data required for study. While framing the questions certain things should be kept in mind:
Is the question necessary? Every question should have some use in providing additional and genuine information.
Is the question complete? The question should have the proper scope to reveal all the information that a researcher needs to know.
Is a single question or multiple questions required? There should not be 'double barreled questions' which combine two questions in one e.g. 'Are the elections this year transparent and according to election commission guidelines'. This is an incorrect method. Instead to obtain the desired information the following two questions should have been asked :
Are the elections this year transparent?
Where the election commission guidelines adhered to completely.
Can the respondent articulate? The respondent may be unable to answer adequately due to his inability to organize his thoughts.
Is the respondent informed? The respondent's information level should be kept in mind i.e. the content of the question should match the knowledge level of the respondent.
Can the respondent remember? The questions should not overtax the respondent s recall ability. No assumption should he made regarding the memory. Take a simple test and answer these questions:
What was the last movie you saw?
Where did you last eat out?
When did you visit a temple?
These questions, although very simple, yet test your recall ability.
Is the respondent willing to answer? This is of relevance in situation where the questions are sensitive exploring an individual's faith, money matters, family life etc.
Determine the Response StrategyOnce the content of questions has been decided upon, the next stage is to decide on the structured response strategy. (close response using fixed alternative questions) or an unstructured response strategy open response using open ended questions). Some of the response strategies are:
Do you own a digital camera?
Which brand do you prefer for buy a digital camera?
What features do you look for in your digital camera?
Video recording facility
Smart physical looks
Free service for 1 year
Large memory capacity
Scale questions may be of the rating or ranking type
Which of the following media has been instrumental in influencing your choice of digital camera? Rank them giving 1 to the most effective media, 2 to the next most and so on.
– Television – Company Brochures – Newspapers – Net Advertising – FM Radio
Of the following factors, rate each factor in terms of the 'most desired', 'some what desired' and 'least desired' that digital camera should possess.
Most Desired Some What Desired Least Desired High Picture Clarity Large Memory Big Screen Zooming Function Picture Editing Function Small Size Video Recording Facility
Determine the Question's Wording - This state is concerned with the phrasing of each question. The researcher needs to use utmost caution in framing the question since a poorly phrased question will either lead to a wrong response of result in refusal to answer. While wording a question the following things should be kept in mind:
Use simple words - The questionnaire is not a test of respondent's vocabulary, hence the words should be simple e.g. 'where, in your opinion, does the forte of this organization lie? Since some may not know the meaning of word 'forte', hence it is better if it is phrased simply as 'Where in your opinion, does the strength of this organizationally?'
Avoid technical jargon - Use of technical words may render even the highly educated respondents helpless in answering the questions.
Avoid using ambiguous questions - words like 'occasionally', 'often', 'sometimes', 'you' etc. are all problem words e.g.
How often do you watch movies in a theatre?
This question is worthless since everybody has a different interpretation of occasionally and sometimes.
Avoid biased wording - Questions that lead respondent towards an answer, give him a clue to the questions, are biased or leading questions. Such questions should be avoided since they distort the intent of the question e.g.
'Do you think government is doing right by allowing FDI in retail sector?'
'Do you consider it appropriate to use low cost accessories for your car?'
These questions force a respondent to think in a particular direction.
The level of personalization should be controlled e.g. 'what should our government do today'?
Increase defence expenditure, even if it means more taxes. The alternative can also be written as
Increase defence expenditure even if you have to pay more tax.
The second alternative is more personalized and both these alternatives will produce different results. There is no fixed rule to say which method is to be used. However one should choose that level of personalization which presents the issue more realistically.
Phase III: Drafting and Refining the Questionnaire
Decide On Question Sequence - From this step, we enter the stage of drafting the questionnaire and the ordering of questions is an important aspect. The following things need to be kept in mind:
Use simple and interesting Question first. It makes the respondent feel comfortable. An alternative is to use 'funnel approach' whereby broad questions are asked first and specific questions are asked later on.
The questions should be arranged in a logical order. Jumping from topic to topic would break the flow of respondent and he would lose interest in filling the questionnaire.
Classification questions should be asked later on. Classification questionnaire the personal questions. The reason for asking classification question before target questions is to avoid alienating the respondent before getting to the heart of study e.g. respondents who readily offer their opinion on preferred car types would most likely balk from responding with same freedom if their income is asked first.
Difficult and sensitive questions should be asked right in the beginning of the questionnaire so as to avoid threatening the respondent.
Branching of questions should be done with care. Branching refers to directing the respondent where to go next in the questionnaire on the basis of his answer to preceding questions. Branching is easier in case of telephone or personal interview schedule but when the questionnaire is to be mailed; branching should be avoided as it may confuse the respondent.
Determine the Physical Characteristics - The physical appearance affects the way the respondents react to the questionnaire. Hence the following points should be observed:
Use a good quality paper with high definition ink so that it can be read easily. The questionnaire should look professional and easy to answer.
Size of the questionnaire is important. Small questionnaire is preferred to a lengthy one provided the small size is not achieved at the expense of appearance. If a small size is achieved by making the questionnaire crowded then it will lead to errors and result in less informative answers.
A questionnaire should be accompanied by an introductory letter. It should introduce the study and male the respondent realizes the worth of his powers. The importance of research and the importance of respondent's replies should be conveyed through the letter.
The instructions should be written clearly and politely. The method of reporting the responses i.e. tick mark, cross or circle should stated clearly. If the respondent is to skip certain questions then 'go to' instructions should be used and if an entire section is to be skipped then different colors for different sections should be used.
Pre-Testing the questionnaires - The process of using a questionnaire on a trial basis on a small group of respondents to determine how the questionnaire is performing is termed as pre-testing. Pre-testing helps to find out the errors in individual questions along with the sequence of questions. Various aspects like the ability of questionnaire to generate respondent's interest, the interpretation of the meaning of questions, the continuity of the questions, the time required to fill the questionnaire can be tested through pre-testing, Pretesting can be classified as follows:
Researcher Pre-testing - This involves a small group of fellow researchers testing the instrument in the initial stages of design. These researchers can provide valuable input to improving the questionnaire.
Respondent Pre-testing - In this, pre-testing is done by a small sample drawn from the target respondent population. The instruments in its ready form are administered to the sample and then inputs can he taken from them on the questionnaire.
Collaborative Pre-testing - When the respondents are told about the role in pre-testing then it becomes a collaborative pre-test. In such a pre-testing detailed probing of each and every question is done and it is usually a time consuming process.
Non-Collaborative Pre-test - When the respondents are not told about their role in pre-testing becomes non-collaborative pre-test. Since they are not being told, their cooperation will be relatively less comprehensive. However this approach has the advantage that it is conducted in situation exactly similar to the real environment.