Godwin V Ong'anya
Dr Harry O. Ododa
25 February 2009
Discuss the questionnaire as a data collection technique
A questionnaire is a set of systematically structured questions used by a researcher to get needed information from respondents. Questionnaires have been termed differently, including surveys, schedules, indexes/indicators, profiles, studies, opinionnaires, batteries, tests, checklists, scales, inventories, forms, inter alia. They are
…any written instruments that present respondents with a series of questions or statements to which they are to react either by writing out their answers or selecting from among existing answers. (Brown 6)
The questionnaire may be self administered, posted or presented in an interview format. A questionnaire may include check lists, attitude scales, projective techniques, rating scales and a variety of other research methods. As an important research instrument and a tool for data collection, a questionnaire has its main function as measurement (Oppenheim 100). It is the main data collection method in surveys and yield to quantitative data. Also, due to provision for open endedness, the instrument may be used to generate qualitative and exploratory data (Dornyei 101)
Measurement specification will depend on several factors. The nature of the population to be surveyed is the major concern. The kind of survey may be factual or analytical. If factual, then complications are less. If analytical, the survey may be hugely value laden as to reduce accuracy. However, it is possible to objectify the subjectivity by designing more complex research questions. The kind of respondents would also play a big role in determining how the survey will be conducted. For instance, there may be slight differences on how to conduct a survey with chief executives of flourishing multinational corporations, as opposed to the aged in a remote set up. Survey into phenomena that is subject to seasonal fluctuation would also vary with one that does not fluctuate.
Value laden surveys need complex questions so as to reduce biasness. Such deal with subjects like social representation, opinion, attitudes, stereotypes, awareness, brand images, precepts and values. Due to their multifaceted nature, they warrant questions that are equally multifaceted. Responses from such subjects are often influenced by the environment, and hence tentative. Findings are difficult to validate since they reflect the state of the mind at a given time in a specific environment.
Three types of data about a respondent may be accessed by use of questionnaire instrument. Factual questions: These include demographic information, socio-economic status, education, etc.
Behavioural questions: deals with both past and present deeds of the respondent;
Attitudinal questions: comprises of world views. It covers people's opinions, attitudes, beliefs and values.
Elements of a Standard questionnaire
- Title: this identifies the domain of the investigation. The respondent is initially oriented to the investigation. It should be captivating enough to attract attention and enthusiasm.
- General introduction: this has a description of the purpose of study as well as the organisation(s) involved. The respondent is assured of anonymity/confidentiality of information volunteered, making clear that there are no wrong or right answers. Honest answers are also requested.
- Specific instructions: this offers succinct demonstration on how to carry on with the business of responding to the questionnaire.
- Questionnaire items: is the main part of the questionnaire schedule, to be clearly separated from the aforementioned parts.
- Additional information: includes the full contact information of the researcher/ administrator. May include a promise that a copy of the summary of the final report would be send to the respondent on request.
- "Thank you" may end the questionnaire.
Before coming up with a questionnaire, a researcher has to come up with a precise operational statement on the variables. The instruments to be used have to be well identified and variables well defined. Several considerations have to be made before designing the questions. These considerations are discussed below.
- MAIN METHOD OF DATA COLLECTION
The case here is the consideration of a questionnaire. This includes a standardized formal interview, the postal, self-administered questionnaire and the group administered questionnaire. Each method has its strengths and weaknesses, and this should help in coming up with an appropriate one to suit a specific survey need.
Mail questionnaires and standardized interviews
A mail questionnaire is one that has been sent to a respondent by a researcher, and the respondent would answer at his own time, basing on his own understanding. It is not as controlled as with the case of a standardized interview. Mail questionnaires however have their own advantages.
- The researcher incurs low cost of collecting data. Just designing a questionnaire and sending it to a respondent;
- Analysing and processing the data is less expensive in terms of both time and material resources;
- There is no likelihood of interviewer bias as the interviewee would be interpreting the questions his own way;
- The questionnaire can be handled by a geographically distant correspondent.
A number of disadvantages abound.
- There is no follow-up mechanism;
- No control n how questions are being answered. Questions may be passed on to another person;
- Response rate is low;
- Some questions may be left unanswered;
- Not suitable for a low literacy society, the old, of small children;
- Not suitable for the visually impaired persons;
- No data (ratings or assessments) based on observation.
As opposed to mail questionnaires, interview schedules have a higher response rate and provide an opportunity for both the interviewer and the interviewee clarifying their points. Follow-ups are possible as ratings and assessments can be taken from observation. There may be no problem with disabled or less literate respondents. The major disadvantage is the huge time and material resources need for the purpose.
Self administered questionnaires are presented to respondents, but the researcher is available to make little clarifications. He does not, however, interpret the questions for the respondents as this may increase interviewer bias. There is some degree of personal contact as the two parties would be interacting. The advantage with this is that response rate is high and clarification are made where need arises.
This is presented to respondents in a group. The group may be one of students or labourers, where they are supposed to respond as a group. These questionnaires will be administered by the researcher, and may be in form of a film being shown, and then the respondents are asked questions to respond. Is such cases, the size and literacy of the group is a vital consideration. Each respondent may be required to respond, and in case of a film, it may only make the responses flow. The respondents will be answering by following a certain pattern as the film progresses. Questions may be read out aloud as respondents answer on their own.
- APPROACH TO RESPONDENTS
It would be better if all questionnaires are returned, but more often this is not the case. A researcher must come up with a formula that would enable him get all questionnaires back whenever possible. There are various approaches to this.
- An advance warning to the respondent informing him of the study;
- Explain to the respondent the method of sampling used, and how this respondent came to be selected;
- The aspect of sponsorship is also important. This is where the researcher presents an identification card and a pamphlet from the research organisation. Alternatively, he may have a cover letter or one of introduction addressed to the respondent from an influential person of the firm;
- The envelope used should be as official as possible, addressed to an individual to be supplied with the questionnaire. It should have some personal touch, as opposed to being so mechanical.
- Create awareness of the project through publicity, where possible;
- Provide incentives, though not to the extent of causing bias of any kind. This incentive would influence the respondent to fill the questionnaire with objectivity and return it;
- Confidentiality should be assured in certain cases. Only the researcher should access all survey data. In case there is need for information about identifiable persons to be published, is should happen with express permission of the of the person in point;
- Reminders may serve to increase the response rates, as some respondents may not complete the questionnaire out of mere forgetting;
- Anonymity is another way of increasing the response rate. In such a case, the personal identification details of a respondent are not taken. The respondent may instead be identified by a code number. Alternatively, the respondent may be assured that at the data processing stage, all identifying information will be destroyed.
- The general appearance of the envelope and the letter would also affect the way the recipient would attend to it. The envelope should be addressed to the respondent personally, and with a stamp, not a franked one. The appearance of the letter should be conservative, and the topic should be of some interest to the respondent. It should not be wordy as this would be boring. The letter should be sent with a return envelope.
- THE BUILD-UP OF QUESTION MODULES
The researcher should be careful in the selection of questions to appear at the start of the questionnaire. The instrument should not start with sensitive information that may put the respondent off before he kicks off in response. The module should be arranged in a systematic way so as to flow as the respondent moves down, with the opening questions maintaining neutrality.
- ORDER OF QUESTIONS WITHIN MODULES
The funnel approach has been fronted as a good example of questionnaire design. When preceded by filter questions, the approach has been widely used, though others sequences may also be considered. The funnel method would start by asking a broad question, then narrows down to specifics. A filter would then exclude some questions from respondents, if a response would have been noted through an answer to a different question (Seymour and Bradburn).
- QUESTION TYPE
There are two main types of questions, open and closed. Open questions are those that afford the respondent an opportunity to speak his mind. It is more detailed with little, if any, interviewer bias. Open response may allow graphic examples, illustrative quotes, and generally provide some unanticipated data. We may fall short of imagination in the range of possible response categories thus the strength of open-ended questions. However, analysis of this data may be difficult as varied responses would have been presented. It is not advisable to have open-ended questions in a professional questionnaire (Dornyei 105). Closed questions are easier to answer, process and analyse. They are mostly 'yes' or 'no' answers, without provision for the respondent giving his explanation (Oppenheim 112).
Open ended questions
These should be guided, however open. They may be of four kinds:
- Specific open questions: these ask questions anticipating factual responses. It may include the respondent's preferences, past activities, his details, inter alia.
- Clarification questions: seek further elucidation after selection of a category. It may be in form of "Please specify."
- Sentence completion: a respondent is asked to complete a sentence. "Democracy is practised well, however.…"
- Short answer questions: comprises more than a sentence, but less than a paragraph.
Types of close-ended items
- Likert scales: this has a statement of which a respondent is expected to rate, for instance from best to worst; or from "strongly agree," to "strongly disagree."
- Semantic differential scales: with this technique, a respondent marks in a continuum between two line adjectives or extreme ends. Summation and averaging s the same as in Likert scales. It is important not to have all the positive poles on one side. Alternating avoids position response.
- Numerical rating scales: involves assigning marks in an attempt to describe a feature of the target. We may calibrate a continuum from excellent to poor; always to never, inter alia. Due to its similarity to semantic differential scales, the two may be interchangeable.
- True-false items,
- Multiple choice,
- Rank order items: involve a list of which the respondent is asked to rank.
Strengths and weaknesses
- Questionnaires are highly versatile. They can be used by a variety of people, in different environments, at different times, targeting a variety of topics for analysis;
- Questionnaires are easy to construct and convenience to work with. May provide for anonymity to respondents, should need arise;
- They are cheap to undertake, both in terms of time and material resources;
- Data collected is easy to process as the questionnaire is straight forward;
- Produces superficial data, touching non comprehensively to specificity- it not good for qualitative data;
- Its investigations are limited owing to the short time a questionnaire is attended to (30 minutes to 1 hour);
- May not be suitable to an illiterate respondent.
Brown, J. D. Using Surveys in Language Programs. Cambridge: Cambridge University Press, 2001.
Dornyei, Zoltan. Research Methods in Applied Linguistics: Quantitative, Qualitative and Mixed Methodologies. 1st Edition. Oxford: Oxford University Press, 2007.
Oppenheim, A. N. Questionnaire Design, Interviewing and Attitude Measurement. London: Pinter Publishers Limited, 1992.
Seymour, Sudman and Norman M. Bradburn. Asking Questions. San Francisco: Jossey-Bass, 1983.