Some types of research do not involve numbers or analysis of past performance but, rather, finding out people’s opinions, feelings, likes and dislikes, and the motivations behind their buying behaviour – in other words, the psychology underlying buying decisions. Research into these areas is known as qualitative research, and it is done in very different ways from quantitative research.
Qualitative research becomes increasingly important as we try to predict further into the future, as the further we try to project, the less reliable are historic results and the more important people’s opinions become.There are problems associated with qualitative research, One is the fact that people often find it very difficult to explain their behaviour or their motivations. Another problem is that many people are reluctant to tell a researcher what makes them do what they do. Sometimes interviewees will try to guess the answers the interviewer is looking for and answer and on other occasions respondents may give answers that they believe are true but are simple versions of the true reasons.
Panel (or jury) methodA wider range of experts drawn from the field in question may make up a panel to discuss the variety of opinions put forward by the group and to come to a decision. This can be time-consuming on occasion, and is arguably only a little more accurate than the executive judgement method. The Delphi method may be used to improve forecasts further.Delphi methodThe Delphi method is not dissimilar from the panel or jury method, but attempts to remove the possibility that each individual’s views may be swayed by other members of the panel and by group pressure. Panel members do not meet but each one makes predictions in isolation and these are then examined to determine results. The method works as follows.1.
Each member of the panel gives his or her opinions separately and without consulting other panel members.2. A nominated person gathers together all of the opinions. At this stage any extreme views will be discarded, and an overall impression of a possible agreement will be established.3.
The initial agreement is distributed to each panel member for further thought. The agreement will be amended at this stage as a result of the comments received.4. The new agreement is sent back to each member for further comments. This process will be repeated as many times as necessary.
Use of surveysThe methods discussed above rely on ‘expert’ opinions and those of people in the trade. It ought to be more revealing to ask customers themselves, and a survey will facilitate this.Data collection using questionnairesA questionnaire is a systematic list of questions designed to obtain information from individuals concerning:* specific events* their attitudes* their values* their beliefs.A good questionnaire will result in a smooth interview, giving the interviewer a precise format to follow and ensuring that he or she obtains exactly the information required in a format that is easy for the researcher to analyse later.Effective design of the questionnaire is vital if it is to deliver the quality of information that firms require. Badly worded and poorly constructed questionnaires can mislead or confuse respondents, or may influence the respondent into giving certain replies. If questions are badly worded and do not stimulate the respondent sufficiently, important information may not be gained. If the questions are too wide-ranging, the variety of answers received may make them difficult to interpret.
Questions that require personal information may make the respondent feel that they show him or her in a bad light and consequently the answers given may not be true, so techniques must be used to attempt to establish the validity of such responses. If questions are too complicated, respondents will quickly be put off and may not reply at all.The following general points should be borne in mind when compiling questionnaires.
* Before writing any questions you should first clarify the following in your own mind – better still write the answers down and keep them in front of you as you prepare your questionnaire, to lend your questionnaire focus.1. The exact information you require and why you need it.
All factors that may be relevant should be noted so that you can ensure the questions you write cover all important areas.2. The target population you wish to examine.
Make sure that this sample is representative and that you question all of the people who are likely to have relevant opinions or information. Remember that your survey findings will not be valid if your sample is not truly representative of the population.3. Whether the questionnaire will be completed independently by the respondent. If so, the language and number of prompts and examples included need to considered.
Make it as easy as possible to complete without help.* Always start with a statement of who you are, which organisation you represent and the purpose of your research.* Use appropriate language – that which is likely to be familiar to respondents. Take care with technical phrases which may only be understood by a few people and avoid using jargon.* Questions need to be precise and clearly stated, so that there is no room for ambiguity or misinterpretation.
Questions such as ‘Do you use the Internet frequently?’ are open to misinterpretation. The respondent may consider once a fortnight to be frequent usage whereas you may feel that this is fairly infrequent and daily usage would be a better description of frequent. Responses to such questions may lead to misleading results.
Your question therefore should be more precise, such as ‘How frequently do you use the Internet?’ This might be followed by a variety of possible responses inviting the respondent to tick the most appropriate. If you are designing a questionnaire that is to be read out by an interviewer, make sure that the questions are sufficiently precise – vague questions may require the interviewer to explain or elaborate further, and this may introduce bias by leading the respondent.* Think carefully and critically about each question to decide if each is necessary to your research. if it is not, do not include it.* Keep the questionnaire as short as possible. Preferably there should be no more than forty questions as more can put off respondents or cause them to give hasty replies to later questions.
Long questionnaires are therefore likely to yield a certain amount of inaccurate data, particularly in the later questions. One way of solving this problem is to use ‘branched’ questions. This uses the responses to early questions to guide respondents so that they only answer later questions that are directly relevant to them. Such a technique would include questions such as ‘Do You own a video recorder? If Yes go to Question 6, if no 90 straight to Question 11., This technique can allow you to include more questions in the questionnaire without requiring every respondent to answer all of them.* Each question should not be too long.
If questions become lengthy and involved, they should be broken down into a series of questions.* Ensure that no bias creeps into your questions that may lead your respondents to answer in ways that they feel are more acceptable. Be careful not to put ideas into their heads by quoting examples – often having been given examples, respondents cannot think of any others. Questions such as ‘In your opinion, which record store (e.g. Virgin or HMV) Offers the best customer service?’ can distort your results, as an unrepresentative number of people are likely to respond ‘Virgin’ or ‘HMV’. Similarly, a question such as ‘Would you agree that the service in this store is good?’ may lead the respondent to agree and is therefore not a valid question.
* Try to avoid personal questions that may reveal the respondent in what they would consider to be a poor light. Respondents may give the answer that makes them look best, rather than the truth.* Be careful not to combine two questions into one, as this may confuse the respondent. A question such as ‘What is your opinion of Kellogg’s Corn Flakes and other Kellogg’s products?’ could leave your respondent feeling bemused. They may not like the Corn Flakes but may enjoy other Kellogg’s products, so will find it difficult to answer the question properly.
It would be much better to split the question into two.* Ensure that the grammar in your questionnaire is correct – you can soon lose the respect of a respondent with poor spelling, punctuation or grammar.* Provide example questions at the start of the questionnaire to illustrate how they should be answered. If you have different types of question, each type should have an example for the respondent to follow.Be careful with open questions, as they can lead to so many different answers that they become difficult to analyse. A question such as ‘What types of music do you listen to most often?’ could lead to such a variety of answers that analysis becomes time consuming.
You would do better to offer a comprehensive list of alternative answers that may be ticked, as this will make analysis much more straightforward. A better question might be:Please indicate with a tick which types of music you listen to regularly (tick all that apply)Classical Easy listening Jazz BluesGolden oldies Popular Heavy metal PunkIndie Rap Dance SwingHip hop Other (please specify)..
Conclude by confirming whether your respondent will be prepared to answer further questions if necessary. Bear in mind that you may want to ask more questions as a result of your initial questionnaire, and having a ready source of potential respondents may be very useful.Types of questionnaire* Structured questionnaires list questions in a logical order and specify particular answers that may be given. They give the interviewer little or no discretion concerning the order of the questions or how they are to be posed.
They often contain mostly questions with pre-coded answers, and not many open-ended questions. This makes them particularly suitable to very large surveys as the results are easier to quantify. However, they are criticised for not providing in-depth responses.* Unstructured questionnaires contain principally open-ended questions and allow the respondents to reply in whatever way they choose, with answers as long as they wish. This method will deliver in-depth responses, but will take much longer to analyse. It is probably more suited to smaller surveys.* Checklists are used in unstructured situations where the information is gathered by discussion rather than by posing specific questions. The checklist acts as a reminder to the interviewer of areas that need to be discussed.
Only experienced interviewers should use the technique.* Self-completed questionnaires are posted or delivered directly to the respondent and left with him or her for completion. They are either posted back after completion or collected by the interviewer.* Interviewer-completed questionnaires are used to structure either face-to-face or telephone interviews. The replies may be recorded on paper, but increasingly the answers will be input straight into a computer.Good questionnaire layoutPutting some thought into layout will make a questionnaire easier for both the interviewer and the respondent. Make sure that the questions follow in a logical order, leading from one to another and not jumping from topic to topic, and that all the questions on a particular subject are grouped together. Start with general questions that establish an overview of the respondent’s opinions or preferences, and then gradually work down to the specifics of particular areas or products.
Easier questions should come at the beginning, and harder ones should come later, once the respondent feels comfortable. The principle is to put the respondent at ease early on, and you will then receive more in-depth replies to trickier questions later.The questionnaire should be laid out attractively – it should not be cramped up with tiny print, but questions should be spread out across the page with good use of space. This is especially important in a self-completion questionnaire as, if it looks intimidating, your respondent may be put off before the start.Types of question* Behavioural questions prompt respondents to reveal what they have done in the past, such as what products or brands they have bought, or where they usually shop. These questions typically use the words ‘Who’, ‘What’, ‘Where’, ‘When’, ‘Why’, ‘Which’ and ‘How’, and they are designed to indicate what might happen to a product in the future.
They should reveal likely patterns of purchasing, and may be used to determine future market size, share and trends.* Attitudinal questions ask about the respondent’s chief likes and dislikes, attitudes and opinions. The intention is to determine the motivation s of buyers, i.e. why they buy what they do, and to determine what the company may need to do to improve its products and attract more customers.
* Classification questions try to classify the respondent in terms such as age, sex, occupation, socio-economic class, particulars of home ownership and marital status. This is useful in order to make sure that the sample used is a representative one, and it may also be used to see if different buying patterns and motivations are associated with different classifications of customer. This should help with successful segmentation of the market and indicate how different segments should be approached.
Uses of different types of questionOpen questions invite respondents to offer their opinions, and allow them to express themselves freely, possibly at length. Closed questions, sometimes referred to as multiple choice questions, only allow the respondent to choose from a number of given responses, and often these may be simply ‘Yes’ or ‘No’. Some questions are referred to as ‘hybrid’ questions, as they contain elements of both open and closed techniques. The above example asking which types of music the respondent listens to is an example of this, as it is principally multiple choice but the final option, ‘Other (please specify)….
… is open-ended.From the respondent’s point of view, closed questions are much easier to answer, and they are also invaluable for the interviewer as they are easy to analyse and results may be quantified relatively simply. It is possible to use some quantitative analysis techniques on responses to such questions.
Closed questions are difficult to design accurately, however – if insufficient options are given the respondent may not be able to answer. They may also lead to a certain level of bias, as naming particular responses can lead the respondent.Open questions can lead to a wide range of replies, and consequently they could be difficult to interpret and analyse. They are simple to design, however, and will not lead the respondent. They are often used where the researcher is not sure what the response is likely to be. However, do not simply use open-ended questions because you cannot be bothered to think of a well structured, closed question; some careful thought at this stage could gather data that is much easier to analyse. A questionnaire full of open-ended questions can, in certain circumstances, give the impression that it has been thrown together in a hurry. Make sure that the open-ended question is truly the best way to ask that question, not just the easiest.