Glossary – terms and meanings

  • Ad Tracking Research monitors people's perceptions and attitudes towards an advertisement over time. The purpose is to see how the brand is perceived and how this changes. It can also be used to examine the usefulness of the advert, and if it is cost effective. Tracking can be continuous (interviews conducted at regular intervals) or at specific points in time (for example before the advert is launched, just after it has been launched and when it has been around for some time). The same sample of respondents may be used to establish continuity, but as they have been asked the questions before and are likely to be aware of the questions about to be asked this can bias the results. Another option is to have separate samples of respondents of the same demographic.
  • Dichotomous questions have just two possible responses. Common dichotomous questions are those with the responses True/False, Male/Female or Yes/No. You can use these types of questions to “screen” respondents to ensure that they are eligible for the survey

    You can also use dichotomous questions to split your sample, so you can direct one half to one set of questions and the second half to a different set. This means you can ask the respondents who have bought your product about their customer satisfaction and those who have not why and what would encourage them to buy.
  • Face to face surveys. Respondents may be approached on the street and asked to complete a survey. Advantages are similar to those of telephone interviewing, but also include being able to target certain groups quickly. It's easy to speak to mothers of toddlers, for instance, since the interviewer can visually assess who they need to speak to. This is also the way to go if you need the opinions of people in a particular locality, since door to door interviewing can be very focussed.
  • Focus groups are a form of qualitative research. A number of respondents are gathered together to generate ideas by discussing a product, service or concept. There are generally six to twelve participants who are overseen by a moderator. Like hall tests, they are often used to try out new concepts or new products.

    Focus groups can yield detailed and interesting feedback that can be further developed by interaction with other members of the group. However it is more difficult to get individual's perceptions in this group setting, and some participants may not be willing to voice their true opinions when amongst strangers.
  • Hall tests are a form of market research where consumers are asked to “try out” specific products in a public space, often a church hall or similar, and give their opinion. Respondents are recruited on the street to ensure a random sample, and then guided to a specified venue where the tests will take place. Respondents are often given incentives to do this. Hall tests are often used to test reactions to products, such as food and drink, magazines, cleaning products etc, before they are released and to identify strengths and weaknesses of a product.

    Hall tests provide a secure and controlled environment which ensures that interviewing conditions remain the same throughout the research. You can also achieve a large number of interviews in a short amount of time and get in-depth comments from respondents. However you do need a high number of interviewers and a sufficient amount of space to conduct this type of research, and it needs to be carefully planned and executed to be successful.
  • Multiple choice questions: In this type of question, respondents can choose one or more responses, as opposed to just one in single choice questions. They may be able to give any number of responses, or they may be told the exact number of responses that are needed (for example “choose three” out of the possible eight).

    These are useful in PR surveys since their effect usually makes response stats look higher, as 80% could think that your product is one of the best on the market in a multiple choice question, but in a single choice question it could fall considerably when respondents choose your lead competitor over you. However the stats resulting from these questions do not add up to 100%.
  • Omnibus surveys: are surveys on a number of different subject areas. Usually they are for different clients who join a big survey with other clients. They can ask as many questions as they like, but because they are “sharing” the survey with others, the interviewees will almost always be the general population. This means that you can't target certain types of people or certain age groups unless you are prepared to interview fewer people.

    The other common limitation is time. You'll usually have to devise the questions on your own, have them ready for a certain deadline, and if you miss it you'll have to wait for the next survey. Also, other topics covered in the omnibus may affect the way respondents answer your questions, and a limit on the number of questions that you can ask may be imposed. But the great advantage of an omnibus survey is the price, which is usually very low because you are splitting the cost with other clients and the sample is largely nationally representative.
  • Online surveys: have become very popular recently because of their low cost. Survey respondents sign up to take part in surveys in return for a small reward, often 10p per survey or entry into a prize draw. Their results may be unreliable, especially if the survey company has allowed people to sign on under several different personas. Survey takers are often participating in several company's surveys, and so the accuracy and thoroughness of their answers may be doubted. The biggest problem is that online surveys can never be fully representative of the population in the same way as a telephone survey because they can only reflect the views of people who have Internet access (still just 70% of the population) and are willing to pro-actively access surveys for the sake of a 10p (or similar) reward.
  • Open ended questions: Most surveys ask a question and offer the interviewee a number of optional answers – closed ended questions. But when questions are “open ended” respondents are not given any possible answers to chose from. They are invited to answer a question in their own words with no prompting. This is a qualitative form of questioning. You can also add an “other” option to a closed end question and direct the respondent to a “please specify” open ended question.

    Some advantages to using open ended questions are that you get a more specific and detailed answer, rather than a response that is pigeon-holed into one of the given responses. You can also benefit by avoiding the bias that you may suffer by suggesting responses and putting words into people's mouths.

    However open ended questions are almost impossible to generate statistics for. In addition you run the risk of people not answering it at all or not properly understanding the question, especially if there is no interviewer present to correct them. They can also be time consuming, both for the interviewee and for the survey company which has to make sense of the answers.
  • Qualitative research: is an interpretative method, not based on hard facts like quantitative research. It is generally considered to be more flexible and in depth, as respondent's answers affect the way that they are questioned and what they are asked, unlike in most quantitative surveys.

    Though you get more detailed results and the “human interest” factor, the data you end up with is textual rather than numerical. This makes it hard to analyse and identify general themes and conclusions. Some example methods of qualitative research are focus groups, in depth telephone or face to face interviews and participant observation.
  • Quantitative research: In a quantitative research project a relatively large number of respondents are asked a standard set of questions. The wording and responses that the respondent must choose from for each question generally stay the same. By keeping things structured, you can analyse the statistics once they have been generated, and trust that the results you gain are legitimate.

    To produce useable results you must ensure that the sample you survey is balanced and nationally representative, and of a sufficient size to give reliable results. 1,000 is generally considered to be the “magic number” as it is large enough to be split regionally and still produce useable statistics. Some example methods of quantitative research are telephone surveys, face to face surveys, postal surveys and online surveys.
  • Straw Polls involve asking people about their opinions on a certain subject in an unofficial and informal way to find out what the general feeling is. Sometimes polls carried out without ordinary voting controls in places are also called “straw polls”. They are not usually nationally representative so you are unlikely to get a good cross section of the population.
  • Telephone surveys. This is the traditional method of conducting surveys, and which is highly regarded by the media. Some companys are moving away from this method in favour of online surveys. The reasons for this are that online surveys are cheaper and online research firms can usually contact a large number of respondents in a short space of time.

    However telephone surveys are regarded as solid and reliable as they stand up to scrutiny and it is easier to verify that the respondent's demographics are correct – such as age, gender and region of the country. You can also reach the full population (99.6% have a mobile or landline), and are not limited to interviewing only those with Internet access, as you are with online surveys.

Please choose a colour in which to view our website:

  • Purple l
  • Blue l
  • Green l
  • Pink l
  • Orange l

Our Clients

Consumer Analysis provide research analysis, polls & consumer surveys