Glossary of Market Research Terms

ad hoc :Applied to single surveys designed for a specific purpose as opposed to continuous, regularly repeated, or syndicated surveys.

advertising research :Any research into advertising except media research, particularly creative research, pre-tests of advertisements, and evaluation of advertising campaigns.

age groups :The age groups most often used in British surveys of adults are those of the National Readership Survey (up to 19, 20/24, 25/34, 35/44, 45/54, 55/64, 65 and over) or broader groups based on these.

agency :Market research companies are often referred to as research agencies.

aided recall :A means of helping people to remember things by reminding them of associated events or by prompting.

analysis :The summarizing of data in a way that is intended to make them more readily comprehensible

analysis of variance :A method of allocating the overall variation of a sample statistic among several variables, to show the strength and statistical significance of the associations with each of these variables.

animatic :A representation on film or video of a television advertisement, using a sequence of drawings or simple cartoon animation. Used for advertisement pre-testing to avoid the production costs of a finished commercial.

appraisal :Monitoring the effectiveness of an individual in carrying out assignments in accordance with training and instructions.

arithmetic mean :The sum of the observed values of a statistic divided by the number of observations.

ascription of adjectives :A method of measuring brand images, by showing a list of brands and asking to which of them a particular adjective or adjectival phrase applies.

association :The psychological technique of free or spontaneous association is used in qualitative research. It consists of eliciting the words or thoughts inspired by a number of stimuli, which might eg comprise pictures or possible names for a new product.

attitude :An attitude, as a basic psychological concept, is a learned predisposition to respond in a consistently favourable or unfavourable manner (~A with respect to a given object.

attribute :A difference between people, households etc. of a qualitative rather than a quantitative kind, eg sex, region or between brands, eg name, packaging material colour. Also used to refer to the individual descriptions which comprise an image battery, even where these are presented in the form of scales.

audience share :The percentage of the total viewing audience viewing a particular television channel.

Audit Bureau of Circulation :The ABC validates circulation claims for newspapers and magazines by collecting sales returns in a set format and carrying out spot checks of publishers' internal auditing procedures.

auditor :In data collection, an individual who on behalf of the organisation calculates volume sales by counting in-store stock and delivery invoices.

Note: The term auditor is a specific term used in market research and is unrelated to auditing within business in general or assessment auditing.

automatic interaction detection :A method of dividing a sample into groups based on analysis of variance, in such a way as to maximize the discriminating power of the groups for some dependent variable. The dependent variable might be e.g the frequency of purchase of a product.

The sample is divided progressively into parts, using demographic or other independent variables to define these parts. At each stage the program searches among the independent variables for the split which maximizes the between-group variance and minimizes the within-group variance for the dependent variable. Each group is subdivided until further splits become statistically insignificant, or sample sizes become too small.

This method produces a hierarchy of variables which are significant predictors. It has become a popular form of multivariate analysis for market research purposes, especially for the selection and definition of target groups.

average issue readership :The average number of people who see an issue of a periodical publication.

awareness: Brand awareness and awareness of advertising are often monitored by means of tracking studies. Unprompted or spontaneous brand awareness is measured by questions such as, 'Which brands of … can you think of?' Prompted awareness or recognition is measured by showing a list of brands and asking, 'Which of these… have you heard of?'

back check :A check by a supervisor or by the field office that an interview has been properly carried out, by telephoning or writing to a proportion of the respondents, or sometimes by means of a personal visit.

base :The base number for percentag, on a typical market research table, appears at the top of each column of percentages. Where both unweighted and weighted bases are shown, the unweighted base is the number of respondents in the sub-sample, and the weighted base is the number actually used for percentag.

Bayesian statistics :An approach to decision-making which combines previously- estimated probabilities with the information derived from a survey or experiment.

before/after test :A survey which is carried out before an event, usually some kind of advertising or promotion, and repeated afterwards, in order to detect and measure its effects.

behaviour :What people do as opposed to what they think. In a marketing context, behavioural research, behavioural segmentation, and behavioural theory are concerned with people's buying and consuming activities.

bias: An aspect of survey design which causes the expected value of an estimate derived from the survey to differ from its true value.

bipolar: Scales with two ends such as 'sweet - sour' may be described as bipolar, whereas a monopolar scale would establish the perceived degree of sweetness, eg from 'extremely sweet' to 'not sweet at all'.

bivariate : Bivariate data consist of observed pairs of values of two variables or attributes, from which it is possible to evaluate the relationship between them.

blind test :A product test in which the identity of the brand is not revealed. The reactions obtained, when compared with reactions to the branded product, provide a measure of the effects of branding.

blink rate :Hidden cameras have been used to record the speed at which people blink while looking at advertisements, packs etc., in the belief that a slower blink rate indicates a greater openness to suggestion.

Box-Jenkins :A method of statistical forecasting, based on analysis of time series.

bracket code :A code which represents two or more answers to a question, eg a single code used for all answers in the range 25-34 to a question inquiring about the respondent's age.

brand :A product or service which has been given an identity; it has a brand name and the added value of a brand image.

brand image :The set of associations which a brand has acquired for an individual.

brand positioning :The position of brands on a map. usually in two dimensions, which represents important factors influencing choice. These factors may include eg price, product attributes, user characteristics, and brand images.

brand switching: Changes in the brand purchased by a consumer within a product field, eg as recorded by consumer panels.

break(down): A sub-group used in analysis. These groups are often based on classification data, eg a breakdown by sex would comprise subgroups of men and of women.

brief(ing) :1. A research brief is a statement from the sponsor setting out the objectives and background of a study, and perhaps the method, timing etc, so that the researcher can plan accordingly.

2. Briefing of interviewers prior to a survey is intended to ensure that they understand fully the task to be undertaken.

Broadcasters' Audience Research Board :Since 1983, BARB has measured television audiences and viewers' reactions to television programmes on behalf of the BBC and the ITCA. The audience measurement system replaced the former JICTAR contract.

Business Statistics Office :One of the main parts of the Government Statistical Service, the BSO collects and publishes a wide variety of statistics on behalf of the UK government.

buying intentions :Questions about the likelihood of buying a product during a future period.

buy-response :Questions as to whether a consumer would be willing to buy a product at a number of different prices form the basis of the GaborGranger method of pricing research. These questions are used to plot a 'buy-response curve', which relates the percentage willing to buy the product, to the price.

CAI :Data capture by means of Computer Assisted Interviewing.

callback :1. A further attempt to contact a pre-selected respondent, telephone number,etc, or to secure co-operation, where the first attempt failed.

2. A secondinterview with the same respondent in the course of a single survey.

3. A call made by a supervisor to check that an interview has been carried out correctly.

canonical analysis :An extension of multiple regression to deal with two or more dependent variables.

CAPI :Data capture by means of Computer Assisted Personal Interviewing.

cartoon test :A projective technique in which respondents are asked to fill in 'speech balloons' in a comic-strip representation of a situation relating to the topic of inquiry.

CATI :Data capture by means of Computer Assisted Telephone Interviewing.

census :Enumeration of all the individuals in a population.

central location :Research carried out by a team of interviewers working at or from a single centre, eg hall tests, telephone interviewing.

Central Statistical Office :The CSO co-ordinates the collection of UK government statistics by the Business Statistics Office and the Office of Population Censuses and Surveys.

chi-square(d) :A statistical test as to whether a sample distribution conforms sufficiently to some other distribution. In market research, it is most often used to test a contingency table for the significance of any association between the two characteristics upon which the table is based.

classification :In market research this relates to a group of questions and observations, usually placed at the end of an interview. These tend to be of fairly standard form, and cover eg age, sex, marital status, household composition, status within the household (as head of household or housewife), social grade, terminal education age, tenure of home.

client :The internal or external client commissioning or contracting the project or parts of the project.

clinic :Research to aid product development, in which a group of respondents are invited to provide opinions, and modifications are made to the product before a further group is invited.

closed question :A question which has a limited number of logical answers (eg 'yes' and 'no'), as opposed to an open or open-ended question.

cluster :A group of neighbouring individuals. A geographical cluster comprises people, households, etc., in a relatively small area. Also used to denote a group of-people with similar attitudes or characteristics, or a group of brands which are perceived to be similar to one another.

cluster analysis :An approach to multivariate analysis which aims to identify groups of individuals that are relatively similar to each other, and dissimilar to individuals in other groups.

cluster sampling :A form of multi-stage sampling in which the final stage is the interviewing of all individuals within the selected group, such as everyone at a selected address.

cIustering/cIustered sampling :Almost all samples of individuals for commercial face-to-face interviews embody clustering to some degree. Instead of the individuals being spread evenly over the area occupied by the population, they are concentrated into a number of neighbouring groups or clusters. It is a consequence of multi-stage sampling, and is done in order to reduce costs.

code :A symbol used to classify data for the purpose of analysis.

Code of Conduct :The joint Code of Conduct of the Market Research Society and the Industrial Marketing Research Association consists of a set of principles and rules with which their members undertake to comply.

code frames :The result of grouping closely similar responses used in coding open-ended questions, created through looking at a proportion of the responses in detail.

coding A: The process of allocating codes, especially where open-ended questions require that this is done after fieldwork has finished.

coding B: The allocation of codes to groups of closely similar responses recorded against open-ended questions, in order to analyse that information together with the rest of the data.

coding frame :A set of codes used to categorize answers, usually to a single question, sometimes to several questions taken together.

computer - assisted telephone interviewing :Telephone interviewing conducted with the aid of a computer. The computer displays to the interviewer the question that is to be asked, together with precoded answers (if any).

concept test :A study intended to obtain reactions to an idea for an advertisement or a product, before investing in production.

confidence limits :The values of a parameter which form the upper and lower boundaries of a confidence interval.

conjoint measurement :A method of evaluating consumer preferences among product concepts which vary in respect of several attributes, based on asking people to rank a number of contrasting combinations in order from the most to the least preferred.

consumer:1. The ultimate user of a product, as opposed to the purchaser.

2. Moregenerally, people or households who use or buy goods and services, as distinct from the producers and distributors.

consumer panel :A sample of individuals whose purchases, product usage, andlor media consumption etc. are recorded over a period.

contact:Somebody contacted in the course of a survey, but who has not necessarily completed an interview.

contact sheet :A record of the contacts and attempted contacts made by an interviewer.

continuous research :Research which is undertaken on a continuing basis, or is regularly repeated at frequent intervals, as opposed to ad hoc surveys. Examples include retail audits, consumer panels, and tracking studies.

continuous scale :A type of diagrammatic or graphic rating scale, which does not restrict the respondent to any specific number of discrete response categories.

control group :A group of individuals who provide a standard of comparison in a test. They are exposed to no test stimuli, or to stimuli for which the results are known.

Copland formula :A method of estimating the cover or reach of a poster campaign in a town. The formula involves the number of 'average' sites used, and two parameters, one of which is a function of the population size.

copy test :A test of advertising copy, intended to discover how well it succeeds in communicating and how consumers react to it.

corporate image :The image of an enterprise as a whole rather than of the particular goods or services which it supplies.

correlation :The interdependence between attributes or variables, particularly the relationship between the values or ranks of two variables.

cover(age) :The cover, coverage or reach of a single advertisement is the percentage of the target audience to whom it is exposed.

creative research :Research applied to the creation (usually) of advertisements. Embraces advertising pre-tests, communication tests, concept tests, copy tests, etc.

creativity groups :1. Extended group discussions in which a variety of projective techniques are employed.

2. Groups using brainstorming and similar methods for product development purposes.

cross-tab(ulation) :A two-dimensional table, based on answers to two of the questions included in a survey, eg brands used by frequency of product use.

data cleaning :The correction of data records to make them consistent in relation to filters and data logic.

data collection :A general term covering the collection of information such as:

• - face to face interviewing in-home, in the street, in a central venue, at place of work;

• - group discussions, depth interviews;

• - telephone interviewing from a telephone centre;

• - telephone interviewing from interviewers’ home;

• - auditors collecting in-store information;

• - evaluators conducting mystery shopping and client service work;

• - respondents completing self-completion/postal questionnaires;

• - through electronic techniques.

data entry :The input of questionnaire information into electronic format ready for survey analysis.

data fusion :The combination of partial information from separate surveys, which have been carried out using independent samples drawn from the same population, to form a single database.

data preparation :A general term covering coding, data entry and editing/data cleaning that prepares data for analysis.

Data processing (DP) :Data preparation and data analysis.

DP executive/analyst :The person with overall responsibility for the project during data preparation and/or data processing stages.

data protection :The Data Protection Act 1984 regulates the maintenance and use of automatically processed personal information.

data record :Data relating to the completed questionnaire of an individual respondent, item/transaction (ie. reportable unit).

database :A set of computerized data available for analysis.

debrief :An informal verbal report by a qualitative researcher, or by an interviewer following completion of pilot interviews.

Delphi method :A method of forecasting which derives a consensus view from a group of experts.

demographics :Demographic variables, eg sex, age, marital status and social grade, normally comprise a large part of the classification data obtained in market research interviews.

depth interview A: An informal face-to-face interview, which is only loosely structured, and appears more as a conversation than as a question-and answer session.

depth interview B: An interview that is not confined to a structured questionnaire.

desk research :As opposed to the collection of primary data via field research, desk research is based on the use of secondary data, eg directories, lists, statistics, reports of past surveys, and published information generally.

diadic :See PAIRED COMPARISON.

diagrammatic scale :A type of rating scale, which is distinguished from numerical and from verbal scales, and is also known as a graphic scale or spatial scale.

diary :A means by which members of panels may record their behaviour over a period of time, eg purchases, consumption, television viewing, journeys.

dichotomous question :A question to which there are only two possible answers (apart from 'don't know', 'no preference', etc.).

direct question :A question asked directly of the respondent about his own behaviour, opinions, etc., as opposed to an indirect question. Most questions used in structured questionnaires are direct.

distribution check :An observational survey of a sample of retail outlets, which measures the presence or otherwise of specified products, brands and pack sizes. Prices and display may also be recorded.

editing A: Procedures for tidying up survey data. These begin with inspections of returned questionnaires to ensure that they conform with sampling requirements and that key questions have been answered.

editing B: A comparison of actual responses with possible responses, as defined by questionnaire structure and logic and then, if necessary, rejecting or amending some responses.

electoral register :The annual Register of Electors, which is based on returns made in October and published in February, is in Britain the sampling frame most commonly used for random samples (except for telephone surveys).

electronic data capture (EDC) :Use of computers to record or assist the recording of primary research data.

 :NOTE: This includes CATI and CAPI versions and other methods.

establishment survey :A survey to establish (usually) the proportion of households possessing certain kinds of appliance, etc.

executive :An individual responsible for running, singly or jointly, a project or parts of a project.

exploratory research :Research undertaken where little is already known about the subject, often as a preliminary to a survey. Relatively quick and cheap methods are usually employed, eg desk research, group discussions, omnibus surveys, and street or telephone interviews on a small scale.

exponential smoothing :A method of smoothing a time series and preparing a short-term projection. Each projected value which produces the smoothed curve is a weighted average of all observations for previous points in time, the weights forming a geometric series which diminishes as the time recedes.

exposure :A single opportunity for a member of the target audience to see an advertisement.

extended groups: Group discussions lasting several hours, often employing projective and other techniques.

F-test :Also known as the F-ratio or variance-ratio test, this will show whether two or more samples with different means could plausibly derive from the same population.

face-to-face interview :This fully descriptive term is to be preferred to the term personal interview, which may sometimes be taken to include both face-to-face and telephone interviews.

factor analysis :A branch of multivariate analysis based on the correlation coefficient, used mainly to investigate the structure of attitudes.

Field based individual :In data collection, an individual who works alone without direct and full time supervision.

field force: The interviewers and supervisory staff employed by or available to a company, usually a market research company.

field research :As opposed to desk research, the collection of primary data from external sources by means of surveys, observation and experiment.

filter :An instruction printed on a questionnaire as to which questions should be asked next, depending on previous answers.

Fishbein model :A model of the relationships between behavioural intentions on the one hand, and beliefs and attitudes on the other.

five-bar gates :A manual method of counting classified data.

forecast :The expected magnitude of some quantity or the estimated probability of an event at a future time.

frequency :The number of occurrences of some particular kind of event, eg answers to a question.

forced editing :The application of rules specified for any data items which are deemed to be incorrect (eg. a multiple answer given when only a single answer is required).

 :NOTE: These rules are written as a computer program and applied to data records relating to a project.

Gabor-Granger method :A method of pricing research, used for new products and variants of or improvements to existing products. Respondents are shown eg a test pack, and asked whether they would be willing to buy it at each of a randomized set of prices.

geo demographics :A method of classifying households based on multivariate analysis of data from the Census of Population. 

grossing up :The scaling-up of the results of a survey. experiment, test market, etc., to the whole population or market.

Group discussion :Two or more respondents discussing an issue collectively.

group discussion (also known as focus groups) :One of the basic methods of qualitative research, often used in exploratory work and when the subject matter involves social activities, habits and status.

hall test A: A test for which people are taken to some fixed location, often a public hall.

hall test B: Research conducted at a central venue (eg. car clinic, product testing, advert testing)

hand tab(ulation) :Sorting questionnaires, counting and tabulating the answers manually rather than by computer.

head of household :That member of a household who either owns the accommodation occupied by the household or is responsible for the rent, or, if the accommodation is occupied rent-free, the person who is responsible for the household having it rent-free.

If, however, this person is a married woman whose husband is also a member of the household, then the husband counts as the head of the household.

hole count :A term deriving from the use of punch cards in data analysis, meaning a simple count of all the codes present in all of the records, without any breakdowns.

hole count

(frequency count) :A summary count of individual data items on the computer file

home audit :A panel of households, used for regular measurement of product purchases and in some cases consumption. Data may be collected by means of diaries, interviewers, and dustbin checks.

hot decking :This technique requires the survey to be sorted into an order so that respondents with similar answers on a range of key variables are placed together.

household :A private household consists of one or more people living together, whose food and other household expenses are usually managed as one unit.

hypothesis :Any supposition, whether or not based on evidence, or an assumption made as a basis for reasoning.

image :People's perceptions or impressions of a product, service, company, person, etc., however these may have been formed. and however much they may reflect reality.

incentive :An inducement to co-operate in a market research study.

indirect question :A question which seeks the respondent's views about other people's behaviour or attitudes, used in qualitative research as a projective technique to uncover ideas which the respondent might otherwise be reluctant to reveal.

in-home test :A product test, usually, conducted in participants' homes rather than at some central location, hall, store, etc.

in-house research :Research conducted by the organization that wants the information, rather than by a research agency etc. acting on its behalf.

ink blot test :The Rorschach and Holtzman ink blot tests, more widely used in clinical work than in market or social research, are projective tests in which the subject is asked to describe what he sees in a number of haphazard shapes.

interlocking quotas :Quotas which specify the numbers of interviews required in each cell of a matrix defined by the specified characteristics, eg with a matrix defined by sex, age and social grade.

interview :A contact with an informant, or sometimes a group of informants, in order to obtain information for a research project.

interviewer :An individual who on behalf of the organisation recruits/interviews respondents face-to-face or by telephone.

 :NOTE: Those who conduct depth recruitment and interviews are treated as interviewers.

interviewer instructions :Directions to the interviewer printed on the questionnaire, including eg filters or skips. They are usually distinguished from the questions by the use of capitals, or sometimes by boldface or italic type.

Interviewer Quality Control Scheme :Established independently in 1986, the IQCS assumed the function of verifying the quality of fieldwork, that had been one of the purposes of the former Interviewer Card Scheme of the Market Research Society.

joint Industry Committee for National Readership Surveys :JICNARS represents the Press Research Council, IPA and ISBA, and controls the National Readership Survey.

judg(e)ment sample :Technically, any non-random sample can be described as a judgment or purposive sample. Statistical theory relating to random sampling cannot properly be employed to calculate confidence limits for estimates derived from judgment samples.

Kish grid/box :A table for use by interviewers in random sample surveys, to select one person from a household. The procedure gives an approximately equal chance of selection to each member of the household.

laboratory test market :Any kind of consumer research or simulation aimed at forecasting the sales of a new product, short of actually selling it in the normal way. is occasionally described as a laboratory test market.

large sample :For the purpose of applying standard statistical tests, a sample is described as large where the sampling distribution is normal, or pproximately so. This is usually taken to mean a sample size of at least thirty, preferably fifty or more.

life-style :A life-style, or lifestyle (the hyphen is often omitted) is a way of life for a community or an individual, expressed particularly in activities, interests and opinions, and to some degree in the products and brands consumed. Various lifestyle classifications have been devised for commercial use.

Likert scale :A type of verbal rating scale in which respondents are asked to indicate the extent of their agreement or disagreement with a series of statements, eg for a number of brands. Also called an agree/disagree scale.

logic data entry :Data entry processes which are programmed for specific projects to check question skips and response range checks.

 :NOTE: Automatic serial numbering alone does not constitute logic data entry.

mapping :Rules or formulae by which the elements in one set can each be made to correspond to a single element in a second set.

market map :A diagram which shows the relative positions of brands in terms of the most important brand characteristics, sometimes used to summarize the findings of attitude research.

market research :Market research has sometimes been distinguished from marketing research, to mean the collection of data about markets by means of surveys.

Market Research Society :The MRS was founded in 1946, and is the incorporated UK professional body for those using survey techniques for market, social and economic research. The Society aims to promote and protect the interests of the profession, and to maintain and improve standards of competence.

market share :The proportion of a market accounted for by a particular brand or supplier, either by volume or by value, or sometimes in terms of the number of consumers.

marketing :The management function responsible for identifying, anticipating and satisfying consumer requirements profitably.

marketing mix :The set of choices made by an organization in respect of those marketing factors which it can control. Important categories in the marketing mix are the so-called four Ps, namely product, price. place, and promotion.

mean :The arithmetic mean. Sometimes refers to other kinds of average.

media :May refer to any means of communication. In an advertising context, includes independent television and radio, newspapers and magazines, poster sites, cinemas, and other means by which advertising is communicated.

media research :Research into readership and media audiences.

media schedule :A plan for an advertising campaign setting out the media to be used, when and how often the advertisement is to appear, and other such details.

median :The middle of a set of numbers, one-half of the numbers being larger and one-half being smaller.

mode :The most frequently occurring value in a set of observations.

moderator A: The leader of a group discussion.

moderator B:  An individual who is responsible for facilitating the interaction of the group discussion members, and for capturing the data generated.

monadic :A monadic product test is one in which each person tests just one product, as distinct from comparative tests, eg diadic and triadic tests.

monopolar :An attitude scale is monopolar, as opposed to bipolar, where it measures one quality only.

motivation research :Small-scale studies aimed at discovering reasons for people's behaviour.

moving average :For a time series, a series of averages such that each covers the same number of successive periods.

MRS freephone :An independent service provided to the public to verify the legitimacy of UK market research agency members.

multi-client research: Also termed syndicated research, this describes studies for which the costs and the findings are shared amongst a number of clients.

multi-phase sampling :A survey in which some information is collected from the whole sample, and additional information is collected from sub-samples, whether at the same time or later.

multiple choice question :A closed question with more than two possible answers apart from 'don't know', sometimes termed a cafeteria question.

multiple regression :A mathematical model in which a dependent variable is represented as a linear function of so-called independent or predicated variables.

multi-stage sampling :A sample which is drawn in stages. The units sampled at each stage are each regarded as being composed of a number of units of the next stage, until the final sampling units are reached, ie the people who are actually interviewed.

multivariate :Multivariate data comprise observations for each of which three or more variate values are recorded. Multivariate statistical methods are those which simultaneously examine the relationships among a number of variables.

mystery shopping :Client service evaluation process involving individuals who on behalf of the organisation collect information by behaving as customers and report their findings as a way of monitoring the quality of performance.

National Readership Survey :A continuous readership survey established in 1954 by the Institute of Practitioners in Advertising, and since 1968 conducted under the aegis of the joint Industry Committee for National Readership Surveys.

nested sampling :See MULTI-STAGE SAMPLING

niche :A small and specialized market segment, that is capable of being profitably exploited.

non-response :That part of a pre-selected sample of named individuals, or of any random sample, from which no response is obtained.

normal distribution :Also known as the normal curve and the Gaussian distribution. A statistical frequency distribution, which is of particular importance in sampling theory, since it is a close approximation to the sampling distributions of many statistics derived from reasonably large samples. It is used to make statements about the confidence limits which may be attached to estimates derived from random samples, and for tests of statistical significance.

normative: Normative beliefs and statements, as opposed to positive beliefs and statements, are those which concern value judgments. They express opinions about what ought to be.

null hypothesis: The central hypothesis in a test of statistical significance.

numerical scale: Any scale which is represented by numbers, as distinguished from diagrammatic and from verbal scales.

objective: A market research proposal incorporates a statement of the information objectives of the project, and these may be explicitly related to marketing objectives.

observation: The alternative to questioning as a way of obtaining primary data.

observational research :Collection of information without the involvement of a respondent.

 :NOTE: Excluding the processes of auditing and mystery shopping.

occupation: The basis of classification by social grade and of other social status scales.

off-the-shelf product or methodology :A research product or methodology developed and made available to meet the needs of more than one client.

omnibus survey: A survey which covers a number of topics, usually for different clients.

open-ended question: As opposed to a precoded question, one where the answer is recorded verbatim, or as fully as practicable, and the answers are coded at a later stage.

opinion poll: A survey of opinions about political, social and other issues of public interest, especially as a basis for forecasting voting behaviour.

opportunity -to -see: A single impact, exposure, or opportunity to see an advertisement is effectively defined by the method of audience measurement.

optimization analysis: In sensory evaluation, optimization analysis can be carried out when sufficient data points are collected on formula variables in a controlled research design.

ordinal scale: Ordinal numbers express precedence (first, second, third, etc) and an ordinal scale is one which produces such a ranking.

organisation :Market research or social research body.

Osgood scale: See SEMANTIC DIFFERENTIAL.

outdoor advertising: A term covering posters, supersites, signs, facias, advertisements on vehicles, and in general all advertising which appears in public places.

Outdoor Site Classification and Audience Research :OSCAR is the UK poster audience reses arch system, launched in October 1985 by the Outdoor Advertising Association.

overclaim: A tendency for respondents to say that they have done some thing more often, have consumed more of something, etc., than is in fact the case.

pack test: Any kind of market research to evaluate a package design, in terms of eg its functional efficiency, its visual impact at the point of sale, the image it conveys of the product, and its influence on sales.

paired comparison: A product test, also known as a diadic test, in which people are asked to compare two products.

panel: A sample of people, households, or retail outlets, or sometimes of other kinds of organization, from which information is obtained on more than one occasion.

penetration: The proportion of a population, or of a subgroup, who have a certain characteristic.

peoplemeter: An audimeter with provision for recording whether individuals are present.

personal interview: Usually taken to mean an interview carried out face-to-face between an interviewer and a respondent.

pilot: A pilot survey is a small-scale replica of a main survey, carried out beforehand in order to reveal the problems likely to be encountered, or to help in the design of the main survey.

placement test: A product test in which samples of the product are given to people to try in their own homes, or wherever the product is normally used.

poll: A sample survey in which political opinions form the subject matter.

population: The whole of the material from which a sample may be taken.

postal survey: Any survey carried out by post, using a self-completion questionnaire or diary. A mail survey in American usage.

poster research: Research to measure audiences for outdoor advertising. See OSCAR.

precoded question: A question for which the respondent is constrained to choose one or more from a set of allowable answers, or where the interviewer is similarly constrained in recording the answer given.

preference test: A product test in which consumers are asked to compare a number of trial products, and to express their preferences.

pre/post: Refers to market research conducted before some event, such as a product launch or advertising campaign, and repeated afterwards.

preselected sample: A sample in which the individuals who are to be interviewed have been selected prior to fieldwork, as distinguished from field sampling in which the interviewer makes the final selection.

pre-test: As opposed to post-test, a test of something before it has been exposed to the public, particularly advertising and promotional material.

price quotation :Response to a client enquiry that is a priced response for specified work.

 :NOTE: This may reference the research design of another project.

pricing research :Any kind of research which aims to show how demand for a product or service will vary with its price.

proposal :Response to client enquiry for work which includes some element of research design.

recruiter :An individual who on behalf of the organisation recruits respondents for depth interviews or group discussions.

 NOTE: Those who conduct depth and group recruitment are treated as recruiters if they do no other type of work.

remote listening-in: In data collection, a method of validating the authenticity of responses and quality of interviewing in the telephone centre by the manager or supervisor listening to both sides of an operator’s calls.

NOTE: This may be by using monitoring equipment of voice recording media.

research universe :The scope of population segment included in the survey.

sampling error: An estimate derived from a sample is usually different from the true value for the population as a whole. The term standard error is often used to refer to sampling error.

score: A numerical value assigned to an observation, eg an answer given in response to a rating scale.

screening: A brief interview for the purpose of locating individuals with certain characteristics who are needed for a survey.

second person review :A review carried out by a second suitably (or similar) skilled individual to confirm the research objective can be met by the proposed design.

segment: A part of a market or population.

segmentation: Division of a market into parts, each of which has identifiable characteristics of actual or potential economic interest. Most often segmentation is in terms either of characteristics of the product or service, or of purchaser/user characteristics.

self- completion: A self-completion questionnaire is one that is completed by the informant rather than by an interviewer.

semantic differential: A commonly-used attitude scaling technique, also known as Osgood scales. It is a bipolar diagrammatic rating scale.

semi- structured:An interview or questionnaire in which many or even all of the questions may have been specified in advance, but the questions are typically of open or open-ended form, and there is extensive use of probing techniques.

sensory evaluation:This is a research technique to assist technical, research and development management design better products. Data are mapped to show where the subject being researched stands against consumer preferences and in comparison with competition.

signed-off :A record kept on file confirming that the specified action(s) have been carried out, by whom, and on what date.

significance test: Analysis of sample data, to show whether or not they support a given hypothesis about the parent population(s).

simple data entry :Data entry process containing no in-built logic checks

simulated test market: See LABORATORY TEST MARKET.

SMART (Salient Multi-Attribute Research Technique): This technique takes the view that a better way to measure interest in a service company's change is in terms of improvements. The aim of the analysis is specialized since it concentrates on identifying areas of highest cost/benefit in service areas.

social grade (or socio-economic class): The socio-economic classification system used by the National Readership Survey, and generally for market research in the UK. The social grade of an informant is based on the occupation or former occupation of the head of the family unit, or in certain circumstances, eg where the head of the family unit is retired or unemployed and has a low income, it may be based on the occupation of the chief wage earner. Usually this person is also the head of the household. The classes are:

Social Grade Social Status Occupation

• A Upper middle class Higher managerial/administrative/ professional

• B Middle class Intermediate managerial/administrative/ professional

• Cl Lower middle class Supervisory or clerical and junior managerial/administrative/professional

• C2 Skilled working class Skilled manual workers

• D Working class Semi and unskilled manual workers

• E Those at lowest levels of subsistence State pensioners or widows (no other earnings), casual or lowest grade workers

The six classes are often combined into four (AB/C1/C2/DE) or into two (ABC1/C2DE).

specialist interviews :Specifically related to executive/business to business/ medical/specialist samples or depth interviewing.

Standard Industrial Classification: A classification of industrial establishments.

Standard Region: One of the ten regions into which Great Britain is divided according to the Registrar General's system.

storyboard: A set of drawings together with a script and a description of sound effects, which sets out the action in a television commercial.

stratification: A technique used in sampling, to ensure that the sample is representative in terms of the factor(s) used for stratification. The population is first divided into a number of sub-groups or strata, eg by geographical area. The required numbers are then sampled from each stratum.

street interview: A brief interview conducted in the street or other public place, usually employing a quota sample.

structured: A structured questionnaire sets out precisely the wording of the questions and the order in which they are to be asked.

supervisor :An individual who on behalf of the organisation undertakes at least one of the following data collection tasks:

• - interviewing and selecting new recruits;

• - training;

• - appraisals/accompaniments;

• - allocating work/progress chasing and/or project and quota control.

NOTE:1 If an individual only validates data collection work they are not deemed to be a supervisor.

NOTE:2 An individual employed full or part-time as an office-based member of staff is not deemed to be a supervisor.

supplier :An individual or organisation that provides products or services that have an impact on the services delivered to the client by the research organisation.

t-test: A test of statistical significance, based on the use of tables of Student's t-distribution. This describes the sampling distribution of the mean for small samples, ie where the sample size is less than 30 or so. The t-test may be used to compute confidence intervals, to test whether an observed sample mean differs significantly from some expected value, and to test the difference between two sample means.

tach istoscope: A device which permits brief glimpses of stimulus material, the exposures being controllable in small steps from .01 second or less.

telephone centre :A central office housing a group of telephone lines used for market research recruitment/interviewing.

time series: A set of sequential values which are dependent on time, ie statistics of quantitative observations of some kind which relate to different periods of or moments in time.

topic guide: A list of the topics to be covered in a depth interview or group discussion.

trade-off analysis: A research method aimed at discovering the most attractive combination(s) of attributes for a product or service. Price may be included as one of the attributes, represented at a number of levels. See also CONJOINT MEASUREMENT.

traffic count: Observation of the flow of vehicles or people past a point or along a route, used in eg poster audience research and transport planning.

variance: A statistical measure of the variability or dispersion of a set of numbers. It is the arithmetic mean of the squared differences between each number and the mean of all the numbers.

variate: A variable which has an associated frequency distribution, eg all the quantities measured in a survey, whether facts or opinions, can be described as variates.

verbal (rating) scale: Any kind of scale for the measurement of attitudes or behaviour in which the permissible answers are expressed verbally, as distinguished from diagrammatic or numerical scales.

verification :A checking process during data coding and data entry stages undertaken by a second person.

viewing labs :Venues specifically arranged to accommodate group discussions.

virgin respondents :Respondents who attend a group discussion for the first time.

visual aid: Anything which is shown to someone as an aid to communication. In market research the term is applied both to the prompt cards etc. used in interviews.

weighted sample: A sample to which post-weighting has been applied, ie weighting after fieldwork.