Glossary of market
research terms

Access panels

A database of individuals who have agreed to be available for surveys of varying types and topics. Rising rates of refusals and non response, make it more difficult to recruit for a single survey, therefore sampling from a pool of potentially willing marketing research respondents can be seen as an appropriate way of saving time and money.

Accompanied shopping

A specialised type of individual depth interview, which involves respondents being interviewed while they shop in a retail store and combines observation with detailed questioning.

Alternative hypothesis

The hypothesis where some difference or effect is expected (i.e. a difference that cannot occur simply by chance).

Ambiguous question

A badly constructed question which results in respondents and researchers reading different meanings into what is being asked, resulting in inappropriate or unexpected answers.

Animatics

A type of stimulus material where key frames for a television advertisement are drawn or computer generated with an accompanying sound track.

Annotation method

An approach taken to analyse qualitative data using codes or comments on the transcripts to categorise the points being made by respondents.

ANOVA

Analysis of variance. A test for the differences among the means of two or more variables.

Area sampling

A type of cluster sampling in which the clusters are created on the basis of the geographic location of the population of interest.

Artificial Intelligence (AI)

The use of computer algorithms and systems to simulate human intelligence and perform tasks such as data analysis or decision making. See Machine Learning, which is a major subset of AI.

Audience's thinking sequence

The sequence of thoughts that people go through when they are being communicated with.

Audits

An examination and verification of the movement and sale of a product. There are three main types: wholesale audits, which measure product sales from wholesalers to retailers and caterers, retail audits, which measure sales to the final consumer, and home audits, which measure purchases by the final consumer.

Bar chart

A chart which uses a series of bars that may be positioned horizontally or vertically to represent the values of a variety of items.

Base

In a cross tabulation, the base defines the total size of the sample or subsamples on which percentages are calculated.

Bayes' theorem

Bayes' theorem is a mathematical formula that provides a way to update the probability of an event based on new information. In market research, Bayes' theorem can be used to estimate the likelihood of a target audience having a certain characteristic, given data on prior behaviours or characteristics. For example, if we have prior data suggesting that 30% of females are likely to purchase a certain product, and we also have data suggesting that 60% of people who purchase that product are female, we can use Bayes' theorem to estimate the likelihood that the product will be purchased.

Beauty parades

The procedure of asking a number of agencies to present their proposals verbally to the client company. The procedure is used to assist clients in selecting the research agency that will undertake a research project.

Behavioural science

Behavioural science is an interdisciplinary field that applies theories and techniques mainly from psychology, but also from economics, sociology, and anthropology. Researchers often use behavioural science to uncover what consumers value and to provide solutions to pricing, choice architecture (nudge theory), perceptions and behaviours that are often difficult to gather using traditional research methods. Also see System 1 and System 2 thinking.

Big data

A term to describe the significant volume and variety of data available to organisations and the increased frequency in which they are generated.

Biometrics

Biometrics refers to the use of physiological or behavioural characteristics to identify and verify a person's identity. These characteristics include fingerprints, facial recognition, voice recognition, and signature analysis. Biometrics is increasingly being used in market research studies as a way to authenticate survey respondents.

Blog

"An abbreviated title for the term web log, meaning a frequent, chronological publication of personal thoughts and ideas. Twitter is a form of microblogging service that allows an individual to publish their blog type opinions and ideas in short Tweets
( text-based messages of up to 140 characters)."

Brand Equity Modelling

Brand Equity is ultimately the strength of the brand. But there are many different views on what Brand Equity is and how to measure it within surveys. Brand Equity modelling is the creation of a Brand Equity measure, and also any Key Driver analysis to determine what drives Brand Equity (e.g. which brand imagery statements are most related to Brand Equity).

Brand Fit

Brand fit measures the fit of a brand’s profile of product qualities with the desirability of these qualities within a segment. For example, if a brand performs well on criteria that are important to a particular segment, then that brand will have a good fit with the segment. A very helpful analysis for optimising brand positioning.

Brand mapping

A projective technique which involves presenting a set of competing brand names to respondents and getting them to group them into categories based on certain dimensions such as innovativeness, value for money, service quality and product range.

Brand personalities

A projective technique which involves respondents imagining a brand as a person and describing their looks, their clothes, their lifestyles, employment, etc.

Brand Price Trade Off

Brand Price Trade Off is a technique used for establishing brand and price preferences. Respondents are presented with a set of branded products priced at the lowest possible price point for each of the brands in question. The respondent is asked to select which products they’d choose. The price of the product selected is increased to the next level, and they are then asked again which product they’d choose. This process is repeated until a product reaches its maximum price and is still selected. Competitive demand curves can be created as well as specific price scenario testing.

Canonical Correlations

A statistical technique used to identify how much one set of independent variables (e.g. Age, Gender, Social Class) drives another set of dependent variables (e.g. Snack Choice). It is particularly useful when there are multiple dependent variables or the variables are categorical (e.g. Age, Gender etc.). It can also be used within segmentation, for example to segment on the relationship between attitudes and behaviours.

CAPI

CAPI stands for Computer-assisted personal interviewing. Where laptop computers, tablets or mobile devices are used rather than paper-based questionnaires for face-to-face interviewing.

Car clinic

A car clinic is a type of market research method that is used in automotive research to gather information about consumer preferences and to identify potential issues or problems with a vehicle before it goes into production. During a car clinic, participants are invited to a facility where they can interact with a pre-production or prototype vehicle, and provide feedback on aspects like design, features, and functionality. Car clinics are typically conducted with a small group of participants.

Cartoon completion

A projective technique which involves a cartoon that the respondent has to complete. For example, the cartoon may show two characters with balloons for dialogue. One of the balloons sets out what one of the characters is thinking or saying, while the other is left empty for the respondent to complete.

CATI

Computer-assisted telephone interviewing. CATI involves telephone interviewers typing respondent's answers directly into a computer-based questionnaire rather than writing them on a paper-based questionnaire.

CATS

Completely automated telephone interviews which use interactive voice technology and require no human interviewer. Respondents answer the closed-ended questions with their touch tone telephone.

Causal research

Research that examines whether one variable causes or determines the value of another variable.

CAWI

CAWI stands for Computer-assisted web interviewing. Also known as online interviewing. CAWI involves a respondent filling in a self-completion questionnaire, with questions and answers appearing on their screen, delivered by the Internet.

Census

Research which involves collecting data from every member of the population of interest.

CHAID (CHi-Squared Automatic Interaction Detection)

CHAID is a type of decision tree technique, based upon significance testing. It can be used to create rules to classify future respondents into identified groups using a number of different questions, or detecting interrelationships between different questions. For example, the combination of age and gender helps to further explain the different types of snacks that people consume.

Chat rooms

A web-based platform that can be used for online focus groups where individuals are recruited who are willing to discuss a subject online usually using text.

Chatbot survey

A Chatbot survey is a traditional survey presented to the respondent in the form of an online conversation. Instead of answering static questions, the respondents are invited to engage in a conversational exchange or chat.

Chi-square

A statistical test which tests the 'goodness of fit' between the observed distribution and the expected distribution of a variable.

Choice-Based Conjoint

Choice-Based Conjoint is a specific type of conjoint analysis where respondents are asked to make a choice between different sets of products/services, to derive the overall appeal of each component part. This is given either as a discrete choice, or as a chip-allocation style response (e.g. number out of 10 next purchases allocated to each product/service).

Closed question

A question that requires the respondent to make a selection from a predefined list of responses. There are two main types of closed questions: dichotomous questions with only two potential responses and multiple response questions with more than two.

Cluster analysis

"Cluster analysis is the statistical term for the creation of segments – the process of dividing markets into groups that are similar to each other, but different to the other groups. There are a number of different ways to segment, the two most common being:

Consumer Segmentations – used to understand which consumers to target and service with distinct marketing propositions, or to tailor brands, products, pricing, communication to specific groups and make more effective use of marketing resource.

Occasion-Based Segmentations – used to understand Needs on different occasions (e.g. Had a coffee to wake up in the morning, or had a social coffee with friends after work) in order to help with new product development in repertoire markets, or for brand positioning for clients with multiple brands."

Cluster sampling

A probability sampling approach in which clusters of population units are selected at random and then all (one-stage cluster sampling) or some (two-stage cluster sampling) of the units in the chosen clusters are studied.

Coding

The procedures involved in translating responses into a form that is ready for analysis. Normally involves the assigning of numerical codes to responses.

Coefficient alpha

See Cronbach alpha

Coefficient of determination

Measure of the strength of linear relationship between a dependent variable and independent variables.

Concept boards

A type of stimulus material which uses a set of boards to illustrate different product, advertising or pack designs.

Confidence level

The probability that the true population value will be within a particular range (result +/– sampling error).

Conjoint analysis

"An abbreviation of “consider jointly”, Conjoint Analysis is a powerful statistical technique to understand what combination of a limited number of attributes or features is most influential in the consumer’s decision-making process.

Conjoint Analysis is a multivariate technique based on the idea that when choosing a product or service, people will trade off features simultaneously. This process is replicated within a respondent exercise – a number of scenarios are shown where respondents are asked to express choice or preference amongst a set of products/services comprising of different combinations of attributes levels. A model is created from the resulting data to simulate “what if” scenarios and to assess to impact of each attribute in the decision-making process. Conjoint analysis is often used in concept testing studies, and pricing research."

Constant sum scales

A scaling approach which requires the respondent to divide a given number of points, usually 100, among a number of attributes based on their importance to the individual.

Construct validity

An analysis of the underlying theories and past research that supports the inclusion of the various items in the scale. It is most commonly considered in two forms, convergent validity and discriminant validity.

Content analysis

The analysis of any form of communication, whether it is advertisements, newspaper articles, television programmes or taped conversations. Frequently used for the analysis of qualitative research data.

Content analysis software

Software used for qualitative research which basically counts the number of times that pre-specified words or phrases appear in text.

Content validity

A subjective yet systematic assessment as to how well a rating scale measures a topic of interest. For example a group of subject experts may be asked to comment on the extent to which all of the key dimensions of a topic have been included.

Continuous research

See Longitudinal research.

Contrived observation

A research approach which involves observing participants in a controlled setting.

Convenience sampling

A non-probability sampling procedure in which a researcher's convenience forms the basis for selecting the potential respondents (i.e. the researcher approaches the most accessible members of the population of interest).

Convergent validity

A measure of the extent to which the results from a rating scale correlate with those from other scales or measures of the same topic/construct.

Cookies

Text files placed on a user's computer by web retailers in order to identify the user when he or she next visits the website.

Correlation

A statistical approach to examine the relationship between two variables. Uses an index to describe the strength of a relationship.

Cost per complete (CPC)

CPC is a common term in quantitative research for the price you pay per completed survey. This calculation usually includes the cost of the sample that will be invited to take the study, any associated fielding costs, plus the cost of using the survey software for the collection.

Critical path method (CPM)

A managerial tool used for scheduling a research project. It is a network approach that involves dividing the research project into its various components and estimating the time required to complete each component activity.

Cronbach Alpha

A statistical test used to measure the split half reliability of a summated rating scale. Also known as coefficient alpha.

Cross tabulation (crosstab)

Cross tabulation (or "crosstab" for short) is a technique used to analyse and compare the relationship between two or more variables. Crosstabs involve creating a table that shows the distribution of one variable (the "row variable") across the different categories of another variable (the "column variable"). The cells in the table show the number and/or percentage of respondents who fall into each combination of categories for the row and column variables. Crosstabs help to quickly identify patterns and relationships in the data, and can be used to compare subgroups of the sample or detect differences and similarities between groups.

Cross-sectional research

Research studies that are undertaken once only involving data collection at a single point in time providing a 'snapshot' of the specific situation. The opposite of longitudinal research.

Customer database

A manual or computerised source of data relevant to marketing decision making about an organisation's customers.

CX

CX stands for Customer Experience. Customer experience refers to how customers feel or perceive all aspects (touchpoints) when they buy goods or services or interact with a business.

Data analysis errors

Non-sampling errors that occur when data is transferred from questionnaires to computers by incorrect keying of information.

Data cleaning

Computerised checks made on data to identify inconsistencies in the data and to check for any unexplained missing responses.

Data conversion

The reworking of secondary data into a format that allows estimates to be made to meet the researcher's needs.

Data fusion

Data fusion is the process of combining multiple data sources into a single, comprehensive representation of information. In the context of market research, data fusion can be used to combine different data sets collected from different sources, such as customer surveys, sales data, and social media analytics, in order to gain a more complete and accurate understanding of consumer behaviour and market trends.

Data level

A data level refers to the type of data to be analysed in a hierarchical survey. An example might be a healthcare study, where you have interviewed a number of doctors who have provided information regarding many patients, who have been prescribed a range of therapies/drugs. Analysis of the survey results could then be based at the doctor, patient or the therapy level.

Data mining

Data mining is the process of discovering patterns and knowledge from large amounts of data, by using techniques such as statistical analysis, machine learning, and database systems. The goal of data mining in market research is to extract useful insights and make data-driven decisions that can inform marketing strategies, product development, and other business initiatives.

Data validation

The verification of the appropriateness of the explanations and interpretations drawn from qualitative data analysis.

Degrees of freedom (d.f.)

The number of observations (i.e. sample size) minus one.

Descriptive statistics

Statistics that help to summarise the characteristics of large sets of data using only a few numbers. The most commonly used descriptive statistics are measures of central tendency (mean, mode and median) and measures of dispersion (range, interquartile range and standard deviation).

Desk research

See Secondary research

Dial test

A dial test is a continuous measurement tool used to gather feedback from participants when watching or listening to content media, such as an advertisements, broadcasts or speeches. Participants indicate their continuous level of agreement/like or disagreement/dislike with the content being presented. In the past this used to be a physical dial (hence the name), but is now usually done online using a slider control. The dial or slider movements are captured and plotted over time to provide an insight into the participants' reactions and help assess their emotional engagement with the content being tested.

Dichotomous questions

Questions with only two potential responses (e.g. Yes or No).

Discriminant analysis

Discriminant analysis uses the responses to a set of questions (e.g. attitudes) to predict existing group membership (e.g. segments). The output can then be used to classify future respondents into the same groups using their responses to the same set of questions. Similar to CHAID, but used when all independent variables are scales.

Discriminant validity

A measure of the extent to which the results from a rating scale do not correlate with other scales from which one would expect it to differ.

Discussion guide

A discussion guide is used to structure and direct focus groups or depth interviews with participants. It serves as a guide for the facilitator or interviewer, providing questions, topics, and prompts to cover during the discussion. The purpose of a discussion guide is to ensure that the conversation stays focused on the research objectives, while also allowing for natural and spontaneous dialogue. Also known as a Topic list.

Disproportionate stratified random sampling

A form of stratified random sampling (See Stratified random sampling) where the units or potential respondents from each population set are selected according to the relative variability of the units within each subset.

Do-It-Yourself (DIY) research

DIY, or "Do-It-Yourself," refers to market research methods that can be conducted without the assistance of professional researchers. This type of market research can help businesses to make informed decisions about their products, services, and marketing strategies and is typically more cost-effective and faster than traditional market research. Examples include templated (standardised) methodologies for product and concept testing. DIY is sometimes also know as self-serve research.

Double opt-in (DOI)

A double opt-in (DOI) panellist refers to a person who has agreed to join a research panel and has confirmed their agreement through a two-step verification process. The first step involves providing their email address or other contact information and agreeing to participate in the panel. The second step involves reconfirming their agreement by clicking on a link sent to their email address or by responding to a text message. This process helps to ensure that the panellist has given explicit consent to participate in the research panel and that their contact information is accurate and authentic.

Double-barrelled question

A badly constructed question where two topics are raised within one question.

Doughnut chart

A form of pie chart which allows different sets of data (e.g. for different years) to be shown in the same chart.

Emotion recognition

Emotion recognition is the use of AI algorithms to detect and interpret human emotions, often through analysis of facial expressions, voice, or text.

Ethnography

Ethnography is a research method made for investigating cultural practices, rituals, consumer behaviour, routines and social norms. It helps to unearth previously unseen opportunities through looking at people's worlds in an authentic way, through putting behaviour at the heart of the research. Ethnographic fieldwork often involves observation and empathetic interviewing and uses video recordings to create ethnographic films.

Eye tracking

Eye tracking is a method of measuring and analysing where and how people look at visual information. It uses a device, such as a camera or sensor, to track the movements of a person's eyes as they view stimuli. In market research, eye tracking is used to understand how people interact with visual information, such as advertisements, product packaging, and website layouts. It can provide insights into what elements of a design are attracting attention, where people are spending the most time looking, and which elements are being overlooked.

Face-to-face survey

Research which involves meeting respondents face-to-face and interviewing them using a paper-based questionnaire, a laptop computer, tablet or mobile device.

Facial tracking

Facial tracking is a method of measuring and analysing facial expressions and movements. It uses a device, such as a camera or sensor, to capture and track the movements and expressions of a person's face. Facial tracking can provide information on a person's emotions, such as happiness, sadness, anger, and surprise, as well as more subtle changes in expression, such as furrowed brows, tightened lips, and raised eyebrows. In market research, facial tracking is used to understand how people are responding emotionally to advertising, products, and other stimuli.

Factor Analysis

Factor Analysis is a statistical technique to examine the similarities between items in order to identify a more concise summary of themes. For example, from a list of 20 statements on car imagery, we may identify a factor on reliability, design, performance, environment and image. This can be useful for ordering statements in a presentation so that similar statements are presented together, or for data reduction purposes e.g. to focus a segmentation or to take account of bias in questionnaire design that included too many statements from one particular theme.

Focus group

A focus group is a research method used to gather qualitative data through group discussions. Participants are brought together in a moderated group setting to discuss a specific topic, product, or service. The moderator guides the discussion, encouraging participants to share their opinions, experiences, and attitudes. The information collected from the focus group is analysed to gain insights into the attitudes, beliefs, and motivations of the target audience.

Funnel sequence

A sequence for ordering questions in a questionnaire based on moving from the generalities of a topic to the specifics.

Gabor Granger

Gabor Granger is a pricing technique used to understand price elasticity for set products. Respondents are asked how likely they are to purchase a product at a number of different price points. The purchase intention measures can be converted to estimated take-up scores and plotted to establish which price point is most suitable. The technique can be extended to present products within a competitive context.

Galvanic Skin Response (GSR)

Galvanic Skin Response (GSR) is a method of measuring changes in the electrical conductance of the skin, which can indicate changes in emotional arousal. It uses a sensor that is placed on the skin, typically on the fingers or palms, to measure changes in conductance caused by sweat gland activity. When a person is emotionally aroused, sweat gland activity increases, which can lead to changes in the electrical conductance of the skin. In market research, GSR is used to understand people's emotional reactions to advertising, products, and other stimuli. It can provide insights into the effectiveness of advertisements and product packaging, as well as the emotional impact of branding, pricing, and other marketing strategies.

Gamification

Gamification, the process of adding game-like elements to a research activity, is increasingly being used as a way to engage participants and collect data in a more interactive way. For example, a survey is likely to be more engaging with respondents and the data will be gathered in a more natural and less intrusive way, if the survey is engaging, interactive and fun. This should lead to better quality data.

GANTT chart

A managerial tool used for scheduling a research project. It is a form of flowchart that provides a schematic representation incorporating the activity, time, and personnel requirements for a given research project.

Geodemographic profiling

A profiling method which uses postal addresses to categorise different neighbourhoods in relation to buying power and behaviour.

Grounded theory

A set of analysis techniques that were developed in the 1960s by two medical sociologists, Glaser and Strauss(1967). It is more commonly used by academic researchers rather than marketing research practitioners in areas where little is known about a subject or where a new approach to understanding behaviour is required. It is a systematic method of generating theory and understanding through qualitative data collection and analysis.

Group moderator

The interviewer responsible for the management and encouragement of participants in a group discussion.

Hall test

Research undertaken in a central hall or venue commonly used to test respondents' initial reactions to a product, package or concept. Respondents are recruited into the hall by interviewers stationed on main pedestrian thoroughfares nearby. Sometimes known as a Central Location Test (CLT).

Holecounts

Holecounts are defined as the number of respondents who gave each possible answer to each question in a questionnaire. Sometimes known as frequency distributions or topline numbers. The term originally comes from the time when punched cards were used and data was represented with physical holes punched out of the cards.

Hypothesis

An assumption or proposition that a researcher puts forward about some characteristic of the population being investigated.

IDI

IDI stands for in-depth Interview (sometimes individual depth interview and also simply a depth interview). An IDI is a qualitative research method using a relatively open, discovery-oriented approach to obtain detailed information about a topic from a participant.

IHUT

IHUT stands for "In-Home Usage Test" which is a market research method that allows consumers to test products in their own homes. Participants in an IHUT study are given a product to use and are asked to provide feedback on its features, benefits, and overall satisfaction. This type of research helps manufacturers and marketers understand how a product will perform in a real-world setting and to identify areas for improvement. Also known as a Placement test.

Impact indices

Impact indices measure the impact any independent variable has on changing a dependent variable. Often used when the independent variables are binary (e.g. Yes/No), it can be used for example to evaluate the impact of different product qualities on preference for the product.

Implicit Association Test (IAT)

An Implicit Association Test (IAT) is a psychological measurement technique used to uncover unconscious biases. It measures a respondent's automatic associations between mental representations of objects in memory and evaluations, attitudes, or stereotypes. The test works by measuring the speed with which people categorise various stimuli, such as words, brands, logos, faces, etc.

Implicit assumption

A badly constructed question where the researcher and the respondent are using different frames of reference as a result of assumptions that both parties make about the question being asked.

Independent samples

Samples in which the measurement of the variable of interest in one sample has no effect on the measurement of the variable in the other sample.

Internal data

Secondary data sourced from within the organisation that is requiring the research to be conducted.

Interquartile range

A measure of dispersion that calculates the difference between the 75th and 25th percentile in a set of data.

Interval data

Similar to ordinal data with the added dimension that the intervals between the values on a scale are equal. That means that when using a scale of 1 to 5, the difference between 1 and 2 is the same as the difference between 4 and 5. However, the ratios between different values on the scale are not valid (e.g. 4 does not represent twice the value of 2).

Interviewer bias

Bias and errors in research findings brought about by the actions of an interviewer. This may be influenced by who the interviewer interviews, how the interview is undertaken and the manner in which responses are recorded.

Interviewer Quality Control Scheme (IQCS)

A quality control scheme for interviewers in the UK. The scheme is aimed at improving selection, training and supervision of interviewers and is jointly run by the Market Research Society, the Association of Market Survey Organisations, the Association of British Market Research Companies, the Association of Users of Research Agencies, and a number of leading research companies.

Judgement sampling

A non-probability sampling procedure where a researcher consciously selects a sample that he or she considers to be most appropriate for the research study.

Key Driver Analysis

"The analysis of the relationship between a dependent variable (e.g. brand strength) and one or more independent variables (e.g. brand imagery statements). Its purpose is to determine whether a relationship exists and the strength of the relationship, and used to help prioritise what to focus on.

There are many different statistical techniques that fall under this term, from correlations to Structural Equation Modelling. Different techniques can be applied depending on the objectives, the type of data and how the results will be used."

Kiosk-based survey

A survey often undertaken at an exhibition or trade show using touch screen computers to collect information from respondents. Such computers can be programmed to deliver complex surveys supported by full colour visuals as well as sound and video clips. They can be much cheaper to administer in comparison with the traditional exit survey undertaken by human interviewers.

Kruskal's relative importance analysis

A type of Key Driver Analysis, Kruskals’ relative importance analysis is as an alternative to other techniques such as ordinary regression analysis, which can give misleading results when there is missing data, or when variables are strongly related to each other (which is typical of research data).

Leading question

A badly constructed question that tends to steer respondents toward a particular answer. Sometimes known as a loaded question.

Least filled quota sampling

Least-filled quota sampling is a variation of quota sampling, where the research team attempts to fill the quotas in the most efficient way possible by selecting respondents from the sub-population that has the lowest representation in the sample. This method is used when the research team wants to make sure that the sample is representative of the population, but also wants to minimise the number of respondents needed to fill the quotas.

Least squares approach

A regression procedure that is widely used for deriving the best-fit equation of a line for a given set of data involving a dependent and independent variable.

Lifetime value

The present value of the estimated future transactions and net income attributed to an individual customer relationship.

Likert scale

A Likert scale is a rating scale which requires the respondent to state their level of agreement with a series of statements about a product, organisation or concept. The scale uses the descriptors Strongly agree; Agree; Neither agree nor disagree; Disagree; and Strongly disagree. It is based on a format originally developed by Renis Likert in 1932.

Linear regression

Linear regression is used to find out the relative importance of different drivers in order to re-create a dependent variable. For example, the influence of brand imagery items onto brand appeal.

Logistic regression

Logistic regression is used to find out the relative importance of different drivers in order to re-create a dependent variable when the dependent variable is binary (e.g. Yes/no or Buy/not buy). It is used when the usual linear regression cannot be used, and is particularly useful in propensity modelling.

Longitudinal research

A study involving data collection at several periods in time enabling trends over time to be examined. This may involve asking the same questions on a number of occasions of either the same respondents or of respondents with similar characteristics. Sometimes known as continuous research. Longitudinal surveys are also known as Trackers.

Machine learning

Machine Learning is a subset of AI that involves training computer models to learn from data and make predictions or decisions. Machine learning is increasingly being used in market research in helping to automate much of the classification and analysis of both structured and unstructured data.

Mall intercept interviews

A mall intercept is a type of market research technique where participants are approached in a shopping mall or other public area and asked to participate in a survey or interview. This type of research allows for the collection of data from a diverse and representative sample of consumers in a specific geographic location. Mall intercepts are often used to gather information about consumer attitudes, shopping behaviours, product usage, and other aspects of consumer behaviour. They can be conducted in person, through the use of mobile or tablet devices, and can provide real-time insights into consumer opinions and preferences. Also known as street interviews.

Market Research Online Communities (MROCs)

Market Research Online Communities (MROCs) refer to a group of individuals who are recruited and engaged by a market research company to participate in ongoing research studies, surveys, and discussions about a specific topic or product. These communities often consist of a panel of individuals who have agreed to provide their thoughts, opinions, and feedback on a regular basis over a set period of time. This type of research is often used to gain deeper insights, track brand sentiment, and understand consumer behaviour.

Marketing research

The collection, analysis and communication of information undertaken to assist decision making in marketing.

MaxDiff (Maximum Difference Scaling)

MaxDiff is a technique used to understand relative importance or appeal amongst a list of features/statements. Respondents are asked to compare sets of typically 4-5 features or statements stating which of these are the most and least appealing/important to them. The results give a % appeal/importance score for each item, and can be used to identify which are most popular. Often used to force differentiation, when simple scales may not work as well (e.g. consumers thinking all attributes are important).

MBC (Menu-Based Conjoint)

Menu-based conjoint is a specific type of conjoint analysis able to handle a variety of menu choice situations in which respondents make from one to multiple choices in the process of building their preferred selection. An example situation to which this would apply would be a fast-food restaurant where it’s possible to choose something from the ‘fixed-menu’ section, with some personalisation e.g. choice of side dish - or instead purchasing a series of single items.

Mean

The arithmetic average which is calculated by summing all of the values in a set of data and dividing by the number of cases.

Measures of central tendency

Measures that indicate a typical value for a set of data by computing the mean, mode or median.

Measures of dispersion

Measures that indicate how 'spread out' a set of data is. The most common are the range, the interquartile range and the standard deviation.

Median

When all of the values in a data set are put in ascending order, the median is the value of the middle case in a series.

Mixed Mode Studies

Research studies that use a variety of collection methods in a single survey ( e.g. using the same questionnaire online and face to face) in order to improve response rates.

Mode

The value in a set of data that occurs most frequently.

Monadic testing

Monadic testing is a survey method where you have several concepts to test, but each respondent only evaluates one concept. Monadic testing is commonly used in the early stages of product or concept development, when researchers are exploring different concepts or ideas and want to gather data on their potential appeal to consumers. See Sequential monadic testing for an alternative approach.

Multi-stage sampling

A sampling approach where a number of successive sampling stages are undertaken before the final sample is obtained.

Multiple discriminant analysis

A statistical technique used to classify individuals into one of two or more segments (or populations) on the basis of a set of measurements.

Multiple regression analysis

A statistical technique to examine the relationship between three or more variables and also to calculate the likely value of the dependent variable based on the values of two or more independent variables.

Multivariate data analysis

Statistical procedures that simultaneously analyse two or more variables on a sample of objects. The most common techniques are multiple regression analysis, multiple discriminant analysis, factor analysis, cluster analysis, perceptual mapping and conjoint analysis.

Mystery shopping

Mystery shopping is a method of testing the quality of services provided to customers. It is also used to assess whether employees are following company procedures or industry guidelines. Mystery shoppers are hired to act as regular customers to get an accurate reading of the shopping/buying process. Mystery shopping can be in-store, over the phone, online, or a combination of these to understand the entire customer journey.

Nat Rep

Nat Rep, short for "national representative sample," refers to a research sample that is intended to be representative of the entire population of a country or region. This is typically achieved by selecting participants at random from a sampling frame that includes all individuals or households in the target population. The goal of using a Nat Rep sample is to ensure that the findings from the research can be generalised to the entire population.

Natural Language Processing (NLP)

NLP is a branch of AI that involves understanding and generating human language, such as voice recognition, text generation and sentiment analysis.

Net Promoter Score (NPS)

Net Promoter Score (NPS) is a measure of customer loyalty and advocacy. It is a metric that is used to gauge the likelihood that a customer will recommend a company's products or services to others. NPS is calculated by asking customers to rate their likelihood of recommending a company on a scale of 0 to 10, and then grouping those responses into three categories: Detractors (0-6), Passives (7-8), and Promoters (9-10). NPS is widely used in many industries and sectors, as it is a simple, quick and easy to understand metric that provides a good overall view of a company's customer satisfaction and loyalty.

Netnography

Netnography is a qualitative research method used to study online communities. It involves the systematic observation and analysis of the interactions, communications, and content generated by a group of people on the internet. Netnography can encompass a range of online platforms, including social media, forums, blogs, wikis, and other virtual communities. The goal of netnography is to gain insights into the beliefs, attitudes, behaviours, and motivations of a specific online population, and to understand how they interact and engage with each other in a virtual setting. Sometimes known as online ethnography and webnography.

Neuroscience

Neuroscience is used in market research to provide an accurate and unbiased measure of consumer response. By using techniques such as electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) , researchers are able to measure brain activity and better understand how people respond to marketing stimuli. By examining brain activity, researchers have been able to identify the most effective types of advertisements, marketing messages and design concepts that are more likely to engage and persuade consumers. The main challenge of using neuroscience in market research is the cost of the equipment and specialists required to run it. However, costs are starting to come down - e.g. portable EEG devices are now available and can be used to measure brain activity outside of a laboratory setting.

Nominal data

Numbers assigned to objects or phenomena as labels or identification numbers that name or classify but have no true numeric meaning.

Non-probability sampling

A set of sampling methods where a subjective procedure of selection is used resulting in the probability of selection for each member of the population of interest being unknown.

Non-response errors

An error in a study that arises when some of the potential respondents do not respond. This may occur due to respondents refusing or being unavailable to take part in the research.

Non-sampling error

Errors that occur in a study that do not relate to sampling error. They tend to be classified into three broad types: sampling frame error, non-response error and data error.

Normal distribution

A continuous distribution that is bell-shaped and symmetrical about the mean. This means that in a study, 68.27 per cent of the observations fall within plus or minus one standard deviation of the mean, approximately 95.45 per cent fall within plus or minus two standard deviations, and approximately 99.73 per cent fall within plus or minus three standard deviations.

Null hypothesis

The hypothesis that is tested and is the statement of the status quo where no difference or effect is expected.

Observation

A data gathering approach where information is collected on the behaviour of people, objects and organisations without any questions being asked of the participants.

Omnibus surveys

A data collection approach that is undertaken at regular intervals for a changing group of clients who share the costs involved in the survey's set-up, sampling and interviewing.

Open-ended question

Questions which allow respondents to reply in their own words. There are no pre-set choices of answers and the respondent can decide whether to provide a brief one-word answer or something very detailed and long. Sometimes known as unstructured questions.

Ordinal data

Numbers that have the labelling characteristics of nominal data, but also have the ability to communicate the rank order of the data. The numbers do not indicate absolute quantities, nor do they imply that the intervals between the numbers are equal. and type in their perceptions or comments taking account of other participants' inputs.

Paired interviews

An in-depth interview involving two respondents such as married couples, business partners, teenage friends or a mother and child.

Panel research

A research approach where comparative data is collected from the same respondents on more than one occasion. Panels can consist of individuals, households or organisations, and can provide information on changes in behaviour, awareness and attitudes over time.

Participant observation

A research approach where the researcher interacts with the subject or subjects being observed. The best-known type of participant observation is mystery shopping.

Participant Validation

A validation technique that involves taking the findings from qualitative research back to the participants/ respondents that were involved in the study and seeking their feedback. If the feedback verifies the explanations and conclusions, then the researcher can be more confident about the validity of the findings.

Passive data

Passive data is data that is collected without actively seeking out the information from a respondent/participant. This type of data is generated naturally through actions such as website visits, online purchases, geolocation, and social media activity. It is usually collected through tracking tools such as cookies, beacons, and log files.

Pearson's product moment correlation

A correlation approach that is used with interval and ratio data.

Perceptual mapping

An analysis technique which involves the positioning of objects in perceptual space. Frequently used in determining the positioning of brands relative to their competitors.

Photo sorts

A projective technique which uses a set of photographs depicting different types of people. Respondents are then asked to connect the individuals in the photographs with the brands they think they would use.

Pictogram

A type of bar chart which uses pictures of the items being described rather than bars.

Pilot testing

The pre-testing of a questionnaire prior to undertaking a full survey. Such testing involves administering the questionnaire to a limited number of potential respondents in order to identify and correct flaws in the questionnaire design. Also known as Soft launching.

Placement tests

The testing of reactions to products in the home and where they are to be used. Respondents are given a new product to test in their own home or in their office. Information about their experiences with and attitudes towards the products are then collected by either a questionnaire or by a self-completion diary.

Population of interest

The total group of people that the researcher wishes to examine, study or obtain information from. The population of interest will normally reflect the target market or potential target market for the product or service being researched. Sometimes known as the target population or universe.

Predictive analytics

Predictive analytics is the use of data, statistical algorithms, and machine learning techniques to identify the likelihood of future outcomes based on historical data.

Price Sensitivity Management (Van Westendorp)

Price Sensitivity Measurement (PSM) is a technique used to understand price preferences. Respondents are asked at what price they would consider a product to be too expensive, too cheap, etc. The optimal price point is determined to be where an equal proportion of respondents have said “too cheap” and “too expensive”. Sometimes known as Van Westendorp.

Primary data

"Data collected by a programme of observation, qualitative or quantitative research either separately or in combination to meet the specific objectives of a marketing research project.
"

Principal Component Analysis (PCA)

PCA is a descriptive method of multivariate analysis that is used to analyse numeric questions (metrics). It converts a set of observations of possibly correlated variables, into a set of values of linearly uncorrelated variables called Principal Components. This technique allows you to reduce the number of correlated variables and decrease the redundant information.

Probability sampling

A set of sampling methods where an objective procedure of selection is used, resulting in every member of the population of interest having a known probability of being selected.

Programme evaluation and review technique (PERT)

A managerial tool used for scheduling a research project. It involves a probability-based scheduling approach that recognises and measures the uncertainty of project completion times.

Projective questioning

Sometimes known as third-party techniques, this is a projective technique that asks the respondent to consider what other people would think about a situation.

Proportionate stratified random sampling

A form of stratified random sampling (See Stratified random sampling) where the units or potential respondents from each population subset are selected in proportion to the total number of each subset's units in the population.

Purchase intent scales

A scaling approach which is used to measure a respondent's intention to purchase a product or potential product.

Qualitative research

An unstructured research approach with a small number of carefully selected individuals used to produce non-quantifiable insights into behaviour, motivations and attitudes.

Quantitative research

A structured research approach involving a sample of the population to produce quantifiable insights into behaviour, motivations and attitudes.

Questionnaire design process

A stepped approach to the design of questionnaires.

Quota sampling

Cells or subsets are selected based on the characteristics of the population being studied, such as age, gender, income etc. A numerical quota in each cell is established and the researcher carries out sufficient interviews in each cell to satisfy the quota.

Range

A measure of dispersion that calculates the difference between the largest and smallest values in a set of data

Ratio data

Actual 'real' numbers that have a meaningful absolute or zero. All arithmetic operations are possible with such data.

Regression

A statistical approach to examine the relationship between two variables. Identifies the nature of the relationship using an equation.

Reliability of scales

Refers to the extent to which a rating scale produces consistent or stable results. Stability is most commonly measured using test-retest reliability and consistency is measured using split–half reliability.

Research brief

A written document which sets out an organisation's requirements from a marketing research project. This provides the specification against which the researchers will design the research project.

River sampling

River sampling invites respondents to take a survey via online banners, ads, promotions, offers and invitations placed on a variety of websites and social media. Once their attention is caught and they click on a link, they are asked several screening questions and then routed to a survey based on their answers. The research team have no idea who will respond. They don’t have any pre-existing socio-demographic information for these respondents, and they can’t contact them again them upon survey completion.

Role playing

A projective technique which involves a respondent being asked to act out the character of a brand.

Routing

When designing a questionnaire, routing refers to the survey logic, that allows you to change / limit the content based on previous answers given by respondents. Sometimes known as branching.

Sample

A subset of the population of interest.

Sampling error

The difference between the sample value and the true value of a phenomenon for the population being surveyed. Can be expressed in mathematical terms: usually the survey result plus or minus a certain percentage.

Scaling questions

Questions that ask respondents to assign numerical measures to subjective concepts such as attitudes, opinions and feelings.

Screening criteria

Screening criteria help to ensure that the participants in a study are relevant, qualified, and representative of the target population. Screening questions are usually asked at the start of the study and may include demographics as well as questions such as brand usage and product ownership. Screening criteria are managed with quota sampling (see quota sampling).

Screening questionnaire

A questionnaire used for identifying suitable respondents for a particular research activity, such as a group discussion.

Secondary data

Information that has been previously gathered for some purpose other than the current research project. It may be data available within the organisation (internal data) or information available from published and electronic sources originating outside the organisation (external data).

Segmentation

The process of dividing markets into groups with people or occasions that are similar to each other, but different to the other groups.

Semantic differential scales

A scaling approach which requires the respondent to rate a brand or concept using a set of bipolar adjectives or phrases (e.g. helpful and unhelpful; friendly and unfriendly). Each pair of adjectives is separated by a seven-category scale with neither numerical nor verbal labels.

Semiotics

Semiotics is the study of symbols and signs and how they communicate meaning. Semiotics in market research usually involves analysing the visual and linguistic elements of product packaging, advertising, and branding to understand the cultural and emotional associations consumers have with these symbols. The goal is to uncover deeper insights into consumer attitudes and behaviour, including why they choose certain products or brands over others. Semiotics helps researchers to understand the unconscious motivations behind consumer choices and how these choices are influenced by cultural and societal values.

Sensory testing

Sensory testing is a method of evaluating the characteristics of a product or food by measuring how it is perceived by the senses of smell, taste, sight, touch, and sound. Sensory testing is used to evaluate the attributes of a product, such as its appearance, flavour, texture, and aroma, in order to assess its acceptability and quality. It can be used to determine how a product compares to its competitors, or to identify opportunities for product improvement. Sensory testing can be conducted in laboratory or real-life settings, and can be done with a small group of trained testers or a large panel of consumers.

Sentence completion

A projective technique which involves providing respondents with an incomplete sentence or group of sentences and asking them to complete them.

Sentiment analysis

Sentiment analysis looks at the emotion expressed in a text. It’s a form of text analytics that uses natural language processing (NLP) and machine learning. It is commonly used to analyse survey responses, user/customer feedback and product reviews. Sentiment analysis is used to determine whether a given text contains negative, positive, or neutral emotions.

Sequential monadic testing

Sequential monadic testing, is a survey method whereby each respondent provides feedback on two or more concepts in the same survey. The concepts should be presented in a random order to avoid order bias. Sequential monadic testing is commonly used in the early stages of product or concept development, when researchers are exploring different concepts or ideas and want to gather data on their potential appeal to consumers. See Monadic testing for an alternative approach.

Share of Shelf (SOS)

Share of Shelf (SOS) is a metric used to measure the proportion of space a product or brand occupies on a retail shelf compared to its competitors. This metric can help understand the visibility and accessibility of a product in the retail environment.

Share of Voice (SOV)

Share of Voice (SOV) is a measure of the proportion of advertising a brand or product receives compared to others in the same category. It's a way to understand how often a brand is advertised in relation to its competitors.

Share of Wallet (SOW)

Share of wallet (SOW) is a metric used to measure a company's market share in terms of customer spending. It represents the proportion of a customer's total spending on a given product or service category that is going to a particular company. Share of wallet is often used in customer relationship management and customer retention strategies, and can be a key indicator of a company's competitiveness in a particular market.

Shelf impact testing equipment

Used to determine the visual impact of new packaging when placed on shelves next to competitors' products.

Simple random sampling

A probability sampling method where every possible member of the population has an equal chance of being selected for the survey. Respondents are chosen using random numbers.

Simulated test markets

A research approach used to predict the potential results of a product launch and to experiment with changes to different elements of a product's marketing mix. Rather than testing in retail stores, simulated test markets rely on simulated or laboratory-type testing and mathematical modelling.

Snowball sampling

A non-probability sampling procedure where additional respondents are identified and selected on the basis of referrals of initial respondents. It tends to be used where the population of interest is small or difficult to identify.

Social media analytics

Social media analytics is the use of data from social media platforms to understand consumer behaviour and sentiment, as well as to track the performance of campaigns or products.

Spearman's rank-order correlation

A correlation approach for ordinal data.

Split half reliability

Measures the internal consistency of a summated rating scale and refers to the consistency with which each item represents the overall construct of interest. The method involves randomly dividing the various scale items into two halves. High correlations between the two halves suggests internal consistency of what is being measured.

Standard deviation

A measure of dispersion that calculates the average distance that the values in a data set are away from the mean. The standard deviation of different sets of data can be compared to see if one set of data is more dispersed than another.

Stapel scales

A scaling approach which is a variation of the semantic differential scaling approach. It uses a single descriptor and 10 response categories with no verbal labels.

Stimulus materials

Materials used in group discussions and individual depth interviews to communicate the marketer or advertiser's latest creative thinking for a product, packaging or advertising to the respondents.

Stratified random sampling

A probability sampling procedure in which the chosen sample is forced to contain potential respondents from each of the key segments of the population.

Structural Equation Modelling

Structural Equation Modelling (SEM) is a statistical technique for testing and estimating causal relationships, using a combination of statistical data and qualitative causal assumptions. Factor analysis, path analysis and regression all represent techniques used within SEM, which also allows the construction of variables which are not measured directly. As an example, it can be used to model and understand the relationship between different aspects of customer satisfaction and how these explain customer loyalty.

Structured observation

A research approach where observers use a record sheet or form to count phenomena or to record their observations.

System 1 & System 2 thinking

Psychologists and behavioural economists have proposed a dual system of the mind: System 1 and System 2. In simple terms these can also be thought of as intuitive and deliberate thought. System 1 operates automatically and quickly, with little or no effort and no sense of voluntary control. System 2 thinking is more deliberate, conscious, and reflective with more effortful thought processes. Researchers have devised a number of methods that specifically measure System 1 responses (e.g. Implicit Association Test, Facial Coding, Galvanic Skin Response, etc).

Systematic sampling

A probability sampling approach similar to a simple random sample but which uses a skip interval (i.e. every nth person) rather than random numbers to select the respondents.

T-test

A hypothesis test about a single mean if the sample is too small to use the Z test.

Tabular method of analysis

A method for analysing qualitative research data using a large sheet of paper divided into boxes.

Test-retest reliability

Measures the stability of rating scale items over time. Respondents are asked to complete scales at two different times under as near identical conditions as possible. The degree of similarity between the two measurements is determined by computing a correlation coefficient

Triangulation

Using a combination of different sources of data where the weaknesses in some sources are counterbalanced with the strengths of others. The term triangulation is borrowed from the disciplines of navigation and surveying, where a minimum of three reference points are taken to check an object's location.

TURF

TURF stands for Total Unduplicated Reach and Frequency, and is used for providing estimates of media or market potential and devising optimal communication and placement strategies. If we take appeal of ice cream flavours as an example, TURF analysis can identify the number of users reached by the combination of ice cream flavours in a client’s range, as well as how frequently they will be consumed - very useful for deciding which flavours of ice cream to have in a range.

Type I error

Rejection of the null hypothesis when it is actually true.

Type II error

Failing to reject the null hypothesis when it is actually false.

Universe

See Population of interest.

Unstructured questions

See Open-ended questions.

User generated content

Online material such as comments, profiles photographs that is produced by end users.

UX

UX stands for User Experience, which refers to the overall experience of a person using a product or service, including the design, functionality, and ease of use. It encompasses the emotions, attitudes, and perceptions that a person has while interacting with a product or service. In market research, UX research is used to understand how people interact with a product, website, or service, and to identify areas for improvement. It can involve usability testing, where users are asked to perform specific tasks while researchers observe and record their interactions. It can also include surveys, interviews, and other methods to gather qualitative and quantitative data on users' attitudes and perceptions.

Validity

Whether the subject requiring to be measured was actually measured.

Verification

Data verification in market research refers to the process of checking and verifying the accuracy and reliability of data collected during research. This includes reviewing data for errors, inconsistencies, and missing information. The goal of data verification is to ensure that the data collected is complete, accurate, and free of any biases or anomalies that could affect the validity of the research findings. Data verification may involve reviewing survey responses and checking for "speeders" (respondents giving answers too quickly) and straight-liners (respondents answering grid/rating questions all in the same way) and plausible verbatim responses.

Viewing rooms

Specialist facilities/locations for group discussions. They are set out in the form of a boardroom or living-room setting with video cameras or a large one-way mirror built into one wall. Some are owned by research agencies, but the majority are independent and available to anyone willing to pay the hourly room-hire rates.

Virtual reality

Virtual reality can be used in market research to create immersive experiences for consumers, allowing researchers to gather data on how consumers interact with products and brands in a simulated environment (e.g. inside a virtual store or car interior).

VoC

VoC stands for Voice of Customer. This is a research method that is used to collect customer feedback. A well-planned, ongoing VoC program can help companies capture how their customers feel about their business, product, or service, giving insights that can help create a stronger customer experience. VoC is usually a key component of an overall customer experience (CX) strategy - see CX above.

Web 2.0

Web 2.0 refers to the second generation of the World Wide Web, characterised by the shift from static websites to dynamic, interactive, and user-generated content. The term was popularised in the early 2000s and refers to a set of technologies and design principles that allow for greater collaboration, communication, and information sharing among internet users. Key features of Web 2.0 include social networking, blogs, wikis, podcasts, video sharing, and online communities. These technologies and platforms allow users to create, share, and comment on content, as well as connect and interact with each other in new and innovative ways. Web 2.0 has had a profound impact on the way people consume and create content online and has been a major driving force behind the growth of the internet and the digital economy.

Web 3.0

Web 3.0 refers to the next generation of the World Wide Web, characterised by the integration of artificial intelligence, machine learning, and the decentralised web, in an immersive manner, in which users have more control over their data and online identity. It is aimed at creating a more intelligent, rich and interconnected web that allows for a more personalised, seamless, and secure user experience.

Web scraping

Web scraping can be used in market research to gather large amounts of data from the internet, such as information on prices, product reviews, or social media sentiment. This data can then be used to gain insights into market trends, consumer behaviour, and competitors' strategies.

Web survey

see CAWI

Weighting

The process of adjusting the value of survey responses to account for over- or under-representation of different categories of respondent. Weighting is used where the sample design is disproportional or where the achieved sample does not accurately reflect the population under investigation.

Word association tests

A projective technique that involves asking respondents what brands or products they associate with specific words. In addition to the direct outputs regarding brand imagery, it is also a very useful technique for building rapport within a group discussion and getting everybody contributing and involved.

Word cloud

A visual depiction of words used by respondents in qualitative research, open-ended survey questions or the content appearing on social media or reviews. The font size of the words, is determined by the number of times a word has been used. The more that a topic or word is talked about, the bigger it is.

Z test

A hypothesis test about a single mean where the sample size is larger than 30.

This glossary has been edited and updated by Richard Collins, and includes content reproduced from the book, Market Research: An Integrated Approach (2012) by Alan Wilson. For use of this information more widely, copyright permission must be sought by the user directly from MRS.

Glossary - What have we missed?

Get the latest MRS news

Our newsletters cover the latest MRS events, policy updates and research news.

Sign up A white arrowA black arrow