A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

A
AAPOR	(See American Association for Public Opinion Research)
ABC	(See Audit Bureau of Circulation)
absolute precision	Absolute precision is precision expressed in units of the variable measured (See precision. Compare relative precision.)
absolute value	The numerical value which ignores the algebraic (+ or -) sign.
absolute zero	A natural zero point, e.g., the point of origin, characteristic of a ratio scale (used to measure height, weight, etc.)
acceptance region	A term used in significance (hypothesis) testing representing values of the test statistic for which we accept (do not reject) the null hypothesis.
accuracy	The degree to which a sample statistic would correspond to the population parameter it is meant to estimate if there were no random error. Accuracy is high when bias is low.
acquiescence bias	A respondent bias reflecting the respondent's tendency to concur with all questions about a topic or position.
acquiescent response set	A biasing condition that can be set up by a series of questions that beg a "YES" response.
ad hoc research	Research designed to help solve a problem of immediate interest. (Compare tracking research)
additive effect	When the joint influence of two or more variables may be determined by simply adding together the individual effects of each of the variables.
ADI	Area of Dominant Influence; a geographic market definition defined by Arbitron for the purpose of analyzing markets based on the predominant television broadcast signals. Arbitron no longer defines ADIs and Nielsen’s Designated Market Areas (DMAs have replaced ADIs as the preferred market designations.
advertising development research	Advertising development research refers to those investigations done prior to the launch of a campaign, to help define the audience and their interests, and to provide input to the creative process
advertising effectiveness research	Advertising effectiveness research refers to research done to measure consumers’awareness of and comprehension of advertising.
advertising tracking study	(See tracking study)
affective component	Acceptance or avoidance responses. Hedonic scales measure affective responses, i.e. degree of pleasantness.
after-only design	A research design involving measurements only after the treatment (e.g. ad campaign) is introduced. The design generally involves a control group and no measure is taken on either the test or the control group before exposure of the independent variable.
agree-disagree scale	(Seescales)
AID	(SeeAutomatic Interaction Detection)
aided recall	Aided recall refers to the mention of a brand name in a brand or advertising recall question; e.g., "Have you ever heard of Microsoft as a provider of word processing software?" Aided recall questions are typically asked after unaided recall questions. (Also calledprompted recall. See recall. Compareunaided recall)
alpha error	One of two types of errors encountered in statistical hypothesis testing. Also referred to as the type I error, it represents the probability of rejecting the null hypothesis when it is true.
alternative hypothesis	The counterpart of the null hypothesis in statistical significance testing. Sometimes called the business hypothesis. The alternative hypothesis is accepted when the null hypothesis is rejected.
AMA	(SeeAmerican Marketing Association)
ambiguous question	An ambiguous question is one that implies multiple questions or could be answered correctly in at least two different ways.
ANA	(SeeAssociation of National Advertisers)
analysis of dependence	Analysis of dependence refers to any of a number of multivariate analytical methods in which one or more variables (dependent variables) are assumed to be dependent upon other variables (independent variables). (Compareanalysis of interdependence)
analysis of interdependence	Analysis of interdependence refers to any of a number of multivariate analytical methods in which a number of variables examined but none are presumed to be dependent on others. (Compareanalysis of dependence)
analysis of variance	Analysis of variance, frequently abbreviated ANOVA, is a statistical test used with interval data to determine if multiple samples come from populations with equal means. Like chi-square, ANOVA tests for significant variation between groups or samples. However, ANOVA requires interval data and signifies differences in sample means.
analytical matrix	An analytical matrix is a row/column format in which the columns are categories (e.g. under 40 years old, 40 years old or older) and the rows are possible answers (e.g. yes, no) Sometimes used in focus group analysis where vertical columns are separate groups and horizontal rows are responses to each question. Allows looking across columns for consistency or differences.
anonymity	Keeping the identity (if known) of a respondent in a study strictly confidential.
ANOVA	(Seeanalysis of variance)
a priori segmentation	A segmentation research strategy in which the market is segmented using a criterion preselected by the researcher. Other characteristics such as demographics are then used to describe the segments.
area sampling	Method of sampling in which a georgraphic area is divided into mutually exclusive and exhaustive subsets and a random sampling of areas is then selected. If all households in the selected areas are used, this is called 'one-stage area sampling.' A 'two-stage area sample' involves a random selection of households within the selected areas.
ARC	(SeeAdvertising Research Foundation)
ARF	Advertising Research Foundation
arithmetic mean	An average, obtained by dividing the sum of a set of observations by the total number of observations in the set.
ASA	(SeeAmerican Statistical Association)
association	The extent to which two or more variables tend to 'go together'. May be expressed in terms of an index of association (e.g., correlation coefficient) or a functional relationship (e.g., regression equation). However, association does not imply causation.
attitude	An attitude is a generalized predisposition to respond positively or negatively to an object or class of objects such as symbols, slogans, persons, products, institutions, etc.
attitude scale	A scale designed to measure a person's attitude towards an object or a class of objects. Examples of attitude scales include the hedonic scale (like extremely thru dislike extremely), the Semantic Differential (rude - courteous), the Likert Scale, etc.
attribute	A quality, character or property used to describe an idea or object being evaluated.
audimeter	An electronic device which is wired to the television receiver and utilized by the A.C. Nielsen Company to record automatically the times the TV set is turned on and off and the stations to which it is turned.
audit	An audit is a review of specific aspects of a market, such as products on store shelves, to determine the number of facings and competitive mix, or a complete advertising audit. Also a broad-based survey, such as a market audit, intended to determine a wide range of basic market facts about a product category.
average	(See mean)
awareness	Awareness refers to the consumer’s knowledge of a brand, advertising execution.
awareness and attitude research	A fundamental type of marketing research often used as a baseline for tracking companies’ positions in the market. Measures awareness of the company and competitors and attitudes toward the company or product.
B BACK TO TOP
back-to-back focused group interviews	Back-to-back focused group interviews are interviews conducted so that one occurs shortly after another and on the same day.
banner	The column or heading over a data set in a computer cross-tab table.
banner point	In cross-tabulations, banner points are the vertical headings which describe data breakouts. For example:college grads,income over $50,000,design engineers. Similar to the verbal heading at the top of a spreadsheet column.
base, sample base, base over	The number over which percentages, averages or other findings are calculated. For example, positives or negatives about a specific company will often be based over those who are aware of the company. Often a subset of the entire sample.
baseline	An initial research project against which future changes are benchmarked
basic research	Basic research is research undertaken to advance the knowledge of methodologies and techniques of research. (Compare applied research.)
before-and-after design	Experimental design in which a variable (e.g. awareness) is measured on a sample (the before stage); then after potential exposure to a stimulus (e.g. newspaper ad), awareness is measured again (the after stage). May have a control group in addition to the experimental group.
before-and-after with control group	(See before-and-after design)
behavioral correlates	Behavior variables (e.g. brand usage, purchase intent) that are correlated with consumer actions and, hence, used to attempt to predict or forecast such actions.
benchmark	A tracking wave or survey designed to measure against the previous baseline. All repetitions of a given study after the first are benchmarks. Also used to describe the ability to measure one company against an industry standard or its own past measures.
benefit segmentation	Breaking out specific subsets of a population based on their needs or desires rather than on measurements such as life-style or demographics.
beta coefficient	The standardized regression coefficient. Can be found either by using standardized data in regression analysis or by mulitplying the regression coefficient for a particula predictor variable by the ratio of the standard deviations of that predictor variable to the criterion variable. Gives a very rough indication of the importance of predictors.
beta weights	(See beta coefficients)
bias	Bias is the research equivalent of sin. Clients and researchers are consistently on the lookout for anything which can corrupt or "bias" data and lead to erroneous conclusions.
blind study	Research where the sponsor/client/brand is masked from respondents or researchers.
bimodal distribution	Different from the familiar bell-shaped curve or normal distribution. Bimodal means that there are two modes of responding instead of having the data cluster around a central point. A bimodal distribution might be seen in a population where people either strongly prefer a product or strongly reject a product - half the population might be rating the product 1-2-3 on a ten-point scale, and the other half of the population might be rating it 8-9-10.
binomial distribution	The distribution of the probability of a specified number of successes in a given number of independent trials, in each of which the probability of success is the same. Useful in performing certain statistical significance tests on dichotomous data.
bipolar scale	Scale frequently used in attitude measurement in which the end points are opposites (e.g. sweet - sour, friendly - unfriendly, etc.).
bivariate analysis	The simultaneous analysis of two variables, such as income and brand expenditure
blind test	A blind test is a taste test in which the samples have been coded so that the brands or types are not known by the respondent. (See alsodouble-blind test)
block	Normally a rectangular piece of land, bounded by four streets. However, a block may also be irregular in shape or bounded by railroad tracks, streams, or other features Blocks do not cross the boundaries of counties, census tracts or block numbering areas (BNA's). Census data are tabulated by block in all urbanized areas but much information is suppressed to protect the confidentiality of census information. A block is the smallest level of census geography.
block group	A combination of numbered blocks that is a subdivision of a census tract or untracted area. Block groups are defined in all areas for which block statistics are prepared. Block groups (BG's) are defined as that set of blocks sharing the same first digit within a census tract or non-tracted area. For example, Block Group '3' within a particular census tract would include any blocks numbered between 300 and 399. Block groups may cross and be split by the boundaries of MCD's, places, and congressional districts. When this occurs statistical summaries are provided for each component or part, but not for the whole BG. Block groups are represented by a one digit numeric code.
block sampling	A method of selecting a sample based on the fact that individuals live in city blocks (or similar land segments) and that these blocks can be easily tested and sampled.
branching	(See skipping)
brand preference and awareness research	Studies conducted to determine the awareness of specific brands or companies and the degree to which the purchasing population prefers or doesn’t prefer a specific product.
brand switching	Brand switching refers to a consumer’s use of more than one brand within a product category. A consumer who uses one brand predominantly, rarely switching, is said to be brand-loyal.
briefing	Typically the process of orienting data collection personnel (interviewers) on a new project.
C BACK TO TOP
call record sheets	A paper and pencil method of tracking results of calls made to homes or businesses. Not the actual survey but a record of terminations, those who were not at home, attempts, language barriers, refusals and so forth. Becoming less common as CATI (see definition) systems automate this record keeping.
callback	Recontacting a respondent to clarify or verify survey information. Done to check quality, correct errors or expand an initial finding.
canonical correlation	Canonical correlation is one approach to correlation analysis
categorical data	Responses which have no numeric relationship to one another. For example, categorizing respondents as brown-eyed, blue-eyed or green-eyed. Same as nominal data.
causal inference	An attempt to determine the nature of the cause-and-effect relationship among two or more variables through the use of experimentation.
causal relationship	A relationship between two or more variables characterized by the condition that changes in one or more of the variables causes changes in another variable.
CATI	(Computer-Aided Telephone Interviewing.) Questionnaires are shown on a computer screen, and the questioning is directed to some degree by computer. Results are usually entered directly into a computer database. Can be administered solely by computer, or the computer can aid a human interviewer.
cell	Often used to refer to a subset of a survey sample or experimental design. "Since there are only ten males in the cell, the sample base is too small to draw a conclusion."
census	A survey designed to reach all units of the population in question. (Compare sample).
census divisions	The Bureau of the Census divides the entire U.S. into nine divisions defined by groups of states. These divisions are: New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central West South Central, Mountain and Pacific. These divisions should not be confused with the four Census Regions, since they are comprised of different state groupings,
census region	The Bureau of the Census divides the entire U.S. into four regions. These regions are: Northeast, North Central (now named Midwest), South and West. Each of these regions is defined by groups of states. These regions should not be confused with the nine Census Divisions, since regions are made up of larger state groupings. The nine Census Divisions can be combined to form four Census Regions.
census tract	Geographic area including households with uniform social and economic characteristics. Tracts generally have between 2,500 and 8,000 residents.
central limit theorem	The central limit theorem states that as the sample size (n) increases, the sampling distribution of means of samples taken from the population of practically any distribution will approach the normal distribution with mean µ (equal to the population mean and variance equal to F2/n, where F2 is the population variance.
central location test (CLT)	Using a central site for conducting interviews. May be either a convenience sample such as traffic in a shopping mall or pre-recruited to specific criteria and invited to an interviewing location.
central tendency	There are three basic indicators: mean or average, mode (most common) and median. The median is the middle response. For example, if there are 19 responses, the median is the tenth response up from the lowest or the tenth down from the highest.
CFA	Common Factor Analysis (Seefactor analysis)
chi-square distribution	A skewed distribution whose shape depends on the number of degrees of freedom. As the number of degrees of freedom increases, the distribution becomes more symmetrical.
chi-square test	A test of statistical significance used for analyzing frequency distributions or contingency tables.
choice modeling	A related technique to conjoint, used when there is a strong need for the results to map closely to actual market response. Also called discrete choice analysis.
clarification	An interviewing process used to improve the quality and clarity of data for open-ended questions. For example, "Tell me more about what you meant by your response."
classification/classification questions	Classification questions are those that ask for respondents’ demographic or other characteristics (age, income, household size, dollars spent in a product category, etc.) so that their responses may be grouped with the responses of other respondents like themselves. Typically classification questions occur late in a questionnaire.
cleaning/data cleaning	The process of taking survey computer printouts and checking data for logical consistency or error. Increasingly automated by computer. Also, dual entry of data and comparing records to eliminate mistakes.
closed-ended	A question with a fixed response set. Perhaps a rating scale, "yes" or "no" choices and so forth. As opposed to an open-ended, free-responding question where the respondent’s verbatim answers are recorded.
cluster analysis	Techniques designed to classify individuals into a relatively small number of exclusive groups. A goal of cluster analysis is to maximize likeness within groups and differences between them to create an understandable topology of the market.
cluster sample	A cluster samples is one drawn in a cluster so the researcher can collect information from that cluster. Example Sales districts are randomly chosen and then all buyers within the chosen district are interviewed. Cluster samples are generally cost-efficient
CMSA	(See Consolidated Metropolitan Statistical Area)
code	A code is a mathematical symbol which represents a particular response to a particular question. For example, in a yes/no question, the response ‘yes’ may be coded ‘1,’ the response ‘no’ coded ‘2,’ and no response coded ‘9.’
codebook	A set of question responses and their associated computer code numbers. Used to assign categories to answers given to each question on a survey questionnaire.
coding	Coding is the process of assigning codes to the acceptable responses in a survey.
coefficient of determination	The percent of the variability in the dependent variable explained by the independent variable.
cohort analysis	Analysis of the activity of a cohort over an extended time period.
collinearity	Collinearity refers to a degree of association between two or more of the independent variables in a regression analysis.
comparative scale	(See scale)
completion rate	The percent of qualified respondents completing an interview or study.
concept test	A concept test is one which presents the idea of a product or communication to consumers to observe their reactions at the concept (pre-production) level
concomitant variation	(See causality)
confidence Interval	The range around a survey result for which there is a high statistical probability that it contains the true population parameter.
confounded	An independent variable and an extraneous variable are confounded when their effects on the dependent variable cannot be distinguished from each other. Often a comparative study using a control group is used to avoid confounding variables.
conjoint analysis	A multivariate technique used to quantify the value that people associate with different levels of product/service attributes. Respondents trade product attributes against each other to establish product (brand) preference and the relative importance of attributes. Based on utility theory and consumer rationality. Better for functional than fashionable brands.
consumer	the term consumer is frequently used to describe people generally, under the assumption that their importance to the researcher is invested in their ability to and interest in consuming products or services.
contact rate	The proportion of respondents who were actually contacted divided by the number of contacts which were attempted.
content validity	The extent to which the questions in the instrument are an adequate representation of the field being investigated. Usually, evaluated by expert opinion.
contrived observation	(See observation)
control group	An experimental sub-group where the independent variable is not manipulated. It is used as a benchmark for comparison with the experimental group.
co-op fee (co-operation fee)	A co-op fee is stipend (money or gift) given to respondents for their participation.
cooperation rate	Refers to the proportion of all contacted households where respondents agree to participate in a survey.
copy testing	Copy testing is a form of advertising testing in which the goal is to measure respondents’ recall and comprehension of the advertising as well as the ad’s uniqueness and believability
correspondence analysis	Maps results using categorical data such as "name anyone" responses. Generally more flexible and easier on respondents than classic multidimensional scaling.
cost per interview (CPI)	The dollar cost of completing one interview in a survey research project.
CPI	(Seecost per interview)
CPS	(See Current Population Survey)
criterion validity	The degree to which a measurement instrument can predict a variable that is designated a criterion.
criterion variables	The variables being predicted or explained in a study. Also known as the dependent variable.
cross-cultural analysis	(See culture)
cross-tabulation	A typical technique used to display research data, similar to a spreadsheet. Used as a basis for analyzing most surveys. Cross-tabs normally sort responses by type (young, aware, heavy users) and display the results in a data matrix.
culture	The total sum of learned beliefs, values and customs that serve to regulate the consumer behaviour of members of a particular society.
Current Population Survey (CPS)	The survey conducted by the Census Bureau which monitors changes between the decennial censuses. Conducted monthly to 60,000 households.
D BACK TO TOP
DAR (day after recall)	The awareness or message content recalled within 24 hours. Normally a measure used for television or radio advertising.
data collection	Also called field work or interviewing. The process of actually collecting market research data. Conjoint Analysis.
data entry	Entering survey data into a database so that it can be analyzed. May be done from a CATI terminal or entered from paper records and questionnaires.
data processing	Organization of data for the purpose of producing desired information; involves recording, classifying, sorting, summarizing, calculating, disseminating and storing data. (The February issue of Quirk's Marketing Research Review has a section devoted to data processing service providers.)
debriefing	The presentation of research findings to clients.
demand bias	Research corruption created when respondents know or suspect that researchers have a particular agenda. One example of demand bias is the effect of knowing who is sponsoring the research.
demographics	The characteristics of respondents participating in research. Typical examples include age, income, sex and so forth.
dependent variable	(See variable)
depth interview	(See in-depth interview)
descriptive data analysis	As opposed to statistical testing. Observing findings, data distributions and relationships between data without applying statistical tests. Reporting and summarizing on these observed findings. Descriptive Statistics.
design	The framework or plan of a research project providing a detailed specification of the best strategy for collecting the relevant information to solve the problem at hand.
design effect	That part of the observed variance in the sampling distribution of an estimate due to certain aspects of the sample design.
Designated Market Area (DMA)	A television market, as defined by NPD/Nielsen, a firm which measures TV audiences.
diadic	Paired or face-to-face. An example of a diadic design would be a paired comparison taste test where two versions of a new french fry might be tested against each other. As opposed to monadic where one version of the french fry would be tested independently of any comparison.
diary	The record kept by a respondent of purchase behavior, likes and dislikes and so forth, over a period of time. Typically diaries are a product of research panels who create them.
dichotomous question	A question that has only two possible responses. An example might be, "Will you vote for candidate A or candidate B?"
partial termination	When a respondent ceases to be involved in an interview and quits the process.
discriminant analysis	An analysis technique where the dependent variable is non-metric (ie nominal or ordinal in nature) and the independent variables are metric (ie interval or ratio in nature).
discriminant validity	The lack of association among constructs that are supposed to be different.
discussion guide	Typically five to ten pages of questions used in a focus group. The questions proceed in chronological order. Discussion guide design is usually worked out between the researchers and the client prior to conducting groups.
disk-by-mail	A research technique where a self-prompting questionnaire is sent to respondents on floppy disk. Respondents use the disk to answer the survey on their computer and send the completed disks back to the researchers. Core Strengths.
disproportionate stratified sample	A disproportionate stratified sample is an appropriate sampling method when certain segments of a population are seen as more important than others, as varying more, or as more expensive to sample. For example, a health insurance company surveys its corporate customers, oversampling firms with larger memberships to reflect their true influence.
distribution	The pattern and frequency of responses to a given question. Descriptive Statistics. Testing Significance.
DK	(See don’t know)
DMA	(See Designated Market Area)
don’t know	The response a person gives when he doesn’t have any other answer. ‘Don’t know’ responses should not be confused with "no answer" responses. In the former case, a respondent was asked a question which he could not answer; in the latter, the respondent was not asked the question.
door-to-door interviewing	Door-to-door interviewing is the practice of going into neighborhoods and to individuals’ homes to interview them in some sort of systematic fashion. The Gallup Poll was originally conducted in respondents’ homes. In the United States, door-to-door interviewing has largely given way to other interviewing methodologies; in many other countries, door-to-door interviewing is still very common. (Also calledin-home interviewing)
double-barreled question	Example: "What did you like about the new product, and how would you improve it?" Also ambiguous question.
double-blind test	A double-blind test is a taste test in which the brand or product types are hidden from both the respondent and the test administrator. (Seeblind test)
drawing a sample	The process of determining in a random or systematic fashion who will be candidates to participate in research.
dual moderator focused group interview	A technique using two researchers to conduct group interviews. Typically used for complex or difficult topics to provide greater depth, breadth and quality of information.
dwelling unit	The term dwelling unit refers to a place where a person or persons live; it may be rented or owned, single-family or multiple-family,