Bookmark and Share
               

GLOSSARY OF MARKET RESEARCH TERMS

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

This is a glossary of the most often used market research phrases and words.  It is meant to help the reader communicate with suppliers, evaluate research proposals and interpret reports.

A

Awareness:  usually used to describe the percentage of people in a market who recognize a brand name, advertising campaign or some other element of the marketing mix.  see unaided awareness, aided awareness, top of mind awareness

Advertising recall:  the percentage of people who claim to have read, seen or heard an advertising campaign or execution and are able to describe some aspect specific to it.

back to top

B

Benchmark:  data from a study that are used as standards of comparison for other studies, waves or dips from a tracking study.  Also, using the performance of a product or advertisement in research as a standard against which others are judged.

Brand awareness:  see awareness

Brand preference:  the percentage of people who claim that a particular brand is their first choice.

Bias:  the tendency of people to answer survey questions a certain way thereby influencing the results of a study.  Bias can be introduced into a research study by an interviewer(interviewer bias), by a person’s history of brand use, by the order in which products are presented for evaluation in a taste test (order bias), by an imbalance of the demographic or brand use characteristics of a sample (sample bias).  Bias is controlled by the design of a study, by sampling criteria, and by statistical procedures used during computer processing of questionnaires.

Blind product test:  the evaluation of a product without identifying the brand name.

Buying attitudes:  the propensity of people to buy a particular brand as measured by a verbal or numerical scale.

back to top

C

Call back interview:  an interview with a respondent who has been contacted at an earlier date and questioned or recruited for a survey.

Central location test (CLT):  a research study that brings respondents to a single place, often a store front or mall site.  The sample is either pre-recruited by mail or telephone or intercept screened at the location and recruited to attend the interview.

Sometimes CLT studies are conducted  in a number of locations dispersed geographically.  Taste tests are often conducted in this manner.

Closed-end question:  a question that forces the respondent to choose one of a number of preset answers that are printed on the questionnaire.  

Cluster analysis:  a statistical formula employed via computer that analyzes the similarities and differences among a group of variables, such as brands, and groups together those that share certain characteristics.  Cluster analysis is useful to determine how a category or market is structured.  The results of a cluster analysis are often plotted on a graph that is called a competitive map. 

Coding:  involves translating the answers to open-end questions into a series of summary statements that are identified by number and or letter.  The number of the code is written next to the answer on the questionnaire and then tabulated along with the “closed end” answers.  These summary statements are called “codes”.  This process allows for the tabulation of the answers to open-end questions by computer.

Competitive map:  a process for analyzing the brands in a market to determine which brands compete most closely with each other.  The process often employs complex  multi-variate statistical techniques, so called because the techniques compare brands across a large number of variables simultaneously.

Competitive set:  those brands or products that are most likely to compete with each other for preference, as determined by responses to specific survey questions.

Concept test:  a technique for pretesting ideas for products or advertising themes.  It usually involves showing a less than finished version of a product or advertisement that is designed to convey the essence of the idea to a sample of people and asking them to evaluate it.

Control (control group):  independent cell or sample in a research study whose responses are used as a basis for comparison with a test group or cell;  e.g.  in an advertising test, an unexposed group serves as the control against which the responses of an exposed group are compared.

Conjoint analysis:  a multidimensional statistical technique used to determine the relationships among a large number of disparate variables.  The technique clusters together those variables that correlate to each other and  is used to predict the attributes that will have the most influence on consumer preferences.

Cross tabulation (cross tab):  an computer tabulation  technique that allows the comparison of data on the same dimension or question among different groups or segments.  For example, brand usage cross-tabbed by sex will reveal the differences between men and women on brand usage.  Sometimes referred to as in-tab.

back to top

D

Day after recall (DAR):  an advertising testing technique that measures the percentage of people who claim to remember seeing or reading an advertisement twenty four hours after having been exposed to it.  The total percentage of people who claim to remember the test advertisement is called claimed recall; the percentage who are able to describe something specific to the advertisement is called proven recall.  There are a number of testing services that provide day after recall measures as part of their technique.  Measurement of day after recall alone is not considered to be a sufficient evaluation the effectiveness of an advertising execution.  It should be accompanied by measures of persuasion and effects on brand imagery.

Dependent variable:  a characteristic or item from a research study that is being analyzed to determine the influence or effect that some other item has on it.  Usually used in the context of a cross tabulation, where the dependent variable (e.g. brand use) is examined by an independent variable (e.g. sex of respondent).  So called because brand use may vary by or be dependent upon the sex of the respondent.

Discriminator:  an attribute or benefit that separates brands in a product category.  Contrast with motivator.

back to top

E

Editing:  the process of reviewing completed questionnaires before tabulating them to determine if they are properly done.  The editing process eliminates or corrects interviewing or recording errors, such as answers to questions that a particular respondent was not qualified to be asked.  May be done both by computer as well as manually.

Evoked set:  a list of brands or products volunteered by respondents in response to specific questions.  Refers to those brands or products that share certain qualities, usually the preferred brands in a category.

back to top

F

Factor analysis:  a multivariate statistical program that examines a large number of variables and groups together those that correlate closely.  Often used to define the types of benefits people desire from a product category.

F-ratio:  a statistic used to measure the amount of variation in separate samples in order to determine if a difference in response between two samples is large enough not to be a chance variation.

Focussed group session (also focus group):  an interviewing method involving six to twelve subjects (respondents) who are gathered in a room and led through a series of probing questions by a skilled moderator about a particular topic or product category.  The technique is useful for exploring ideas and concepts in the early stages of new product or advertising development. 

Frequency distribution:  a chart or graph showing the number of times a variable is present or occurs across some dimension, such as time or space;  for example, the frequency of use of a product in a one month period among a sample of people is displayed on a frequency distribution.

back to top

G

Gap analysis:  a technique for analyzing opportunities and weaknesses whereby brands are arrayed on a two dimensional “map” defined by “x” and “y” axes.  The axes represent product qualities (e.g., quality, speed, value, status).  Brands are plotted on these axes.  The position of brands on the map reveals spaces (gaps) which represent areas of potential opportunity or weakness.

Group session:  see focussed group session

back to top

H

Halo effect:  a bias whereby a one favorable characteristic influences the overall judgement of an individual.  Brand preferences are the most common source of this bias.  Ratings of a product or brand on a series of attributes may be “inflated” by a favorable attitude toward the brand being rated.   Methodological techniques exist for controlling this bias in the design of a study.   There are also analytical and statistical techniques for removing the “halo effect” from data.

Home use test (HUT):  a method for evaluating products, whereby product samples are given to people to use in their homes under natural conditions for a period of time ranging from a few days to several months.  Participants are recontacted after  a predetermined period of time and questioned about their opinions of the product.

Home use tests are particularly useful for assessing a new product’s appeal before entering the marketplace.   Products may be tested blind (unidentified) or branded.

Hedonic evaluation (rating):  evaluation of a product’s pleasure-giving.  Usually refers to sensory evaluations (taste, smell, touch).  Ratings often taken in the course of a central location or in-home product test.

back to top

I

Importance rating:  used in a survey,  a scale designed to measure the degree of importance each of a number of variables of a product category.  Usually verbal (e.g. extremely important, very important, somewhat important, not too important, not at all important)

Incentive:  usually a financial reward given to people for participating in a survey or focussed group session.

Incidence:  the proportion of qualified respondents in the general population to be sampled.  affects the cost of a research study.

Independent variable:  a factor that is not dependent for change on other factors.  Also referred to as a predictor variable.  Usually used in the context of statistical analyzes conducted to determine the influence of a single quality or attribute (the independent variable) on a number of others (dependent variables).

Interviewer bias:  the influence on survey responses resulting from a prejudice or characteristic of the person conducting the interview.  Usually determined by comparing subsets of survey responses among interviewers, which reveals unusual response patterns.  Typically it is the responsibility of the fieldwork organization to prevent and/or detect interviewer bias in survey results before providing finished tabulations.

back to top

L

Leverage analysis:  method for determining the relative influence of each of a number of attributes or benefits on brand preferences.  Often conducted with a regression equation, but also may be done with a variety of analytical techniques.

Life-style segmentation:  process for dividing people into subgroups based on shared

activities, opinions and interests.  Used for targeting media or brand position.

back to top

M

Mall intercept interview:  method of sampling whereby an interviewer selects people for a survey from the pedestrian traffic passing a particular location.  The technique introduces many kinds of bias into survey sampling.

Mail survey:  survey technique that employs self-administered questionnaires returned to the survey center via the mail.  May be a two-way mail survey (distribution and return) or a one-way mail survey (distribution only or return only).

Mapping:  generally, the technique of plotting brands or attributes on a two or three dimensional graph where the positions of the variables visually represent their interrelationships.  also see perceptual mapping

Marginals:  the initial computer tabulations from a survey, unannotated and unchecked for errors.  Also referred to as a data dump.

Market segmentation study:  method for dividing populations into subgroups who share attitudes, behaviors, lifestyles or demographic characteristics.  Used for strategic purposes of selecting targets, positioning brands or identifying new product opportunities.

Minimarket test:  controlled sale market testing in small, non-representative panels of stores involving placement of product in distribution.  Essentially used for the measurement of product demand based on in-store presence.

Motivator:  an attribute or benefit that influences the use of product category.  Contrast to discriminator, which is an attribute or benefit that separates brands in a category.

Multivariate analysis:  statistical process used to  analyze simultaneously the relationships of a large number of items (variables)  to each other and to some independent variable.

Multidimensional scaling:  the process of arraying a large number of items (variables) on a map or matrix, where each item is plotted based on its relationship to each other.  Used to represent graphically how people perceive the brands or products in a category and how they distinguish among them.  The results are often displayed on a perceptual map.

Multiple regression analysis:  statistical technique for determining the relative influence of each of a number of dependent variables (e.g. such brand image dimensions as value, status, taste) on an independent variable (e.g. brand most likely to buy).

back to top

O

One-on-one interview:  qualitative in-depth questioning of a single respondent by a skilled interviewer, who usually follows a guide but does not use a structured questionnaire.

Open-end question:  one for which a respondent can answer in any manner he/she chooses, with no pre-structured alternatives.  e.g.  Why do you feel that way?  The use of such questions requires coding of the answers for tabulation.  Opposite of a closed-end question.

back to top

P

Paired comparison:  design for a product, ad or package test whereby the respondent is asked to evaluate two products (ads, etc.) relative to each other.

Perceptual map:  method of plotting brands on a two or three dimensional graph where the distances among them are calculated by a multivariate statistical formula and represent similarities and differences on specific attributes or benefits.  Provides a visual way to determine competitiveness among brands and opportunities for new brand positions. 

Persuasion test:  usually a method of evaluating advertising whereby the ability of an advertisement to increase the likelihood of purchase of those exposed to it is measured. 

Probability sample:  technique of survey respondent selection whereby each member of the population in the sample frame has an equal or known chance (probability) of being selected.  syn. random sample.  Can be used only when potential respondents are drawn from a known universe in a systematic way  (e.g. every nth person in a  census tracts or telephone household).

Psychographics:  term coined to represent the values, self image and general belief systems of people.  The counterpoint to demographics.

back to top

Q

Qualitative research:  market research that is not statistical in nature.  Provides anecdotal views of behavior, attitudes and perceptions that cannot be extrapolated to a larger population.  Usually takes the form of focussed group sessions or one-on-one in-depth interviews.

Quantitative research:  market research that is statistical in nature.  Provides information that is based on the analysis of frequency of responses to survey questions.  Results presented in the form of percentages.  Requires large and systematically drawn samples (in excess of 100 interviews) with interviews conducted individually using structured questionnaires.   May be in person, by telephone or via mail surveys.

Q-sort:  technique for survey respondents reviewing a large number of statements printed individually on cards and placing them in piles depending upon their degree of relevance to them.  Often used for a personality inventory.

Quota sampling:  method of structuring the sample of a research study based on obtaining a predetermined number of people with specific characteristics (e.g age, sex, brand use).

back to top

R

Random-digit dialing:  method for selecting telephone numbers to call for a survey.  Numbers are randomly generated, thereby permitting unlisted numbers to be called.

Many telephone interview suppliers have computer programs that generate the numbers to be called, dial them, and connect an interviewer only when a person is on the line, thereby increasing the efficiency of the interviewing process.  Usually referred to as CATI, for Computer Assisted Telephone Interviewing.

Random sample:  see probability sample

Range of error:  the variation that is calculated around a survey statistic.  Usually expressed as plus or minus x %points.  Permits the establishment of confidence levels in the results of a survey.  Can be done only with a probability sample.

Recall testing:  method for evaluating advertising that measures the ability of people to recall seeing or reading the test advertising some time after exposure.

Regression analysis:  see multiple regression analysis

Research design:  the overall plan for a study, specifying the nature of the sample, the structure of the study and the analysis and action standards to be followed.

Respondent:  the person who is the subject of an interview or takes part in a research study.

Recruiting:  soliciting participation in a research study from qualified respondents, usually in advance of the interview date.

Reliability:  quality of the data from a research study;  refers to the replicability of the results should the survey be repeated in an identical manner.  Affected by field controls,  sampling techniques and quality of interviewing.

back to top

S

Sample:  the people selected to participate in a research study.  see probability sample, quota sample.

Sampling plan:  the method for selecting respondents to be interviewed. 

Sampling error:  occurs when the people interviewed in a survey are not representative of the population intended to be surveyed.  Size of sampling error decreases as sample size increases.  also see standard error.

Sample weighting:  statistical method for balancing survey data to correct for over or under-representation of specific demographic groups.  Usually based on age, gender, household income.  Requires census data to make accurate application of weights.

Screener:  questions asked to determine if a person is qualified to be interviewed for a particular study.  Sample criteria are used to create the screening questionnaire.

Segmentation:  process of dividing a sample into subgroups that share certain characteristics or qualities;  e.g. age, gender, ethnicity, lifestyle, product or brand usage.

Used for strategic targeting and positioning.

Self-administered questionnaire:  survey that is filled out by the respondent without the aid or presence of an interviewer.

Semantic differential scale:  technique for measuring perceptions of brands or products whereby two words that represent opposite extremes of a particular attribute or benefit are used.  Respondent is asked to place his/her opinion in one of seven places between the two words based on their perception or belief.  e.g. bitter vs. sweet.

Sequential monadic:  design for a study whereby two products are evaluated individually by the same person.  The order of evaluation is rotated from respondent to respondent to control for order bias.  In some instances, the second evaluation is a paired comparison with the first product.

Significance (test):  statistical measure of the probability that a difference between two statistics has not occurred by chance.  Indicates the likelihood that the difference would reoccur a number of times greater than chance given an infinite number of repetitions of the test.  Several types of significance test are used depending upon the nature of the statistics being compared.  t test and chi-square are the most commonly used.

Single source data collection:  method for obtaining a wide variety of information from a single respondent, usually over a long period of observation.  Provides an opportunity to observe the associations among various behaviors, attitudes and beliefs. 

Skip pattern:  survey questionnaire design element that describes which questions a respondent will be asked based on his/her answers to previous questions.

Standard error:  measure of the amount a statistic may be expected to differ by chance from the true value of the statistic.  see sampling error.

Statistical significance:  see significance testing.

Strategic research:  studies whose purposes are to determine such brand or product issues as market target, brand position, buying incentive, pricing strategy, brand extension strategy.

back to top

T

Top of mind awareness (recall):  the proportion of people who first mention as brand or advertising campaign (unaided) in response to a survey question.  Used as a measure of brand or advertising salience.

Tracking study:  continuous survey reported out at intervals or taken in waves or dips

for the purpose of measuring progress of advertising or marketing campaigns.

Trade-off analysis:  questioning technique designed to determine how much of a particular quality or attribute a respondent is willing to sacrifice in order to get more of another.  Used to help determine priorities for new product development or brand positioning strategy.

back to top

U

Unaided brand awareness:  the proportion of people who mention a brand unprompted when asked to name the brands in a category.

Unaided recall:  the proportion of people who mention a brand or advertising campaign without prompting.  Determined by an open-end question, e.g.  what brand or brands do you remember seeing or reading advertising for in the last week?

Universe:  the total population that is being measured or studied in a survey or test usually described in terms of demographics and brand/product use.

back to top

V

Validation:  process of verifying the interviews taken in a survey by recontacting a small proportion of original respondents.

Validity:  truthfulness of data from research.  The aspect that the data predicts or represents what it purports to represent.

Variable:  an individual item in a research analysis, usually referred to as dependent or independent (which see).

Variance:  amount of divergence from a measure of central tendency in a distribution of data.  Used to evaluate confidence levels for differences between two distributions

(analysis of variance).

Volumetrics:  analysis of consumption quantities by respondents.  Used to determine heavy user segments and the value of particular groups or subsegments.

back to top

 
   
 

Chudnoff Associates, Inc.
7 Plaza Nine
Manalapan, NJ 07726
Phone: 732.863.1100
Fax: 732.294.7799

Email Us Now!
© 2012, CHUDNOFF ASSOCIATES - All Rights Reserved
   
Home   |   What we offer   |   Company   |   Team   |   Clients   |   Research Glossary   |   Blog   |   Sitemap   |   Contact