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 |
