Aided recall. Respondents are asked if they remember a commercial for the brand being tested.
Alternative hypothesis. A competing hypothesis to the null.
Attitude. A learned predisposition to respond in a consistently favourable or unfavourable manner with respect to a given object.
Audit. A formal examination and verification of either how much of a product has sold at the store level (retail audit) or how much of a product has been withdrawn from warehouses and delivered to retailers (warehouse withdrawal audits).
Balanced scale. Scale using an equal number of favourable and unfavourable categories.
Banner. The variables that span the columns of the cross-tab; generally represents the subgroups being used in the analysis.
Before-after design. Experiment where a measurement is taken from respondents before they receive the experimental treatment condition; the experimental treatment is then introduced, and the post-treatment measurement is taken.
Before-after with control design. Experiment that adds a control group to the basic before-after design; the control group is never exposed to the experimental treatment.
Between-group variations. Between-group differences in scores for groups that were exposed to different treatments - represents "explained" variation.
Blind testing. Tests where the brand name of the product is not disclosed during test.
Cartoon completion test. Projective technique that presents respondents with a cartoon of a particular situation and asks them to suggest the dialogue that one cartoon character might make in response to the comment(s) of another cartoon character.
Causality. Relationship where a change in one variable produces a change in another variable. One variable affects, influences, or determines some other variable.
Chi-square test statistic. Measure of the goodness of fit between the numbers observed in the sample and the numbers we should have seen in the sample, given the null hypothesis is true.
Cognition. A person's knowledge, opinions, beliefs and thoughts about the object.
Comparative scaling, (non-metric scaling) Scaling process in which the subject is asked to compare a set of stimulus objects directly with one another.
Comparison product test. Designs where a consumer rates products by directly comparing two or more products.
Concept board. Illustration and copy describing how the product works and its end-benefits.
Concept evaluation tests. Concept tests designed to gauge consumer interest and determine strengths and weaknesses of the concept.
Concept screening test. Concept tests for screening new product ideas or alternative end-benefits for a single product idea.
Concept test. Collection of information on purchase intentions, likes/dislikes and attribute rating in order to measure the relative appeal of ideas or alternative positioning and to provide direction for the development of the product and the product advertising.
Concept. An idea aimed at satisfying consumer wants and needs.
Concept/construct. Names given to characteristics that we wish to measure.
Confidence interval. Range into which the true population value of the characteristic being measured will fall, assuming a given level of certainty.
Confounds or confounding variables. Extraneous causal factors (variables) that can possibly affect the dependent variable and, therefore, must be controlled.
Connotative meaning. The associations that the name implies, beyond its literal, explicit meaning; the imagery associated with a brand name.
Constant sum scale. Procedure whereby respondents are instructed to allocate a number of points or chips among alternatives according to some criterion - for example, preference, importance, and so on.
Constitutive definition. Specifications for the domain of the constructs of interest so as to distinguish it from other similar but different constructs.
Continuous rating scale. (graphic rating scale) Procedure that instruct the respondent to assign a rating by placing a marker at the appropriate position on a line that best describes the object under study.
Control test market. Method in which the entire test market project is handled by an outside research company.
Copy recall. Percentage of respondents in the programme audience that correctly recalled copy elements in the test commercial.
Cross-price elasticity of demand. The percentage of change in demand for one product divided by the percentage change in price of the second product, assuming that all other factors affecting demand are constant.
Diary panels. Samples of households that have agreed to provide specific information regularly over an extended period of time. Respondents in a diary panel are asked to record specific behaviours as they occur, as opposed to merely responding to a series of questions.
Delphi method. A method of forecasting based on asking a group of experts for their best estimate of a future event, then processing and feeding back some of the information obtained, and then repeating the process; on the last set of responses, the median is usually chosen as the best estimate for the group.
Dependent variable. A variable whose value is thought to be affected by one or more independent variables. For instance, sales (dependent variable) are likely to be a function of advertising, availability, price, degree of competitive advantage, customer tastes, etc.
Depth interview ("one-on-one"). Sessions in which free association and hidden sources of feelings are discussed, generally through a very loose, unstructured question guide, administered by a highly skilled interviewer. It attempts to uncover underlying motivations, prejudice, attitudes toward sensitive issues, etc.
Dollar metric scale. (graded paired comparison) Scale that extends the paired comparison method by asking respondents to indicate which brand is preferred and how much they are willing to pay to acquire their preferred brand.
Double-barrelled questions. Questions in which two opinions are joined together.
Dummy magazine test. A realistic-looking test format using a dummy magazine that systematically varies the advertisements in such a way that some families receive magazine containing the test ad and other (matched) families receive a dummy magazine containing no ads at all.
Duo-trio designs. Test where a respondent is given a standard product and asked to determine which of two other products is more similar.
Electronic process. Review of the questionnaires for maximum accuracy and precision.
Ethnography. The systematic recording of human cultures.
Experimental design. A contrived situation designed so as to permit the researcher to manipulate one or more independent variables whilst controlling all extraneous variables and measuring the resultant effects on a dependent variable.
Filter question. A question that is asked to determine which branching question, if any, will be asked.
Focus group interview. Interview in which the interviewer listens to a group of individuals, who belong to the appropriate target market, talk about an important marketing issue.
Forced itemised test. Procedure in which a respondent indicates a response on a scale, even though he or she may have "no opinion" or "no knowledge" about the question.
Frequency distribution. The number of respondents who choose each alternative answer as well as the percentage and cumulative percentage of respondents who answer.
Funnel sequence. The procedure of asking the most general (or unrestricted) question about the topic under study first, followed by successively more restricted questions.
Gross incidence. Product/category use incidence for the entire population.
Hypothesis. An assumption or guess the researcher or manager has about some characteristic of the population being sampled.
Independent variable. A variable over which the researcher is able to exert some control with a view to studying its effect upon a dependent variable. For instance, an experiment may be conducted where the price (independent variable) of a dozen boxed carnations is varied and the sales (dependent variable) is observed at each price set.
Internal secondary data. Data available within the organisation - for example, accounting records, management decision support systems, and sales records.
Interval data. Measurements that allow us to tell how far apart two or more objects are with respect to attributes and consequently to compare the difference between the numbers assigned. Because the interval data lack a natural or absolute origin, the absolute magnitude of the numbers cannot be compared.
Itemised (closed-ended) questions. Format in which the respondent is provided with numbers and/or predetermined descriptions and is asked to select the one that best describes his or her feelings.
Itemised rating scaling. The respondent is provided with a scale having numbers and/or brief descriptions associated with each category and asked to select one of the limited number of categories, ordered in terms of scale position, that best describes the object under study.
Judgemental sampling. Studies in which respondents are selected because it is expected that they are representative of the population of interest and/or meet the specific needs of the research study.
Judgemental data. Information generally based on perceptions or preference may give better indications of future patterns of consumption.
Jury of expert opinion. A method of forecasting based on combining the views of key executives.
Laboratory experimental environment. Research environment constructed solely for the experiment. The experiment has direct control over most, if not all, of the crucial factors that might possibly affect the experimental outcome.
Likert scale. Scaling technique where a large number of items that are statements of belief or intention are generated. Each item is judged according to whether it reflects a favourable or unfavourable attitude toward the object in question. Respondents are then asked to rate the attitude towards the object on each scale item in terms of a five-point category labelled scale.
Line marking. Similarity judgements recorded by making a mark on a 5-inch line anchored by the phrases "exactly the same" and "completely different".
Line marking/continuous rating non-comparative scale. Procedure that instructs the respondent to assign a rating by placing a marker at the appropriate position on a line that best describes the object under study. There is no explicit standard for comparison.
Loaded questions. Questions that suggest what the answer should be or indicate the researcher's position on the issue under study.
Loadings. Weightings that give the correlation of the attribute with respect to the dimension.
Magnitude estimation. Scale in which respondents assign numbers to objects, brands, attitude statements, and the like so that ratios between the assigned numbers reflect ratios among the objects on the criterion being scaled.
Mail diary services. General term for services involving a sample of respondents who have agreed to provide information such as media exposure and purchase behaviour on a regular basis over an extended period of time.
Mail surveys. Data-collection method that involves sending out a fairly structured questionnaire to a sample of respondents.
Mall-intercept personal survey. Survey method using a central-location test facility at a shopping mall; respondents are intercepted while they are shopping.
Market segment. Subgroups of consumers who respond to a given marketing-mix strategy in a similar manner.
Maturation. Threat to internal validity; refers to changes in biology or psychology of the respondent that occur over time and can affect the dependent variable irrespective of the treatment conditions.
Measurement. Process of assigning numbers to objects to represent quantities of attributes.
Monadic products test. Designs where a consumer evaluates only one product, having no other product for comparison.
Mortality. Threat to internal validity; refers to the differential loss (refusal to continue in the experiment) of respondents from the treatment condition groups.
Nominal data. Measurement in which the numbers assigned allow us to place an object in one and only one of a set of mutually exclusive and collectively exhaustive classes with no implied ordering.
Non-comparative scaling (monadic scaling). Scaling method whereby the respondent is asked to evaluate each object on a scale independently of the other objects being investigated.
Non-probability samples. Form of sampling where there is no way of determining exactly what the chance is of selecting any particular element or sampling unit into the sample.
Non-response error. Error that occurs because not all of the respondents included in the sample respond; in other words with non-response, the mean true value (on the variable of interest) of the sample respondents who do respond may be different from the entire sample's true mean value (on the variable of interest).
Non-sampling error. Degree to which the mean observed value (on the variable of interest) for the respondent of a particular sample agrees with the mean true value of the particular sample of respondents (on the variable of interest).
Observational methods. Observation of behaviour, directly or indirectly, by human or mechanical methods.
Optical scanning. Direct machine reading of numerical values or alphanumeric codes and transcription onto cards, magnetic tape, or disk.
Order bias. Condition whereby brands receive different ratings depending on whether they were shown first, second, third, etc.
Ordinal data. Measurement in which the response alternatives define an ordered sequence so that the choice listed first is less (greater) that the second, the second less (greater) than the third, and so forth. The numbers assigned do not reflect the magnitude of an attribute possessed by an object.
Over-registration. Condition that occurs when a sampling frame consists of sampling units in the target population plus additional units as well.
Paired comparison designs. Tests where a consumer directly compares two products.
Paired comparison scale. Scale that presents the respondent with two objects at a time and asks the respondent to select one of the two according to some criterion.
Primary data. Data collected for a specific research need; they are customised and require specialised collection procedures.
Print ad tests. Attempts to assess the power of an ad placed in a magazine or newspaper to be remembered, to communicate, to affect attitudes, and ultimately, to produce sales.
Probability sampling designs. Samples drawn in such a way that each member of the population has a known, non-zero chance of being selected.
Project proposal. A written description of the key research design that defines the proposed study.
Projective techniques. A class of techniques which presume that respondents cannot or will not communicate their feelings and beliefs directly; provides a structured question format in which respondents can respond indirectly by projecting their own feelings and beliefs into the situation while they interpret the behaviour of others.
Proportional allocation. Sampling design guaranteeing that stratified random sampling will be at least as efficient as SRS. The number of elements selected from a stratum is directly proportional to the size of the stratum.
Purchase intent scale. Procedure attempting to measure a respondent's interest in a brand or product.
Q-sort scale. Rank order procedure in which objects are sorted into piles based on similarity with respect to some criterion.
Qualitative research methods. Techniques involving relatively large numbers of respondents, which are designed to generate information that can be projected to the whole population.
Quota sampling. Design that involves selecting specific numbers of respondents who possess certain characteristics known, or presumed, to affect the subject of the research study.
Random sampling error. Error caused when the selected sample is an imperfect representation of the overall population; therefore, the true mean value for the particular sample of respondents (on the variable of interest) differs from the true mean value for the overall population (on the variable of interest).
Random sources of error. Denoted by X, component made up of transient personal factors that affect the observed scale score in different ways each time the test is administered.
Range. Differences between largest and smallest values of distribution.
Rank-order scale. Scale in which respondents are presented with several objects simultaneously and requested to "order" or "rank" them.
Ratio data. Measurements that have the same properties as interval scales, but which also have a natural or absolute origin.
Recall. Measures of how many people remember having seen the test ad both on an unaided and aided basis.
Related samples. The measurement of the variables of interest in one sample can affect the measurement of the variable in some other sample.
Residual. An error term representing the difference between the actual and predicted values of the dependent variable.
Response error. Error that occurs because respondents (who do respond) may give inaccurate answers, or a respondent's answers may be misrecorded.
Response rates. The total number of respondents sent questionnaires who complete and return them, expressed as a percentage.
Sample two-stage cluster sampling. Design in which the clusters at the first stage are selected by SRS; at the second stage the sampling units are selected probabilistically by SRS from each sample cluster so that with clusters of equal size the same fraction of sampling units is drawn from each sample cluster.
Sample. A subset of the target population from which information is gathered to estimate something about the population.
Sampling frame. An explicit list of individuals or households that are eligible for inclusion in the sample.
Sampling interval. Computed by taking n/N together with r, the first chosen element to be included in the sample, determines which elements will be included in the sample.
Sampling units. The elements that make up the population.
Sampling variable. Variable that represents the characteristic of the population that we wish to estimate.
Sampling. Identification of a group of individuals or households (or institutions or objects) that can be reached by mail, telephone, or in person, and that possess the information relevant to solving the marketing problem at hand.
Scale transformation. Procedures for transforming data by one of a number of simple arithmetic operations to make comparisons across respondents and/or scale items.
Secondary data. Data that have been collected for another project and have already been published. Sources can be in-house or external.
Selection bias. Threat to internal validity; refers to the improper assignment of respondents to treatment conditions.
Semantic differential scale. Semantic scale utilising bi-polar adjectives as end points.
Sentence completion. Projective technique whereby respondents are asked to complete a number of incomplete sentences with the first word or phrase that comes to mind.
Simple one-stage cluster sampling. One-step design in which the first stage clusters all sampling units are selected by SRS, and within each selected cluster all sampling units are chosen.
Simple random sampling. Design guaranteeing that every sample of a given size as well as every individual in the target population has an equal chance of being selected.
Simple weighting. Procedure that attempts to remove non-response bias by assigning weights to the data that in some sense account for non-response.
Simulated test market. Method whereby various groups of pre-selected respondents are interviewed, monitored and sampled about the new product; in addition, respondents may be exposed to various media messages in a controlled environment.
Single-stage cluster sample. One step design where, once the sample of clusters is selected, every sampling unit within each of the selected clusters is included in the sample.
Snowball design. Sample formed by having each respondent, after being interviewed, identify others who belong to the target population of interest.
Split-halves. Scale items split in terms of odd-and even-numbered Hems or randomly. Standard
deviation. Index of variability in the same measurement units used to calculate the mean.
Standard error (s). Indication of the reliability of an estimate of a population parameter; it is computed by dividing the standard deviation of the sample estimate by the square root of the sample size.
Stapel scale. Procedure using a single criterion or key word and instructing the respondent to rate the object on a scale.
Store audits. Studies that monitor performance in the marketplace among dollar and unit sales/share, distribution/out of stock, inventory, price, promotional
Stratified sampling. Design that involves partitioning the entire population of elements into sub-population, called strata, and then selecting elements separately from each sub-population.
Survey. A method of gathering information from a number of individuals (the respondents, who collectively form a sample) in order to learn something about a larger target population from which the sample was drawn.
Syndicated research services. Market research suppliers who collect data on a regular basis with standardised procedures. The data are sold to different clients.
Systematic sampling. Design whereby the target sample is generated by picking an arbitrary starting point (in a list) and then picking every nth element in succession from a list.
Systematic sources of error. Denoted by X, component made up of stable characteristics that affect the observed scale score in the same way each time the test is administered.
Target population. Set of people, products, firms, markets, etc., that contain the information that is of interest to the researcher.
Telephone surveys. Survey that involves phoning a sample of respondents drawn from an eligible population and asking them a series of questions.
Telescoping. Condition that occurs when a respondent either compresses time or remembers an event as occurring more recently than it actually occurred.
Test markets. A system that allows the marketing manager to evaluate the proposed national marketing program in a smaller, less expensive situation with a view to determining whether the potential profit opportunity from rolling out the new product or line extension outweighs the potential risks.
Thematic apperception test (TAT). Projective technique presenting respondents with a series of pictures or cartoons in which consumers and products are the primary topic of attention.
Third person/role playing. Projective technique that represent respondents with a verbal or visual situation and asks them to relate the feelings and beliefs of a third person to the situation, rather than to directly express their own feelings and beliefs about the situation.
Top-down approach. Process of breaking down clusters: or, at the beginning, all respondents belong to one segment, and then respondents are partitioned into two segments, then three segments, and so on until each respondent occupies his or her own segment.
Tracking. System for measuring the key sales components of customer awareness and trail and repeat purchases.
Trade-off procedure. Technique where the respondent is asked to consider two attributes at a time - to rank the various combinations of each pair of attribute descriptions from a most preferred to least preferred.
Treatment. A reference to an independent variable that has been manipulated by the researcher. For example, a researcher may be investigating the customer benefits of three prototype packaging designs in order to determine which design to use. The independent variable which is manipulated is product packaging.
Treatment. Term for that independent variable that has been manipulated.
Triangle designs. Tests where a respondent is given two samples of one product and one sample of another and asked to identify the one that differs.
Two-tail hypothesis test. Test used when the alternative hypothesis is non-directional - the region of rejection is in both tails of the distribution.
Type I error. Situation occurring when the null hypothesis is in fact true, but is nevertheless rejected on the basis of the sample data.
Type II or beta error. Situation occurring when we fail to reject the null hypothesis (HO), when in fact the alternative (HA) is true.
Unaided questions. Questions that do not provide any clues to the answer.
Unaided recall. Respondents are asked if they remember seeing a commercial for a product in the product category of interest.
Unbalanced scale. Scale using an unequal number of favourable and unfavourable scale categories.
Unfinished scenario story completion. Projective technique whereby respondents complete the end of a story or supply the motive for why one or more actors in a story behaved as they did.
Unstructured interview. Method of interviewing where questions are not completely predetermined and the interviewer is free to probe for all details and underlying feelings.
Utility scale values. Ratings that indicate how influential each attribute level is in the consumer's overall evaluations.
Validation. Procedure where between 10 and 20 percent of all respondents "reportedly" interviewed are recontacted by telephone and asked a few questions to verify that the interview did in fact take place.
Validity. Refers to the best approximation to truth or falsity of a proposition, including propositions concerning cause-and-effect relationships.
Word association. Projective technique whereby respondents are presented with a list of words, one at a time, and asked to indicate what word comes immediately to mind.
No comments:
Post a Comment