Marketing Research Exam Study Guide: Chapters 11, 12, 13, and 14

Flashcards covering questionnaire design principles, sampling methodologies, descriptive and inferential statistics, and statistical hypothesis testing errors.

Questionnaire

The primary tool used in survey research to collect responses to research questions.

Relevance

A criterion of questionnaire design requiring that every question collects information directly helping to answer the research question or solve the current marketing problem.

Accuracy

A criterion of questionnaire design requiring that information is reliable and valid, obtained through simple, understandable, unbiased, unambiguous, and nonirritating wording.

Open-ended questions

Questions in which respondents answer in their own words, especially useful in exploratory research when possible responses are not yet known.

Fixed-alternative questions

Questions in which respondents choose from a limited set of provided responses, making answers faster to complete and easier to compare and analyze.

Simple-dichotomy question

A fixed-alternative question that requires the respondent to choose one of two alternatives, such as yes or no.

Determinant-choice question

A multiple-choice question that requires respondents to choose exactly one answer from several provided alternatives.

Frequency-determination question

A fixed-alternative question that asks respondents about how often an activity is performed.

Checklist question

A fixed-alternative question that allows respondents to select multiple answers from a list of options.

Likert scale

A scale question format where respondents indicate their level of agreement or disagreement with a given statement.

Semantic differential scale

A rating scale question that asks respondents to evaluate an object, brand, or concept between opposing bipolar descriptions.

Stapel scale

A scale question that uses a numerical rating scale to evaluate a characteristic, typically placed around a descriptive word.

Mutually exclusive categories

Response category guidelines requiring that answer choices do not overlap, ensuring a respondent can fit into only one category.

All-inclusive categories

Response category guidelines requiring that answer options cover almost all reasonable answers, often using options like "Other" or "Not applicable."

Leading question

A question that suggests or implies a particular answer, potentially creating a bandwagon effect and threatening research validity.

Loaded question

A question phrased with emotionally charged wording or suggesting a socially desirable answer, prompting respondents to present themselves favorably.

Counterbiasing statement

A preamble or statement reassuring respondents that a potentially sensitive or embarrassing behavior is not abnormal.

Ambiguous question

A question containing vague terms such as "often," "frequently," or "soon," which respondents may interpret in different ways.

Double-barreled question

A question that improperly asks about two or more separate issues at the same time, making respondent answers ambiguous and difficult to interpret.

Questions that make assumptions

Questions that presuppose a respondent has had a particular experience or already holds an opinion on a subject.

Unaided recall

A memory measurement approach asking respondents to remember an event, advertisement, or detail without being provided any clues.

Aided recall

A memory measurement approach that provides a clue or prompt to help respondents recall an event or advertisement.

Telescoping

A timing-related memory error where respondents remember an event as having occurred more recently than it actually occurred.

Squishing

A timing-related memory error where respondents believe a recent event happened longer ago than it actually did.

Order bias

Distortion that occurs when the sequence in which questions or answer choices appear influences how respondents answer.

Funnel technique

A question-ordering approach where general questions are asked before specific questions to establish a respondent's broad frame of reference.

Anchoring effect

The phenomenon in which the first concept evaluated sets a baseline or comparison point that influences subsequent evaluations.

Split-ballot technique

A method where different respondents receive different versions of a question (such as positive versus negative wording) to assess wording effects.

Randomized response techniques

A survey procedure where respondents are randomly assigned to answer either a sensitive question or an ordinary, non-embarrassing question to protect privacy.

Breakoff

The premature termination of a survey when a respondent stops answering before completing the entire questionnaire.

Filter question

A question designed to determine whether a particular upcoming section or topic applies to a respondent.

Branching

A survey flow mechanism that directs respondents to different paths or sections based on their specific answers.

Skip logic

An automated survey feature that directs respondents past irrelevant questions according to their answers to filter questions.

Multiple-grid question

A matrix table layout that arranges several related questions alongside a uniform set of response categories in a grid.

Speeders

Online survey respondents who complete an assessment so rapidly that the truthfulness or quality of their data is questionable.

Heat map question

A survey question technology that tracks and visually highlights which portions of an image or advertisement attract the most respondent attention.

Status bar

A visual indicator displayed during an online survey that shows how much of the questionnaire is finished or remains.

Prompting

An automated alert notifying a respondent that they have skipped a required question or entered implausible information.

Piping

A survey technology feature that dynamically inserts an answer provided in an earlier question into the wording of a later question.

Pretesting

The trial administration of a questionnaire to a small sample resembling the target population to identify and fix flaws before full deployment.

Back translation

A translation validation process where a survey is translated into a target language and then independently translated back into the original language to detect meaning differences.

Population (universe)

The entire, complete group of entities or individuals being studied in marketing research.

Target population

The specific group or segment within the broader population that the researcher specifically seeks to investigate and understand.

Sample

A selected subset of a population observed or measured to infer characteristics of the larger population.

Census

An investigation that attempts to collect data from every individual element in the entire population.

Sampling frame

The actual list or operational database of elements from which a sample is selected, also referred to as the working population.

Sampling frame error

Error occurring when certain population elements are omitted from the sampling frame or when the frame fails to represent the target population accurately.

Sampling unit

The individual element or group of elements that is subject to selection within the sampling process.

Primary sampling unit (PSU)

The initial sampling unit chosen during the first stage of a multistage sampling design.

Reverse directory

A directory organized by city and street address or by telephone number, helpful for selecting samples in specific geographic locations.

Online panel

A pre-recruited group of individuals who have agreed to regularly participate in online marketing research studies.

Sample statistic

A summary metric calculated from sample data, used to estimate or make inferences about a population parameter.

Population parameter

A true measured characteristic or numerical summary of an entire specific population.

Statistical error

The numerical difference between a calculated sample statistic and its true population parameter counterpart.

Random sampling error

Error arising purely from chance variations in which specific elements are randomly drawn into a sample, decreasing as sample size increases.

Systematic (nonsampling) error

Error resulting from flaws in research design, list execution, questioning, or recording that cannot be resolved merely by increasing sample size.

Self-selection bias

A bias that emerges when individuals choose whether to participate in a study, often attracting volunteers with strong opinions who do not represent the broader population.

Opt-in

The formal act of giving explicit permission to receive specified emails or survey invitations from an organization.

Nonresponse error

A form of error occurring when selected sample units are unreachable or refuse to cooperate, potentially diminishing sample representativeness.

Nonresponse bias

A bias that occurs when the propensity to respond to a survey is systematically related to specific respondent traits or opinions.

Probability sampling

A sampling approach based on chance selection procedures where every population element has a known, nonzero, equal probability of selection.

Nonprobability sampling

A sampling technique where the probability of selecting any specific element is unknown, relying on arbitrary or pragmatic selection criteria.

Convenience sampling

A nonprobability sampling method where units or individuals are selected because they are conveniently and readily accessible to the researcher.

Judgment sampling (purposive sampling)

A nonprobability sampling technique in which an expert chooses sample elements based on personal judgment about what constitutes a representative sample.

Quota sampling

A nonprobability method where targets are established for specific demographic subgroups and interviewers fill these quotas using nonrandom selection.

Snowball sampling

A nonprobability sampling technique where initial respondents refer additional individuals possessing the desired characteristic, useful for rare populations.

Simple random sampling

A fundamental probability sampling method in which every single element in the population has an equal opportunity of being selected.

Systematic sampling

A probability sampling design where every $n\text{th}$ element on a list is chosen after initiating selection from a random starting point.

Stratified sampling

A probability sampling technique where the population is partitioned into homogeneous subgroups (strata), and samples are randomly drawn from each stratum.

Proportional stratified sampling

Stratified sampling where the number of units selected from each stratum corresponds directly to that stratum's proportional representation in the target population.

Disproportional stratified sampling

Stratified sampling where stratum sample allocations deviate from population proportions to satisfy specific analytical goals, such as adequately analyzing a small subgroup.

Cluster sampling

A probability sampling method where the primary sampling unit is an intact group or cluster of elements rather than an individual person or object.

Multistage area sampling

A cluster sampling procedure conducted over multiple geographic levels, combining probability methods across states, counties, neighborhoods, and households.

Descriptive statistics

Statistical measures that organize, summarize, and describe the characteristics of an observed sample or population.

Inferential statistics

Statistical procedures that utilize sample findings to deduce properties or draw generalized conclusions regarding a population.

Metrics

Summary numerical figures used to assess and compare characteristics of an observed sample against a population benchmark.

Frequency

The exact number of times a particular observation, score, or response option occurs within a data set.

Frequency distribution

A structured table or summary that displays how frequently each distinct value or score occurs in a collection of observations.

Relative frequency distribution

A percentage distribution formed by dividing the frequency of each value by total observations and multiplying by $100$.

Top-box score

The percentage of respondents selecting the most favorable answer category on a rating scale.

Bottom-box score

The percentage of respondents selecting the least favorable rating option, often used to pinpoint dissatisfaction and guide managerial intervention.

Central tendency

A statistical concept describing the middle or central point around which observations in a distribution cluster.

Mean

The arithmetic average computed by summing all observation values and dividing by the total count of observations; susceptible to outliers.

Median

The middle score or 50th percentile of an ordered set of observations, resistant to extreme values and outliers.

Mode

The specific value or category that appears most frequently in a distribution, usable for data below the interval measurement level.

Dispersion

The degree of spread, scattering, or variability found among values in a data distribution.

Range

The difference calculated by subtracting the lowest observed value from the highest observed value in a data set.

Interquartile range (IQR)

The difference between the first quartile and third quartile, encompassing the middle $50\%$ of observations in a distribution.

Deviation score

The difference obtained by subtracting the distribution mean from an individual observation ($\text{observation} - \text{mean}$).

Variance

A statistical measure of sample dispersion that equals zero when all observed data points are identical to the mean.

Standard deviation

A fundamental metric of spread equal to the square root of the sample variance, denoted by $S$.

Sequential sampling

A sampling strategy where researchers review preliminary pilot study findings before deciding on the size of a subsequent, larger sample.

Generalizability

The extent to which statistical results derived from a sample can be reliably applied to the broader target population.

Type I error ($\alpha$)

A false positive error occurring when a researcher incorrectly rejects a true null hypothesis, mistakenly concluding an effect exists when it does not.

Type II error ($\beta$)

A false negative error occurring when a researcher fails to reject a false null hypothesis, thereby missing a true population difference or effect.