| Helpful Definitions
A - M, N - Z
A | B | C | D | E | F | G | H | I | J | K | L | M | N - Z
A
Administrative Control - Term applied to studies relying on questionnaires and referring to the speed, cost, and control of the replies afforded by the mode of administration.
Aided Question - Question where the respondent receives some prompting, such as a list of product names.
Alternative (scientific) Hypothesis - States there is a change, impact, difference, or relationship between two variables.
Analysis of Variance (ANOVA) - Procedure that tests for significant mean differences in situations where there are three or more categories or treatment conditions, under the assumption that the populations are normally distributed.
Area Sampling - Form of cluster sampling in which areas (neighborhoods, census tracts) serve as the primary sampling units. The population is divided into mutually exclusive and exhaustive areas using maps, and a random sample of areas is selected.
B
Banner - A series of cross-tabulations between a criterion, or dependent, variable and several explanatory variables in a single table.
Baseline - An initial research project against which future changes are benchmarked.
Branching Questions - A technique used to direct respondents to different places in a questionnaire based on their response to the question at hand.
C
Causal Research - Research design in which the major emphasis is on determining a cause-and-effect relationship.
Central-Limit Theorem - Theorem that holds that if simple random samples of size n are drawn from a parent population with mean m and variance s^2, then when n is large, the sample mean will be approximately normally distributed with the mean equal to m and variance equal to (s^2)/n. The approximation will become more and more accurate as n becomes larger.
Chi-square Goodness-of-Fit Test - Used to test if a sample of data came from a population with a specific distribution.
Chi-square Test - Nonparametric procedure used to test if the standard deviation of a population is equal to a specified value. There are two types of chi-square tests: test for goodness of fit and test for independence.
Chi-square Test for Independence - Used to test the relationship between two different categorical factors for contingency.
Class Interval - By grouping the values of a distribution into classes, the data is easier to read and to analyze. The class interval is the width of each grouping of data.
Closed-Ended Questions - Questions characterized by the condition that responses are limited to a set of given options or ranges, such as rating scales. Also called fixed-alternative questions.
Cluster Sample - A probability sample distinguished by a two-step procedure in which (1) the parent population is divided into mutually exclusive and exhaustive subsets, and (2) a random sample of subsets is selected.
Coding - Technical procedure by which data are categorized; it involves specifying the alternative categories or classes into which the responses are to be placed and assigning code numbers to the classes.
Coefficient of Determination - The square of the coefficient of correlation. Measures goodness-of-fit (how well or tightly the data fit the estimated model). Takes on values between 0 and 1, with values closer to 1 implying a better fit.
Cohort - The aggregate of individuals who experience the same event within the same time interval.
Comparative Rating Scale - Scale requiring subjects to make their ratings as a series of relative judgments or comparisons rather than as independent assessments.
Completely Randomized Design - Experiment design in which independent random samples are drawn from each of the populations of interest.
Computer-Assisted Interviewing - The conducting of surveys using computers to manage the sequence of questions in which the answers are recorded electronically through the use of a keyboard.
Confidence Interval - The range around a survey result for which there is a high statistical probability that it contains the true population parameter.
Confidence Level - The probability that a particular confidence interval will include the true population value.
Conjoint Analysis - A method for establishing respondents’ utilities or evaluations based on the preferences they express for combinations of product attributes and features. Price is typically one of the attributes included.
Constant Sum Method - A type of comparative rating scale in which an individual is instructed to divide some given sum among two or more attributes on the basis of their importance to him or her.
Content Validity - Approach to validating a measure by determining the adequacy with which the domain of the characteristic is captured by the measure: it is sometimes called face validity.
Contingency Table - Two-dimensional table used to display the data of a chi-square test for independence.
Continuous Variables - Variables that can theoretically have an infinite number of values between adjacent units on the scale. Successive refinements of the measuring instrument yield increasingly precise values of a continuous variable. Examples are time, weight, and height.
Convenience Sample - Nonprobability sample sometimes called an accidental sample because those included in the sample enter by accident, in that they just happen to be where the study is being conducted when it is being conducted.
Correlation - The statistical technique used to measure and describe a relationship between two variables. It measures three characteristics of the relationship between two variables: direction (positive or negative), form (straight line or curved line), and degree (how strong or weak the relationship is).
Cross-Sectional Study - Investigation involving a sample of elements selected from the population of interest at a single point in time.
Cross Tabulation - A method of analyzing data that lets the analyst look at the responses to one question in relation to the responses to one or more other questions.
D
Data - Raw, unsummarized, and unanalyzed facts and figures.
Data Mining Software - uses sophisticated decision-making processes to search raw data for patterns and relationships that may be significant.
Degrees of Freedom - A measure of how precise an estimate of variation is. It is defined as (n-1), or one less than the sample size.
Deontology - An ethical or moral reasoning framework that focuses on the welfare of the individual and that uses means, intentions, and features of the act itself in judging its ethicality; sometimes referred to as the rights or entitlements model.
Dependent (or Correlated) Sample t-test - Used to compare two means taken from the same sample, such as in a pretest/posttest situation.
Descriptive Statistics - Concerned with the mathematical techniques that are used to describe collected sample data.
Dichotomous Question - Fixed-alternative question in which respondents are asked to indicate which of two alternative responses most closely corresponds to their position on a subject.
Discrete Variables - Variables of which there is no possible value between adjacent units on the scale; there are separate, distinct measurements of a given characteristic. Examples are the frequency of any event, gender, and hair color. Nominal and ordinal variables are always discrete.
Discriminant Analysis - Statistical technique employed to model the relationship between a dichotomous or multichotomous criterion variable and a set of p predictor variables.
Disguise - Amount of knowledge about the purpose of a study communicated to the respondent by the data-collection method. An undisguised questionnaire, for example, is one in which the purpose of the research is obvious from the questions posed, whereas a disguised questionnaire attempts to hide the purpose of the study.
Disproportionate Stratified Sampling - Stratified sample in which the individual strata or subsets are sampled in relation to both their size and their variability; strata exhibiting more variability are sampled more than proportionately to their relative size, while those that are very homogeneous are sampled less than proportionately.
Distribution - A given set of values or measures of the same characteristic. An example is a list of the scores students in one class made on their last test.
Distribution of Sample Means - The collection of sample means for all the possible random samples of a particular size that can be obtained from a given population.
Double-Barreled Question - A question that calls for two responses and thereby creates confusion for the respondent.
E
Element - Term used in sampling to refer to the objects on which measurements are to be taken, such as individuals, households, business firms, or other institutions.
Equivalence - Measure of reliability that is applied to both single instruments and measurement situations. When applied to instruments, the equivalence measure of reliability is the internal consistency or internal homogeneity of the set of items forming the scale; when applied to measurement situations, the equivalence measure of reliability focuses on whether different observers or different instruments used to measure the same individuals or objects at the same point in time yield consistent results.
Experience Survey - Interviews with people knowledgeable about the general subject being investigated.
Exploratory Research - Research design in which the major emphasis is on gaining ideas and insights; it is particularly helpful in breaking broad, vague problem statements into smaller, more precise sub problem statements.
External Validity - One criterion by which an experiment is evaluated; the extent to which the observed experimental effect can be generalized to other populations and settings.
F
F-Test - Method used to test whether two samples come from populations with equal standard deviations, or to decide whether a new process, treatment, or test reduces the variability of the current process.
Factorial Experiment - Method for selecting the treatments to be included in an experiment. A complete factorial experiment is one in which observations are made for every combination of the factor levels.
Fill-in-the-blank Questions - A type of open-ended question that requests a single word or limited response set.
Focus Group - Personal interview conducted among a small number of individuals simultaneously; the interview relies more on group discussion than on directed questions to generate data.
Frame Error - Error that occurs when a sample drawn from a population differs from the target population.
Frequency - The number of times that a value appears in a distribution.
Frequency Distribution - Presents the variable values and their frequency of occurrence.
Friedman F-Test - Nonparametric test comparable to two-way ANOVA that is used to compare k populations when a randomized block design is conducted.
Graphic Rating Scale - Scale in which individuals indicate their ratings of an attribute by placing a check at the appropriate point on a line that runs from one extreme of the attribute to the other.
H
Halo Effect - Problem that arises in data collection due to the tendency to be influenced by a single positive or outstanding trait in forming an overall favorable evaluation.
Halo Error - A type of rating error that occurs when an employee receives the same rating on all dimensions regardless of his or her performance on individual ones.
Histogram - Form of bar chart on which the values of the variable are placed along the X-axis and the absolute frequency or relative frequency of occurrence of the values is indicated along the Y-axis.
Hit Rate - Measure used to assess the results of a discriminant analysis by measuring the proportion of the objects that were correctly classified by the discriminant function(s) in the group to which they actually belong.
Homogeneity - A type of rating error that occurs when a rater gives all employees a similar rating regardless of their individual performances.
Hypothesis Testing - Statistical procedures that allow researchers to use sample data to draw inferences about the population of interest.
I
Incidence - The percentage of the population or group that qualifies for inclusion in the sample using some criteria.
Independent Samples t-test - Used to compare the means and check for significant differences between the means of two independent samples.
Inferential Statistics - Concerned with the mathematical techniques that allow the use of collected sample data to be generalized to the population.
Information - Data that has been converted into a meaningful and useful context for the receiver.
Information Control - Term applied to studies using questionnaires and concerning the amount and accuracy of the information that can be obtained from respondents.
Instrument Variation - Any and all changes in the measuring device used in an experiment that might account for differences in two or more measurements.
Internal Validity - One criterion by which an experiment is evaluated; the criterion focuses on obtaining evidence demonstrating that the variation in the criterion variable was the result of exposure to the treatment or experimental variable.
Internet Samples - Classified as unrestricted, screened, and recruited. In unrestricted Internet samples, anyone who desires can complete the online questionnaire. Screened Internet samples impose quotas based on some desired sample characteristics and screen respondents before presenting the questionnaire. Recruited Internet samples recruit respondents in targeted populations by phone, mail, e-mail, or in person before directing qualified participants to the online questionnaire.
Interquartile Range - The distance between the upper and lower quartiles.
Interval Scales - Have the attributes of magnitude and equal intervals. Each unit represents the same magnitude of the variable being measured across the whole range of the scale. Also, zero is an arbitrary point on the scale, which means that comparisons about how many times greater one score is than another cannot be made. Scoring on IQ tests and most standardized tests are measured on interval scales. An example of an interval scale is an attitudinal scale of agree/disagree ranked responses: strongly disagree, disagree, neither agrees nor disagrees, agree, and strongly agree.
Item Nonresponse - Source of nonsampling error that arises when a respondent agrees to an interview but refuses or is unable to answer specific questions.
Itemized Rating Scale - Scale in which individuals must indicate their ratings of an attribute or object by selecting one from among a limited number of categories that best describes their attitude toward the attribute or object.
J
Judgment Sample - Nonprobability sample that is often called a purposive sample; the sample elements are handpicked because they are expected to serve the research purpose.
K
Kolmogorov-Smirnov Test - Nonparametric statistical test used to determine whether some observed pattern of frequencies corresponds to some expected pattern; also tests whether two independent samples have been drawn from the same population or from populations with the same distribution.
Kruskal-Wallis H-Test - Nonparametric test used to determine if three or more independent samples come from populations having the same distribution. A nonparametric version of ANOVA.
L
Leading Question - A question framed to give the respondent a clue about how he or she should answer.
Level of Significance - Probability of making a Type I Error.
Literature Search - Search of statistics, trade journal articles, other articles, magazines, newspapers, and books for data or insight into the problem at hand.
Longitudinal Study - Investigation involving a fixed sample of elements that is measured repeatedly through time.
Lower Quartile - The 25th percentile of a data set.
M
Mail Questionnaire - Questionnaire administered by mail to designated respondents with an accompanying cover letter to be returned by mail by the subject to the research organization.
Mall Intercept - A method of data collection in which interviewers in a shopping mall stop a sample of those passing by to ask them if they would be willing to participate in a research study.
Mann-Whitney U-statistic - Test equivalent to Wilcoxon rank sum test used for comparing two populations based on independent random samples.
Marketing Research - Function linking the consumer to the marketer through information used to identify and define marketing opportunities and problems; generate, refine, and evaluate marketing actions; monitor marketing performance; and improve understanding of marketing as a process.
Maturation - Changes in subjects that take place during the experiment that are not related to the experiment but may affect their response to the experimental factor.
Mean - Commonly known as the arithmetic average. It is computed by adding all the scores in a distribution and dividing by the total number of scores. It can also be thought of as the amount each individual would get if the total were divided equally among all individuals in the distribution.
Measurement Scales - There are four types: nominal scales, ordinal scales, interval scales, and ratio scales. These scales are the levels at which elements can be measured.
Measures of Central Tendency - Identify what is seen as a middle point of a data set. The goal is to find a single value that provides a reasonably accurate description of the entire group of scores. There are three measures of central tendency: mode, median, and mean.
Median - The middle number when the measurements in a data set are arranged in ascending or descending order. It is less sensitive than mean to extreme scores because it doesn't respond to the value of the scores, only to their location in a distribution.
Method Variance - The variation in scores attributable to the method of data collection.
Middle Quartile - The median of a data set.
Mode - The value in a distribution that occurs most frequently. It is the only measure of central tendency that can be used for a nominal level value. Mode has three specific limitations: a given data set may not have a mode, a data set may have more than one mode, and mode is not the most representative value of a distribution, as it emphasizes data concentration.
Moderator - The person asking the questions and directing the discussion in a focus group.
Multichotomous Question - Fixed-alternative question in which respondents are asked to choose the alternative that most closely corresponds to their position on the subject.
Multidimensional Scaling - Approach to measurement in which people's perceptions of the similarity of objects and their preferences among the objects are measured, and these relationships are plotted in a multidimensional space.
Multivariate - Problem of analysis in which there are two or more measures of each of n sample objects, and the variables are to be analyzed simultaneously.
Mystery shoppers - Researchers posing as customers to observe the quality of service offered by retailers.
A | B | C | D | E | F | G | H | I | J | K | L | M | N - Z |
