Helpful Definitions
A - M, N - Z

A - M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

N
Negatively Skewed Distribution - Refers to a distribution where the mode has the largest value, followed by the median, then the mean. The graph of this type of distribution will show most of the scores piled up on the right hand side of the distribution, with a tail tapering to the left.

Nominal Scales - At the lowest level of measurement, nominal scales are qualitative in nature. They do not have magnitude or any other mathematical attribute; they describe a state of being. They are measured in terms of non-numerical qualities, dividing characteristics into categories. Examples are gender, hair color, and religion.

Noncoverage Error - Nonsampling error that arises because of a failure to include some units or entire sections of the defined survey population in the actual sampling frame.

Nonobservation Error - Nonsampling error that arises because of nonresponse from some elements designated for inclusion in the sample.

Nonparametric Statistics (Distribution-Free) - Methods developed to be used in cases when nothing is known about the parameters of the variable of interest in the population. These methods compare the probability distributions of the sampled populations rather than the specific parameters (such as means or variances).

Nonprobability Sample - Sample that relies on personal judgment somewhere in the element-selection process and therefore prohibits estimating the probability that any population element will be included in the sample.

Nonresponse Error - Nonsampling Error that represents a failure to obtain information from some elements of the population that were selected and designated for the sample.

Nonsampling Errors - Errors that arise in research that are not due to sampling; nonsampling errors can occur because of errors in conception, logic, misinterpretation of replies, statistics, and arithmetic; errors in tabulating or coding; or errors in reporting the results.

Null Hypothesis - States there is no change, effect, difference, or relationship between two variables.

O
Observation - Method of data collection in which the situation of interest is watched and the relevant facts, actions, or behaviors are recorded.

Observation Error - Nonsampling error that arises because inaccurate information is secured from the sample elements or because errors are introduced in processing the data or in reporting the findings.

Office Error - Nonsampling error that arises in the processing of the data because of errors in editing, coding, entering, tabulating, or some other part of the analysis.

Open-Ended Questions - Questions characterized by the condition that respondents are free to reply in their own words rather than being limited to choosing from among a set of alternatives.

Operational Definition - Definition of a construct that describes the operations to be carried out in order for the construct to be measured empirically.

Ordinal Scales - Measure magnitude and are quantitative in nature. They allow one to rank order the elements being measured according to whether they possess more, less, or the same amount of the variable being measured. Respondents are frequently given a list of several factors and asked to rank them in order of importance.

Outlier - Observation so different in magnitude from the rest of the observations that the analyst chooses to treat it as a special case.

Overcoverage Error - Nonsampling error that arises because of the duplication of elements in the list of sampling units.

P
Paired Difference Experiment - Experiment in which observations are paired and differences are analyzed.

Parameters - Calculations made when dealing with a population. A parameter is always descriptive, because data is being collected on every element in the set, and therefore no inferences can be drawn.

Parametric Tests - Class of statistical tests used when the variable is measured on at least an interval scale.

Pearson's Correlation (Coefficient of Correlation) - Measures the degree and direction of the linear relationship between two variables. Takes on values between -1 and 1. A perfect (strongest) correlation is indicated by -1 or 1, and 0 indicates no correlation (weakest).

Percentile Ranking - The rank or percentile rank of a particular score is defined as the percentage of individuals in the distribution with scores at or below the particular value. When a score is identified by its percentile rank, the score is called a percentile.

Performance of Objective Tasks - Method of assessing attitudes that rests on the presumption that a subject's performance of a specific assigned task will depend on the person's attitude.

Personal Interview - Direct, face-to-face conversation between a representative of the research organization (the interviewer) and a respondent, or interviewee.

Population - Refers to the total group of elements that conforms to some designated specifications.

Positively Skewed Distribution - Refers to a distribution where the mean has the largest value, followed by the median, then the mode. The graph of this type of distribution will show most of the scores piled up on the left hand side of the distribution, with a tail tapering to the right.

Power - Function associated with a statistical test indicating the probability of correctly rejecting a false null hypothesis.

Pragmatic Validity - Also called predictive validity or criterion-related validity. Approach to validation of a measure based on the usefulness of the measuring instrument as a predictor of some other characteristic or behavior of the individual.

Precision - Desired size of the estimating interval when the problem is one of estimating a population parameter.

Pretest - Use of a questionnaire on a trial basis in a small pilot study to determine how well the questionnaire works.

Primary Data - Information collected specifically for the purpose of the investigation at hand.

Primary Source - Originating source of secondary data.

Probability Sample - Sample in which each population element has a known, nonzero chance of being included in the sample.

Proportionate Stratified Sampling - Stratified sample in which the number of observations in the total sample is allocated among the strata in proportion to the relative number of elements in each stratum in the population.

Q
Qualitative Marketing Research - Research that yields an in-depth understanding about an issue. Qualitative research typically focuses on a small number of people. Since these people are interviewed in-depth, interviews tend to be longer and are often unstructured. An outline of discussion points, rather than a questionnaire is often used. This type of research also tends to be conducted in person, either in focus groups or one-on-one interviews, although the Internet is a growing medium. The Process of Marketing and Opinion Research

Quantitative Marketing Research - Research used to statistically estimate the viewpoints of a population providing estimates of percentages or averages. This research usually employs large samples and takes small amounts of a respondent’s time. Telephone, mail, and intercept, door-to-door, Internet or Web surveys, central location tests, mystery shopping and in-home use studies are all used in quantitative research.

Quota Sample - Nonprobability sample chosen in such a way that the proportion of sample elements possessing a certain characteristic is approximately the same as the proportion of the elements with the characteristic in the population; each field worker is assigned a quota that specifies the characteristics of the people he or she is to contact.

R
Random Error - Error in measurement due to the transient aspects of the person or measurement situation.

Randomized Block Design - Experiment design in which units are arranged into homogeneous groups in order to eliminate differences between groups and increase experimental accuracy.

Randomized Response Model - Interviewing technique in which potentially embarrassing and relatively innocuous questions are paired, and the question the respondent answers is randomly determined.

Range - The distance between the largest score and the smallest score in a distribution.

Rating Scales - Questions that ask respondents to assign specific values to a characteristic in order to establish a common basis for comparison. There are four types of rating scales, each allowing a certain level of analysis on the data collected: nominal, ordinal, interval, and ratio.

Ratio Scales - At the highest level of measurement, they have the attributes of magnitude, equal intervals, and absolute zero. Measurements can range from nonexistent to existent, but never negative. Examples are height, age, weight, and income.

Refusals - Nonsampling error that arises because some designated respondents refuse to participate in the study.

Relative Frequency - A proportion of the frequency of an interval in relation to the total number of frequencies of the distribution.

Reliability - Similarity of results provided by independent but comparable measures of the same object, trait, or construct.

Research Design - Framework or plan for a study that guides the collection and analysis of the data.

Research Process - Sequence of steps in the design and implementation of a research study, including problem formulation, determination of sources of information and research design, determination of data collection method and design of data collection forms, design of the sample and collection of the data, analysis and interpretation of the data, and the research report.

Response Rate - Measure used to evaluate and compare interviewers in terms of their ability to induce contacted respondents to participate in the study.

S
Sample - Refers to a subset of elements selected from a population. A sample should be randomly selected and representative of the population.

Sampling Frame - The list of population elements from which the sample will be drawn.

Sample Survey - Cross-sectional study in which the sample is selected to be representative of the target population and in which the emphasis is on the generation of summary statistics such as averages and percentages. Also called a field survey.

Sampling Control - Term applied to studies relying on questionnaires and concerning the researcher's dual abilities to direct the inquiry to a designated respondent and to secure the desired cooperation from that respondent.

Sampling Distribution - Distribution of values of some statistic calculated for each possible distinguishable sample that could be drawn from a parent population under a specific sampling plan.

Sampling Error - Difference between the observed values of a variable and the long-run average of the observed values in repetitions of the measurement.

Sampling Units - Nonoverlapping collections of elements from the population.

Secondary Data - Statistics not gathered for the immediate study at hand but for some other purpose.

Secondary Source - Source of secondary data that did not originate the data but secured them from another source.

Select-a-Response Question - Questions offering a set of options for a respondent to chose from as a response.

Selection Bias - Contaminating influence in an experiment occurring when there is no way of certifying that groups of test units were equivalent at some previous time.

Self-Report - Method of assessing attitudes in which individuals are asked directly for their beliefs about or feelings toward an object or class of objects.

Sentence Completion - Questionnaire containing a number of sentences that subjects are directed to complete with the first words that come to mind.

Sequence Bias - Distortion in the answers to some questions on a questionnaire because the replies are not independently arrived at but are conditioned by responses to other questions.

Sequential Sample - Sample formed on the basis of a series of successive decisions. If the evidence is not conclusive after a small sample is taken, more observations are taken; if still inconclusive after these additional observations, still more observations are taken. At each stage, a decision is made about whether more information should be collected or whether the evidence is sufficient to draw a conclusion.

Simple Linear Regression - The statistical technique for finding the best fitting straight line for a set of data. This technique is called regression and the resulting line is called the regression line or least squares line.

Simple Random Sample - Probability sample in which each population element has a known and equal chance of being included in the sample and in which every combination of n population elements is a sample possibility and is just as likely to occur as any other combination of n units.

Simple Regression and Correlation - Procedures that examine the relationship between two interval-ratio variables for the same elements.

Simple Tabulation - Count of the number of cases that fall into each category when the categories are based on one variable.

Skip Pattern - The logical organization of an interview or questionnaire so that questions are asked only of those who fit certain criteria. Other respondents are directed elsewhere in the questionnaire depending on the criteria they meet.

Snowball Sample - Judgment sample that relies on the researcher's ability to locate an initial set of respondents with the desired characteristics; these individuals are then used as informants to identify still others with the desired characteristics.

Spearman's Rank Correlation Coefficient - Nonparametric measure of dependence between two variables based on the correlation between the ranks of the observations.

Spurious Correlation - Condition that arises when there is no relationship between two variables but the analyst concludes that a relationship exists.

Spurious Noncorrelation - Condition that arises when the analyst concludes that there is no relationship between two variables but, in fact, there is.

Stability - A technique for assessing the reliability of a measure by measuring the same objects or individuals at two different points in time and then correlating the scores; the procedure is known as test-retest reliability assessment.

Standard Deviation - The positive square root of the average squared distance of the population or sample values from the mean. It is the most widely accepted measure of dispersion.

Standard Error - Standard deviation of the distribution of sample means. It is the standard distance between the sample means and the population means by chance (how much error to expect by chance).

Standard Error of Estimate - Term used in regression analysis to refer to the absolute amount of variation in the criterion variable that is left unexplained or unaccounted for by the fitted regression equation.

Standard Normal Distribution - The distribution of z-scores, which always has a mean of 0 and a standard deviation of 1.

Standardized Scores (z-scores) - To standardize a distribution, scores (x values) are transformed into z-scores by using the mean and standard deviation. A z score specifies the precise location of each x value within a distribution. The z-scores will form a standardized distribution that can be directly compared to other distributions that also have been transformed into z-scores. A z score has two components: sign and magnitude. The sign of the z-score (+ or -) signifies whether the score is above the mean (+) or below the mean (-). The numerical value of the z-score specifies the distance from the mean by counting the number of standard deviations between the score and the mean.

Statistics - Calculations made when dealing with a sample. Statistics can be used to describe the sample, make inferences to the population (estimate parameters), or test hypotheses.

Stereotyping - The tendency to assign an individual to a group or broad category and then attribute generalizations about the group to the individual.

Stratified Sample - Probability sample that is distinguished by the two-step procedure in which (1) the parent population is divided into mutually exclusive and exhaustive subsets, and (2) a simple random sample of elements is chosen independently from each group or subset.

Structure - Degree of standardization imposed on the data collection instrument.

Symmetrical Distribution - When the values in a distribution are graphed, the right hand side of a symmetrical distribution is a mirror image of the left hand side. In a symmetrical distribution with one mode, all three measures of central tendency are all on one value.

Systematic Error - Error in measurement that is also known as constant error, since it affects the measurement in a systematic way.

Systematic Sample - Probability sample in which every kth element in the population is designated for inclusion in the sample after a random start.

T
T-test (one sample) - Hypothesis test about a single mean if the sample is too small to use the Z-test.

Telephone Interview - Telephone conversation between a representative of the research organization, the interviewer, and a respondent or interviewee.

Testing Effect - Contaminating effect in an experiment occurring because the process of experimentation itself affected the observed response.

Tracking Study - Research that repeatedly measures the same variables and determines their changes over time.

Type I Error - Rejecting a null hypothesis that is actually true (concluding that a treatment does have an effect when in fact the treatment has no effect).

Type II Error - Failing to reject a null hypothesis that is actually false (concluding that a treatment has no effect when in fact there actually is an effect).

U
Unaided Questions - Questions where respondents search their awareness for an answer. The respondent is not prompted.

Unbiased - Used to describe a statistic when the average value of the statistic equals the population parameter it is supposed to estimate.

Univariate - Problem of analysis in which there is a single measurement on each of n sample objects or there are several measurements on each of the n observations, but each variable is to be analyzed in isolation.

Upper Quartile - The 75th percentile of a data set.

V
Validity - Term applied to measuring instruments reflecting the extent to which differences in scores on the measurement reflect true differences among individuals, groups, or situations in the characteristic that it seeks to measure, or reflect true differences in the same individual, group, or situation from one occasion to another, rather than constant or random errors.

Variability - The degree of differences from a score. It provides a quantitative measure of the degree to which scores in a distribution are spread out or clustered together. If differences are small, the variability is small. If differences are large, the variability is large. It also provides information about how much error to expect if you are using a sample to represent a population. The three measures of variability are range, standard deviation, and variance.

Variance - The statistical measure of how similar a population is in a characteristic being studied. It is the average squared distance of all measurements from the mean. It is the square of the standard deviation.

Verbatim - Respondents' true and actual responses.

W
Wilcoxon Rank Sum Test - For data that can be ranked in order of magnitude, then this nonparametric test for independent samples can be used to test the hypothesis that the probability distributions associated with the two populations are equivalent. A nonparametric version of the t-test.

Wilcoxon Signed Rank Test - Nonparametric test used to compare two probability distributions when a paired difference design is used.

Z
Z-test - Hypothesis test about a single mean if the sample is large enough and drawn from a normal population.

A - M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z