Glossary

"Don't know" (DK) responses	A response provided by respondents when they have insufficient knowledge to answer the question, when the instrument fails to provide an understandable operational definition for a construct, when the respondents have not formed a judgment on an issue, are reluctant to provide an answer, or feel the issue is too unimportant to formulate an answer.



3-D graphics	A statistical presentation technique that adds interest to obscure data and permits you to compare three or more variables from the sample in one chart; types: column, ribbon, wire frame, and surface line.



A priori contrasts	A special class of tests used after the null hypothesis was rejected with the F test to compare measurements of specific populations.



Action Research	A research approach that emphasizes that interplay between research and specific actions to achieve desired changes.



Active factors	Those independent variables (IV) the researcher can manipulate by causing the subject to receive one treatment level or another.



Administrative questions	A type of measurement question that identifies the respondent, interviewer, interview location, and conditions (usually nominal data).



Alternative hypothesis	That a difference exists between the sample parameter and the population statistic to which it is compared; the logical opposite of the null hypothesis used in significance testing (notation: HA).



Analysis of variance (ANOVA)	A statistical test for testing the null hypothesis that the means of several populations are equal; test statistic is the F ratio; used when you need k independent samples tests.



Applied research	Uses theories to investigate real phenomena that are of relevance for practitioners. Applied research often builds upon basic research. It has an emphasis on solving practical problems.



Arbitrary scales	Universal practice of ad hoc scale development used by instrument designers to create scales that is highly specific to the practice or object being studied.



Area charts	A statistical presentation technique used for time series and frequency distributions over time; a.k.a. stratum or surface charts.



Area sampling	A type of cluster sampling usually applied to a population in a specific spatial area with well-defined political or natural boundaries but without a detailed sample frame; population is divided into homogeneous clusters from which a single-stage or multistage sample is drawn.



Artefact correlations	Where distinct groups combine to give the impression of one.



Asymmetrical relationship	A relationship in which we postulate that change in one variable (independent variable) is responsible for change in another variable (dependent variable).



Audience	The intended reader of the secondary source; one of the five factors used to evaluate the value of a secondary source.



Authority	The credibility of a secondary source as indicated by the credentials of the author and publisher; one of five factors used to evaluate the value of a secondary source.



Backward elimination	In modelling and regression, one of the methods for sequentially adding or removing variables; begins with a model containing all independent variables and removes the variable that changes R²the least; see also backward elimination and stepwise selection.



Bar charts	A statistical presentation technique that represents frequency data as horizontal or vertical bars; vertical bars are most often used for time series and quantitative classifications (histograms, stacked bar, and multiple variable charts are specialized bar charts).



Bar code	Technology used to simplify the researcher's role as a data recorder, involving a label reading device that scans electronically codes (stripes) on product labels and service documents.



Basic research	Attempts to understand processes and their outcomes, which form the fundaments for explaining levels of and changes in characteristics, attitudes and behaviour. Furthermore, theory development is mostly part of basic research. It is mostly conducted within university and research institutes. In business research, basic research often provides the theoretical framework used in applied research. Basic research is also called pure research.



Bayesian statistics	An approach that goes beyond sampling data for making decisions by including all available data – specifically subjective probability estimates based on general experience rather than on specific data collected (see Appendix B).



Beta weights	Another term for standardized regression coefficients, where the size of the number reflects the level of influence on Y that an individual X exerts.



Bibliographic database	An electronic database containing the bibliographic information plus abstract and increasingly full texts. EBSCO is an example of such a database. A researcher often sets up his/her own databases with all the articles and books they have read or referred to.



Bibliography	A list of the bibliographic details (usually authors name, date of publication, title of book or article, title of journal, pages, publisher, editors) of the sources used, including those not referred to in the text. Most scientific journals but also thesis guidelines ask for a list of references, which contains only the sources referred to in the text.



Bivariate correlation analysis	Measures of correlation that use non-contiguous variables and that distinguish between independent and dependent variables.



Bivariate normal distribution	An assumption of correlation analysis, that data are from a random sample of a population where the two variables are normally distributed in a joint manner.



Blind	A condition that exists when subjects do not know if they are being exposed to the experimental treatment. See also double blind.



Blocking factors	A variable on which the researcher can only identify and classify a subject – not manipulate (e.g. gender, age, customer status, etc.); these are often the classification variables within a questionnaire.



Box plot	An EDA technique that reduces the detail of the stem-and-leaf display but provides a visual image of the variable's distribution location, spread, shape, tail length, and outliers; a.k.a. box-and-whisker plot.



Branched questions	A type of measurement question that determines the respondent's path (question sequencing) in a study; the answer to one question assumes other questions have been asked or answered, and directs the respondent to answer specific questions that follow and skip other questions.



Buffer question	A type of neutral measurement question designed chiefly to establish rapport with the respondent (usually nominal data or qualitative data).



Business research	Refers to studies dealing with phenomena in the business world including nonprofit and (semi) governmental organizations to guide their decisions.



Call-backs	Procedure involving repeated attempts to make contact with a targeted respondent to ensure that the targeted respondent is reached and motivated to participate in the study. They are important to reduce the non-response error.



Case study	Emphasizes the full and detailed contextual analysis of a single or fewer events or conditions and their interrelations for a single subject or respondent. It often relies on multiple sources to crosscheck the obtained information.



Case study research	A research approach that investigates the phenomenon in question in its context. It is the most suitable research if the number of variables that needs to be considered is very large and if the phenomena and its antecedents can not yet be clearly distinguished.



Categorization	A type of scale in which the respondents put themselves or property indicants in groups or categories.



Causal hypothesis	See explanatory hypothesis.



Causal method	See causal study.



Causal study	Attempts to reveal the relationship between variables (A produces B or causes B to occur).



Causation	The essential element of causation is that A produces B or forces B to occur. Causation is always based on an inductive [induction] conclusion and are therefore a probabilistic statement, as we can not account for all possible, imaginable and unimaginable processes which may drive the relation between A and B.



Cells	In a cross-tabulation, a subgroup of the data created by the value intersection of two (or more) variables, where each cell contains the count of cases as well as the percentage of the joint classification.



Census	A sample that contains all element of the population. For example for the UK census all people (~59 millions) living in the UK are interviewed.



Central limit theorem	For sufficiently large samples (n > 30), the sample means of repeatedly drawn samples will be distributed around the population mean approximately in a normal distribution.



Central tendency	A measure of location, most commonly the mean, median, and mode.



Central tendency (error)	An error that results because the respondent is reluctant to give extreme judgments, usually due to lack of knowledge.



Centroid	A multi-variate mean.



Checklist	A measurement question that poses numerous alternatives and encourages multiple responses, but where relative order of those responses is not important (nominal data).



Chi-square (X²) test	A non-parametric test of significance used for nominal measurements.



Chi-square-based measures	Tests to detect the strength of the relationship between the variables tested with a chi-square test: phi, Cramer's V, and contingency coefficient C.



Classical statistics	An objective view of probability in which the decision-making rests totally on an analysis of available sampling data where a hypothesis is rejected or false based on the sample data collected.



Classification	In data mining, using a set of reclassified examples to develop a model that can group or classify the population of records at large.



Classification question	A type of measurement question that provides sociological-demographic variables for use in grouping respondent's answers nominal, ordinal, interval or ratio data).



Closed question/response	A type of measurement question that presents the respondent with a fixed set of choices (nominal, ordinal or interval data).



Cluster analysis	A technique that identifies homogeneous subgroups or clusters of study objects or people, then displays the relevant clusters in a diagram (endogamy); viewing the data by segments or groups of data cases.



Cluster sampling	A sampling plan that involves dividing the population into clusters or subgroups, and then drawing a sample from each subgroup in a single-stage or multi-stage design.



Clustering	A data mining technique that assigns each data record to a group or segment automatically by clustering algorithms that identify the similar characteristics in the data set and then partition them into groups.



Code of ethics	A comprehensive source that contains the firm's policies with respect to ethical conduct [ethics]; effective codes are regulative, protect the public interest, are behaviour specific, and are enforceable.



Codebook	Contains the coding rules for assigning numbers or other symbols to each variable; a.k.a. coding scheme.



Coding	Assigning numbers or other symbols to answers to responses that can be tallied and grouped into a limited number of classes or categories.



Coefficient of determination	R²= the amount of common variance in X and Y, two variables in association; is the ratio of the line of best fit's error over that incurred by using mean value of Y.



Collinearity	The situation when two or more of the independent variables are highly correlated; causes estimated regression coefficients to fluctuate widely, making interpretation difficult; aka multicollinearity.



Communalities	In factor analysis, the estimate of the variance in each variable that is explained by the factors being studied.



Communication approach	A study approach involving questioning or surveying people (by personal interview, telephone, mail, computer, or some combination of these) and recording their responses for analysis.



Comparative scale	A scale where the respondent evaluates an object against a standard using a numeric, graphical, or verbal scale.



Computer-administered telephone survey	A study conducted wholly by computer contact between respondent and interviewer, where questions are either appear on the computer or are voice-synthesized and data are tallied continuously.



Computer-assisted personal interviewing (CAPI)	A personal, face-to-face interview where the researcher may be guided by computer-sequenced questions, where data may be entered as responses are given, or where visualization techniques may be provided digitally to each participant.



Computer-assisted telephone interviewing (CATI)	A study conducted wholly by telephone contact between respondent and interviewer where interview is software-driven, usually in a central location with interviewers in acoustically isolated interviewing carrels; data are tallied, as they are collected.



Concealment	A technique in an observation study where the observer is shielded from the subject to avoid behaviour modification by the subject caused by observer presence; this is accomplished by one-way mirrors, hidden cameras, hidden microphones, etc.



Concept	A generally accepted bundle of meanings or characteristics associated with certain events, objects, conditions, or situations.



Conceptual scheme	The interrelationships between concepts and constructs.



Concordant	When a subject that ranks higher on one ordinal variable also ranks higher on another variable, the pairs of variables are concordant.



Confidence interval	The combination of interval range and the degree of confidence. We are confident (to the stated degree) that the true value of mean lies within the interval.



Confidentiality	A privacy guarantee to retain validity of the research, as well as to protect respondents. It can refer to the sponsor, i.e. the sponsor of the study remains unknown to the respondents, and to the respondents, i.e. the information provided by the respondents will only be revealed in a form that guarantees that an individual respondent cannot be identified and linked to the information provided.



Confirmatory data analysis	An analytical process guided by classical statistical inference in its use of significance and confidence.



Confirmatory factor analysis (CFA)	A statistical procedure, which differs from exploratory factor analysis in that common factors are assumed to be uncorrelated, observed variables are affected by only some of the common factors, and the variables, which define the construct, are researcher defined.



Confounding variable	An extraneous variable that influencing the relation between independent [independent variable] and dependent variable, similar to a moderating variable.



Conjoint analysis	A technique that uses input from non-numeric independent variables to secure part worth's that represent the importance of each aspect in the subject's overall assessment; used to measure complex decision making (e.g. consumer purchase behaviour) that requires multi attribute judgments; produces a scale value for each attribute or property. See also factorial survey.



Consensus scaling	Scale development by a panel of experts evaluating instrument items based on topical relevance and lack of ambiguity.



Construct	A definition specifically invented for an image or idea for a given research project.



Construct validity	The degree to which a research instrument is able to measure or infer the presence of an abstract property. See also validity.



Content analysis	A technique used to analyse qualitative data, such as texts by counting the occurrence of certain key words, the length of sentences and the text etc.



Content validity	The degree to which a research instrument provides adequate coverage of the topic under study. See also validity.



Contingency tables	A cross-tabulation table constructed for statistical testing, with the test determining whether the classification variables are independent.



Continuous variable	A variable that can take any value in a given range. Income, temperature and age are examples for a continuous variable.



Control	The ability to replicate a scenario and dictate a particular outcome; the ability to exclude, isolate, or manipulate the influence of a variable in a study; a critical factor in inference from an experiment, implies that all factors, with the exception of the independent variable (IV), must be held constant and not confounded with another variable that is not part of the study.



Control dimension	In quota sampling, a descriptor used to define the sample's characteristics (e.g. age, education, religion, etc.).



Control group	A group of subjects (respondents) that is not exposed to the independent variable IV being studied but still generates a measure for the dependent variable DV. Comparing the outcomes of the DV for the control and experimental group allows an assessment of the IV-DV relationship.



Control variable	A variable that is assumed or discounted in a study; in data analysis, a variable introduced to help interpret the statistically significant [significance] relationship between variables.



Convenience samples	A low-cost but less reliable non-probability sample where element selection is unrestricted or left to those elements easily accessible by the researcher. For example, if you want to investigate bookshops and select those you know.



Correlation hypothesis	Variables occur together in some specified manner without implying that one causes the other (See also hypothesis).



Correlation matrix	A table used to display coefficients for more than two variables.



Cramer's V (used with chi-square)	A measure of association for nominal, non-parametric variables; ranges from zero to +1.0 and used for larger than 2 × 2 chi-square tables; does not provide direction of the association or reflect causation.



Criterion variable	Alternative term for dependent variable.



Criterion-related validity	The success of measures used for prediction or estimation; types are predictive and concurrent. See also validity.



Critical path method (CPM)	A scheduling tool for complex or large research proposals that cites milestones and time involved between milestones.



Critical value	The dividing point(s) between the region of acceptance and the region of rejection; these values can be computed in terms of the standardized random variable due to the normal distribution of sample means.



Cross-sectional study	The study is conducted only once and reveals a snapshot of one point in time.



Cross-tabulation	A technique for comparing two classification variables (usually nominal data variables).



Cumulative scaling	A scale development technique in which scale items are tested based on a scoring system, where agreement with one extreme scale item results also in endorsement of all other items that take a less extreme position.



Data	Facts (attitudes, behaviour, motivations, etc.) collected from respondents or observations (mechanical or direct) plus published information; categorized as primary and secondary.



Data analysis	Editing and reducing accumulated data to a manageable size, developing summaries, looking for patterns, and applying statistical techniques.



Data entry	The process of converting information gathered by secondary or primary methods to a medium for viewing and manipulation; usually done by keyboarding or optical scanning.



Data fields	A single element of data (e.g. a single answer to a question).



Data files	Sets of data records or sets of data cases.



Data mart	Intermediate storage facilities that compile locally required information.



Data mining	Data mining is a technique to detect relationships and patterns in very large databases, often organized in data warehouses [data warehousing]. It is a tool combining exploration and discovery with confirmatory analysis.



Data visualization	The process of viewing aggregate data on multiple dimensions to gain a deeper, intuitive understanding of the data.



Data warehousing	A data warehouse is an electronic repository for databases that organizes very large amounts of data into categories to facilitate retrieval, sorting and interpretation. It is an accessible archive of information to support dynamic organizational intelligence applications.



Database	A collection of data organized for computerized retrieval that defines data fields, data records, and data files.



Debriefing	Explains the truth to participants and describes the major goals of the research study and the reasons after deception has been used.



Deception	When respondents are told only part of the truth or the truth is fully compromised to prevent biasing respondents or to protect sponsor confidentiality. Balancing deception and informed consent is an important ethical issue.



Decision tree models	A data mining technique that segregates data by using a hierarchy of if-then statements based on the values of variables and creates a tree-shaped structure that represents the segregation decisions; used with interval or categorical data.



Deduction	A form of inference in which the conclusion must necessarily follow from the reasons given. If the reasons (premises) are true and the conclusion follows necessarily from the reasons, the deduction is valid (See also induction).



Dependency techniques	Those techniques where criterion or dependent variables and predictor or independent variables are present (e.g. multiple regression, MANOVA, discriminate analysis, etc.).



Dependent variable (DV)	The variable measured, predicted [predictive study], or otherwise monitored by the researcher, expected to be affected by a manipulation of the independent variable. Notation usually DV.



Descriptive hypothesis	States the existence, size, form or distribution of some variable (See also hypothesis).



Descriptive statistics	Display characteristics of the location, spread, and shape of an array of data.



Descriptive studies	Descriptive studies sketch the current state of a phenomenon. Attempts to describe or define a subject, often by creating a profile of a group of problems, people or events, through the collection of data and the tabulation of the frequencies on research variables or their interaction; the study reveals who, what, when, where, or how much; the study concerns a univariate question or hypothesis in which the research asks about or states something about the size, form, distribution, or existence of a variable. They distinct from reporting studies by providing interpretations of the information found.



Dichotomous question	A measurement question that poses two opposing responses (nominal or ordinal data).



Dichotomous variable	A variable that only can take two values. Examples are gender (female or male) or yes – no variables, such as having children, being a foreign company. It generates nominal or ordinal data.



Dictionary	Dictionaries are books explaining the meaning of a word in other words. Next to general dictionaries, many specialized dictionaries exists, such as dictionaries for acronyms or specialized dictionaries for financial terms, business terms etc. Dictionaries are especially useful if you want to know a good definition of a term.



Direct observation	When the observer is physically present and personally monitors and records the behaviour of the subject.



Direct questions	Questions that ask the respondent to provide her/his own view on a matter.



Directory	Directories are books containing names, addresses and often further data. The phone directory is very simple, but could still be used to sample people in a certain city. Specialized directories include for example all companies in a specific sector and might provide next to contact address details further basic information on the firm.



Discordant	When a subject that ranks higher on one ordinal variable ranks lower on another variable, the pairs of variables are discordant; as discordant pairs increase over concordant pairs, the association becomes negative.



Discriminate analysis	The joining of a nominal dependent variable with one or more independent internal or ratio variables into an equation that is used to predict the classification of a new observation.



Disguised question	A measurement question designed to conceal the question's and study's true purpose.



Distribution (of data)	The array of value counts from lowest to highest value, resulting from the tabulation of incidence for each variable by value.



Double blind (study)	A condition that exists when neither the researchers nor the subjects know when a subject is being exposed to the experimental treatment (IV). See also blind.



Double movement of reflective thought	The sequential use of induction and deduction in research reasoning to develop a plausible hypothesis.



Double sampling	A procedure for selecting a sub sample from a sample for further study; a.k.a. sequential sampling or multiphase sampling. Such a sampling design is especially useful if you are only interested in a specific sub-group of the total population, but do not know which subject of the population belongs to the specific subgroup.



Double-barrelled question	A type of multiple questions that includes two or more questions in one that the respondent might need to answer differently to preserve the accuracy of the data.



Dummy variable	Nominal variable used in multiple regressions and coded 0, 1 as all other variables must be interval or ratio measures.



Editing	A customary first step in analysis for detecting errors and data omissions, correcting them when possible, and certifying that minimum data quality standards are achieved.



Effect size statistics	It is used in meta analysis and describes the statistics which transform effect statistics of different empirical studies into one single effect.



Encyclopaedia	Encyclopaedias can be used to find some background information on a topic. For example, if you conduct an international study, you might use them to find the basic background information on the countries involved in your study. Encyclopaedias exist in printed form but also in electronic form. The former look nicer and more prestigious on the shelf, the latter are more convenient as they include advanced search options and a lot of cross-references.



Environmental control	Holding constant the (physical) environment of the experiment.



Equal-appearing interval scale	An expensive, time-consuming type of consensus scaling which results in an interval rating scale for attitude measurement, a.k.a. Thurston scale.



Error term	The deviations of the actual values of Y from the regression line representing the mean value of Y for a particular value of X.



Ethics	Norms or standards of behaviour that guide moral choices about research behaviour.



Ethnographic studies	Research approach that emphasizes the description and interpretation of the social world by primary qualitative [qualitative research] information collection, such as direct observation.



Event sampling	The process of selecting some elements or behavioural acts or conditions from a population of observable behaviour or conditions to represent the population as a whole.



Ex post facto design	Researchers have no ability to manipulate the variables, but can report what has happened or is happening to the variables. This design is very common in business research.



Executive summary (final report)	This document is written as the last element of a research report and is either a concise summary of the major findings, conclusions and recommendations or can be a report-in-miniature covering all aspects in abbreviated form.



Executive summary (proposal)	An informative abstract providing the essentials of the proposal without the details.



Experience survey	An exploratory technique where knowledgeable experts share their ideas about important issues or aspects of the subject and relate what is important across the subject's range of experience; usually involves a personal or phone interview.



Experiment	Studies involving intervention (manipulation of one or more variables) by the researcher beyond that required for measurement to determine the effect on another variable.



Experimental treatment	The manipulated independent variable.



Explanatory hypothesis	Statement that describes a relationship between two variables with respect to some case, one variable leads to an effect on the other variable (a.k.a. causal hypothesis).



Explanatory studies	Explanatory studies go beyond descriptive studies. They attempt to explain the reasons for phenomena by using theories or at least derived hypotheses [hypothesis] and provide answers to questions starting with why or how?



Exploration	See exploratory study.



Exploratory data analysis (EDA)	A process whereby the actual data patterns guide the data analysis or suggest revisions to the preliminary data analysis plan.



Exploratory study	The process of collecting information to formulate or refine management, research, investigative, or measurement questions; loosely structured studies that discover future research tasks, including developing concepts, establishing priorities, developing operational definitions, and improving research design; a phase of a research project where the researcher expands understanding of the research dilemma, looks for ways others have addressed and/or solved problems similar to the research dilemma or research question, and gathers background information on the topic to refine the research question.



Extemporaneous presentation	An oral presentation technique made from minimal notes or an outline, with a more conversational style.



External validity	When an observed causal relationship can be generalized across persons, settings, and times.



Extra linguistic behaviour	The recording of vocal, temporal, interaction, and verbal stylistic behaviours of human subjects.



Extraneous variable	Variables to assume (because they have little effect, or their impact is randomised) or exclude from a research study; notation: EV.



F ratio	The result of an F test, done to compare measurements of k independent samples.



Factor	In experiments it denotes an independent variable (IV); these are divided into treatment levels for the experiment.



Factor analysis	Techniques for discovering patterns among the variables to determine if an underlying combination of the original variables (a factor) can summarize the original set.



Factor scales	Types of scales that deal with multidimensional content and underlying dimensions, such as scalogram, factor, and cluster analyses, and metric and non metric multidimensional scaling.



Factorial survey	A survey technique in which subjects are asked to assess pre-described situations, which are made up from different combinations of vignettes. A.k.a. vignette studies (See also conjoint analysis).



Falsification	Refers to the intentional change or fabrication of data and results. It is a serious ethical offence.



Field condition	The actual environmental conditions where the research studies occurs.



Field experiment	A study that occurs under the actual environmental conditions where the dependent variable occurs and is measured.



Filter question	A question used to qualify the respondent's knowledge about the target questions of interest.



Findings nondisclosure	A type of confidentiality, when the sponsor restricts the researcher from revealing the findings of the research project to third parties and does not allow publishing of the results (see also nondisclosure).



Five-number summary	The median, upper and lower quartiles, and the largest and smallest observations of a variable's distribution.



Fixed sum scale	A scale where the respondent assigns mostly 100 points to different continuous or discrete categories, e.g. how much time he spends on average per day on specified activities, such as paid working, commuting, child care, household work, leisure. It generates interval data.



Focus group	An information collection approach widely used in exploratory studies involving a panel of subjects led by a trained moderator that meets for 90 minutes to two hours; the moderator uses group dynamics to explore ideas, feelings, and experiences on a specific topic; can be conducted in person or via phone.



Follow-up questions	Questions that are asked to motivate the respondents to elaborate deeper on an answer given.



Forced ranking scale	A scale where the respondent orders several objects or properties of objects; faster than paired comparison to obtain a rank order.



Formal study	Begins with a hypothesis or research question and involves precise procedures and data source specifications; tests the hypothesis or answers the research questions posed.



Format	How the information is presented and how easy it is to find a specific piece of information within a secondary source; one of five factors used to evaluate the value of a secondary source.



Forward selection	In modelling and regression, one of the methods for sequentially adding or removing variables; starts with the constant and adds variables that result in the largest R²increase; see also backward elimination and stepwise selection.



Fractal based transformation	A technique that can work on gigabytes of data and offers the possibility of identifying tiny subsets of data that have common characteristics.



Free-response question	A type of measurement question to which the respondent provides the answer without the aid of an interviewer (either in phone, personal interview or self-administered surveys); aka open-ended question (generates all kind of data).



Frequency table	A device for arraying data from lowest value to highest value, with columns for percent, percent adjusted for missing values, and cumulative percent.



Funnel approach	A type of question sequencing that moves the respondent from general to more specific questions and is designed to learn the respondent's frame of reference while extracting full disclosure of information on the topic (nominal, ordinal, interval or ratio data).



Fuzzy logic	An extension of conventional (Boolean) logic that handles the concept of partial truth – with truthvaluesbetween"completelytrue"and"completely false"; used in more complex data mining.



Gamma	Uses a preponderance of evidence of concordant pairs vs. discordant pairs to predict association; the gamma value is the proportional reduction of error when prediction is done using preponderance of evidence (values from –1.0 to +1.0).



Genetic algorithms	Optimisation techniques for search and identification of meaningful relationships.



Goodness of fit	A measure of how well the regression model is able to predict Y.



Graphic rating scale	A scale where the ratter places his or her response along a line or continuum; the score or measurement is its distance in millimetres from either end point.



Halo effect (error)	A systematic bias that the rater introduces by carrying over a generalized impression of the subject from one rating to another.



Handbook	A handbook is a collection of articles or facts around a topic. Those containing articles usually provide a good overview over the current state of the field and are useful to identify the most prominent articles and debates in the field.



Histogram	A bar chart data display technique that groups data values into equal intervals; especially useful for revealing skewness, kurtosis, and modal



Holdout sample	A portion of the sample – usually 1/3 or 1/4 – is set aside and only the remainder is used to compute the estimating equation. The equation is then used on the holdout data to calculate R² for comparison.



Hypothesis	A statement formulated for empirical testing; a tentative or conjectural declarative belief or statement that describes the relationship between two or more variables. One distinguishes descriptive [descriptive hypothesis], u, explanatory [explanatory hypothesis] and relational hypotheses [relational hypothesis].



Independent variable (IV)	The variable manipulated by the researcher, which causes an effect or change on the dependent variable. Notation IV.



In-depth interview	A type of interview, usually unstructured and in an unconstrained environment, which encourages the respondent to talk extensively, sharing as much information as possible.



Index	You find the index at the end of a book, it tells you on which page a specific term or name is mentioned. Index is also used for the word bibliography.



Indirect observation	When the recording of data is done by mechanical, photographic, or electronic means. For example if you study videotapes showing how children interact in a group.



Indirect questions	Questions that do not ask for what the respondent thinks, but what is generally thought according to her/him.



Induction	To draw a conclusion from one or more particular facts or pieces of evidence; the conclusion explains the facts.



Inferential statistics	Includes the estimation of population values and the testing of statistical hypotheses.



Informed consent	Respondent gives full consent to participation after receiving full disclosure of the procedures of the proposed survey. (See also deception)



Interaction effect	The influence that one moderating factor has on the relationship between an independent and a dependent variable.



Intercept	One of two regression coefficients, β₀, is the value for the linear function when it crosses the Y-axis or the estimate of Y when X is zero.



Intercept interview	A face-to-face communication that approaches passers-by to participate in an interview in a centralized location, such as a shopping centre.



Interdependency techniques	Those techniques where criterion or dependent variables and predictor or independent variables are not present (e.g. factor analysis, cluster analysis, multidimensional scaling, etc.).



Internal consistency	A characteristic of measurement in which an instrument measures consistency among responses of a single respondent. See also reliability.



Internal database	Internal databases are archives of information that are kept within an organisation and that are usually not freely accessible to everyone. Sometimes they are organized in data warehouses [data warehousing].



Internal validity	The ability of a research instrument to measure what it is purported to measure; when the conclusion(s) drawn about a demonstrated experimental relationship truly implies cause.



Interpreting questions	In such questions the interviewer repeats the answers of the respondents and asks whether the interviewer's interpretation of the answer is correct.



Interpretivism	A research philosophy built upon the principles that the social world is constructed and given subjective meanings by humans, and that the researcher is driven by interests and part of what is observed. It assumes that social phenomena can only be understood if one looks at the totality and how people give meaning and interpret the social world. (See also positivism)



Interquartile range (IQR)	A calculated statistic using the largest and smallest values in a variable's distribution that measures the distance between the first and third quartiles of the distribution; a.k.a. midspread; the distance between the hinges in a box plot.



Interrogation/communication study	The researcher questions the subjects and collects their responses by personal or impersonal means.



Interval data	Data with order and distance but no unique origin; data which incorporate equality of interval (the distance between one measure and the next measure); e.g. temperature scale.



Intervening variable	A factor that affects the observed phenomenon but cannot be seen, measured, or manipulated, thus its effect must be inferred from the effects of the independent [independent variable] and moderating variables on the dependent variable; notation: IVV.



Interview guide	A guide that states which topics should be covered in the interview and which information is sought.



Interview schedule	An alternative rarely used term for the questionnaire used in an interview (phone or in-person communication approaches to collecting data).



Interviewer error	Error that results from interviewer influence of the respondent; includes problems with motivation, instructions, voice inflections, body language, question or response order, or cheating via falsification of one or more responses.



Introductory questions	General questions that are asked to collect basic information on the respondent and that are usually easy to answer.



Investigative question	Questions the researcher must answer to obtain the information necessary to answer the research question. (See also management research question hierarchy)



Item analysis scaling	Scale development where instrument designers develop instrument items and test them with a group of respondents; individual items are analysed to determine those, which highly discriminate between persons or objects; e.g. Likert scale and summated scale.



Judgment sampling	A type of purposive sampling in which the researcher arbitrarily selects elements to conform to some criterion.



k independent samples test	The parametric test used when interval or ratio measurements are taken from three or more samples (ANOVA), and the non-parametric test used when nominal (chi-square) or interval (Kruskal-Wallis) measurements are taken from three or more samples.



k related samples test	The parametric (ANOVA) and nonparametric tests (Cochran Q for nominal measurements and Friedman for ordinal measurements) used when comparing measurements from more than two groups from the same sample or more than two measures from the same subject or respondent.



Kurtosis	A statistic that measures a distribution's peakedness or flatness (ku); a neutral distribution has a ku of 0, a flat distribution is negative and a peaked distribution is positive.



Laboratory conditions	Studies that occur under conditions that do not simulate actual environmental conditions.



lambda	A measure of how well the frequencies of one nominal variable offer predictive evidence about the frequencies of another variable; values (vary between zero to 1.0) show the direction of the association.



Leading question	A measurement question that assumes and suggests to the respondent the desired answer (nominal, ordinal, interval or ratio data).



Leniency (error)	An error that results when the respondent is consistently an easy or reluctant rater, for example the rater is very optimistic in his judgment, or very pessimistic.



Letter of transmittal	An element of the final report, this letter refers to the authorization for the project and any specific instructions or limitations placed on the study and states the purpose and scope of the study; not necessary for internal projects.



Level of significance	The probability of rejecting a true null hypothesis. See also significance.



Likert scale	A variation of the summated rating scale, this scale asks a rater to agree or disagree with statements that express either favourable or unfavourable attitudes toward the object. The strength of attitude is reflected in the assigned score and individual scores may be totalled for an overall attitude measure.



Line graphs	A statistical presentation technique used for time series and frequency distributions over time.



Linearity	An assumption of correlation analysis, that a straight line passing through the data cloud can describe the collection of data.



Linguistic behaviour	The observation of human verbal behaviour during conversation, presentation, or interaction; may also include content analysis.



LISREL	A technique (linear structural relationships) useful in explaining causality among constructs that cannot be directly measured, by analysing covariance structures.



Literature review	Summarizes and interprets recent or historically relevant research studies, company data or industry reports that act as the basis for the proposed study.



Literature search	The process of collecting information (facts, articles, books, etc.) relevant to the research problem. Literature search is an important phase at the start of your research to get acquainted with the field.



Loadings	In principal components analysis, the correlation coefficients between the factor and the variables.



Longitudinal study	The study is repeated over an extended period of time, tracking changes in variables over time; includes panels or cohort groups.



Mail survey	A relatively low cost self-administered [self-administered questionnaire] study, where the respondent receives a questionnaire by mail and also returns it by mail. Email is also increasingly used to distribute mail surveys.



Main effects	The average direct influence that a particular treatment has on the DV independent of other factors.



Management dilemma	A symptom of a management problem or an early indication of a management opportunity; a problem or opportunity that requires a management decision.



Management question	The management dilemma restated in question format; categorized as "choice of objectives", "generation and evaluation of solutions", or "trouble-shooting or control of a situation". (See also management research question hierarchy)



Management report	A report written for the non-technically oriented manager or client.



Management research question hierarchy	Process of sequential question formulation that leads a researcher from management dilemma to investigative questions.



Manager-researcher relationship	Describes the responsibilities of and conflicts between the manager contracting for the research and the firm providing or conducting the research process.



Mapping rule	Developing and applying a set of rules for assigning numbers to empirical events.



Marginal(s)	A term for the column and row totals in a cross tabulation.



Market-basket analysis	The most common form of association, which studies patterns of products purchased together; used to change store layout, adjust inventories or target promotional campaigns.



Matching	A process analogous to quota sampling for assigning subjects to experimental and control groups by having subjects match every descriptive characteristic used in the research; used when random assignment is not possible; an attempt to eliminate the effect of confounding variables that groups subjects so the confounding variable is present proportionally in each group.



Mean	The arithmetic average.



Measurement	Assigning numbers to empirical events in compliance with a mapping rule.



Measurement question	The questions asked to the respondents or the observations that must be recorded. (See also management research question hierarchy)



Measures of location	Another term for measure of central tendency in a dispersion of data (mean, mode, median).



Measures of shape	Statistics that describe departures from the symmetry of a distribution; a.k.a. moments, skewness, and kurtosis.



Measures of spread	Statistics that describe how scores cluster or scatter in a distribution; a.k.a. dispersion or variability (variance, standard deviation, range, interquartile range, and quartile deviation).



Median	The midpoint of a distribution of data, where half the cases fall above and half the cases fall below.



Meta analysis	A technique to quantitatively analyse and summarise different empirical studies on the same research problem.



Method of least squares	A procedure for finding a regression line that keeps errors (deviations from actual value to the line value) to a minimum.



Metric/non-metric measures	Refers to statistical techniques using interval and ratio measures (metric) and ordinal and nominal measures (non-metric).



Missing data	Information that is missing about a respondent or data record; should be discovered and rectified during data preparation phase of analysis; e.g. miscoded data, out-of-range data, or extreme values.



Mode	The most frequently occurring value in an array of data; data may have more than one mode.



Model	A representation of a system that is constructed to study some aspect of that system or the system as a whole.



Moderating variable	A second independent variable, believed to have a significant contributory or contingent effect on the originally stated relationship between independent and dependent variable. The moderating variable is also called interacting variable. Notation: MV.



Monitoring	See observation study.



Multicollinearity	The situation where some or all of the independent variables are highly correlated; a.k.a. collinearity.



Multidimensional scaling (MSD)	A scaling technique for objects or people where the instrument scale seeks to measure more than one attribute of the respondents or object; results are usually mapped; develops a geometric picture or map of the locations of some objects relative to others on various dimensions or properties; especially useful for difficult-to-measure constructs.



Multiphase sampling	See double sampling.



Multiple case study	A case study research that relies on more than one case, i.e. the phenomenon in question is investigated in different settings; each setting is a single case.



Multiple choice question	A question that asks the respondent to select from a pre-defined one answer alternative (see multiple choice – single response scale) or all appropriate answer alternatives (see multiple choice – multiple response scale).



Multiple choice-multiple response scale	A scale that offers respondent multiple options and solicits one or more answers (nominal or ordinal data); a.k.a. checklist.



Multiple choice-single response scale	A measurement question that poses more than two responses but seeks a single answer, or one that seeks a single rating from a gradation of preference, interest or agreement (nominal or ordinal data); aka multiple choice question.



Multiple comparison (post hoc) procedures	Tests of significance on comparison measures done after the results are compared; tests use group means and incorporate the MSerror term of the F ratio.



Multiple question	A question that requests so much content that it would be better if separate questions were asked. See also double-barrelled question.



Multiple rating list	A numerical scale where raters circle their responses and the layout allows visualization of the results (generates interval data).



Multiple regression	A descriptive tool used to (1) develop a self-weighting estimating equation by which to predict values for a dependent variable from the values of independent variables, (2) control confounding variables to better evaluate the contribution of other variables, or (3) test and explain a causal theory.



Multivariate analysis	Those statistical techniques that focus upon and bring out in bold relief the structure of simultaneous relationships among three or more phenomena.



Multivariate analysis of variance (MANOVA)	A technique that assesses the relationship between two or more dependent variables and classificatory variables or factors; frequently used to test differences among related samples.



Negative leniency (error)	An error that results when the respondent is consistently a hard or critical rater.



Neural networks	Collections of sample processing nodes plus their connections, resulting in a nonlinear predictive model that resembles biological neural networks.



Nominal data	Data without the properties of order, distance, or origin but capable of being partitioned into mutually exclusive and collectively exhaustive categories.



Non-contact rate	Ratio of potential Non-contacts (no answer, busy, answering machine, and disconnects) to all potential contacts.



Nondisclosure	Various types of confidentiality involving research projects, including sponsor, findings and purpose nondisclosures.



Non-parametric tests	One of the two general classes of significance tests, these tests use data derived from nominal and ordinal measurements and must meet three other assumptions: independence of observations, normally distributed populations, and equal variances.



Non-probability sampling	A non-random and subjective procedure where each population element does not have a known non-zero chance of being included, as the probability of selecting population elements is unknown.



Non-resistant statistics	A statistical measure that is susceptible to the effects of extreme values and does not represent typical values well under condition of asymmetry; e.g. mean, standard deviation.



Non-response error	Error that develops when an interviewer cannot locate the person with whom the study requires communication or when the targeted respondent refuses to participate; especially troublesome in studies using probability sampling.



Non-verbal behaviour	Observation of human behaviour without the use of conversation between observers and subjects (e.g. body movement, facial expressions, exchanged glances, eye blinks).



Normal probability plot	A diagnostics tool that compares the observed values with those expected from a normal distribution.



Null hypothesis	That no difference exists between the sample parameter and the population statistic to which it is compared; notation: HO.



Numerical scales	A scale where equal intervals separate the numeric scale points, while verbal anchors serve as labels for the extreme points.



Objects	Concepts of ordinary experience, like people, books, autos, genes, or peer-group pressures.



Observation	The full range of monitoring behavioural and non-behavioural activities and conditions (including record analysis, physical condition analysis, physical process analysis, nonverbal analysis, linguistic analysis, extra linguistic analysis, and spatial analysis).



Observational checklist	A measurement instrument where observed data are recorded; analogous to a questionnaire in a communication study.



Observational study	A monitoring approach to collecting data where the researcher inspects the activities of a subject or the nature of some material without attempting to elicit responses from anyone.



Observer drift	A source of error in an observation study caused by decay in reliability or validity of recorded observations over time that affects the coding of categories.



One-sample tests	Tests that involve measures taken from a single sample.



One-tailed test	A directional test of a null hypothesis that considers only one possibility: that the sample parameter is not the same as the population statistic.



Open-ended question	A type of measurement question in which the respondent provides the answer without the aid of an interviewer (either in phone, personal interview, or self-administered [self-administered questionnaire] surveys); aka unstructured or free response question (nominal, ordinal, or ratio data).



Operational definition	A definition for a variable stated in terms of specific testing criteria or operations, specifying what must be counted, measured, or gathered through our senses.



Operationalised	The process of transforming concepts and constructs into measurable variables suitable for testing.



Optical character recognition (OCR)	Software programs that transfer printed text into a computer file in order to edit and use the information without re-keying the data.



Optical mark recognition (OMR)	Software that uses a spreadsheet style interface to read and process data from user-created forms.



Optical scanning	A data-entry process whereby respondent answers are recorded on computer-readable forms then scanned to form a data set; reduces data handling and the errors that accompany such data handling.



Ordinal data	Data with order, but no distance or unique origin; data capable of determining greater than, equal to, or less than status of a property or an object.



Ordinal measures	Measures of association between variables generating ordinal data. See also gamma, Somer's d, Spearman's rho, tau b, tau c.



Outliers	Data points that exceed +1.5 the interquartile range (IQR).



Pace	A measure of comprehensibility, the rate at which the printed page presents information to the reader; it should be slower when the material is complex, faster when the material is straightforward.



Paired-comparison scale	The respondent chooses a preferred object between several pairs of objects on some property; results in a rank ordering of objects.



Panels	A technique for longitudinal survey work using the same respondents repeatedly over time, using personal, phone, and computer interviewing as well as self-administered[self-administered questionnaire] survey techniques; the use of mail-delivered diaries is common.



Parametric tests	One of the two general classes of significance tests, these tests use data derived from interval and ratio measurements.



Pareto diagram	A statistical presentation technique that includes a bar chart of frequency statistics in bar chart form, ordered from most to least, plus the cumulative percentage at each variable level indicted as a line chart.



Participant initiated error	They occur if participants do not fully answer the questions or even lie either on purpose or because they do not have the knowledge.



Participant observation	When the observer acts as both observer and participant with the subjects; the observer can be known or concealed.



Path analysis	The use of regression to describe an entire structure of linkages that have been advanced by a causal theory.



Path diagram	A diagram that presents predictive and associative relationships among constructs and indicators in a structural model.



Pattern recognition	A technique used in data mining to structure vast amount of information.



Pearson correlation coefficient	The r symbolizes the estimate of linear association based on sampling data and varies over a range of +1 to –1; the prefix (+, –) indicates the direction of the relationship (positive or inverse), while the number represents the strength of relationship (closer to 1, the stronger the relationship; 0 = no relationship); and the p represents the population correlation.



Peer-reviewed	This term refers to the process how articles are selected for a journal. Peer-reviewed means that peers (other notable scientists) evaluate the quality of the article, suggest improvements and give their opinion on whether the article should be published in the journal. The process of peer reviewing ensures the quality of research published.



Personal interview	A face-to-face, two-way communication initiated by an interviewer to obtain information from a respondent.



Phi (used with chi-square)	A measure of association for nominal, nonparametric variables; ranges from zero to +1.0 and is used best with 2 × 2 Chi-square tables; does not provide direction of the association or reflect causation.



Physical condition analysis	The collection of data from the observation of current conditions, including inventory, signs, obstacles or hazards, cleanliness, etc.



Physical trace	A type of observation that collects measures of wear data (erosion) and accretion data (measures of deposit) rather than direct observation (e.g. a study of trash).



Pictographs (geo-graphics)	A statistical presentation technique that uses pictorial symbols to represent frequency data rather than a bar in a bar-type chart; the symbol has an association with the subject of the statistical presentation and one symbol represents a specific count of that variable.



Pie charts	A statistical presentation technique that uses sections of a circle (slices of a pie) to represent 100% of a frequency distribution of the subject being graphed; not appropriate for changes over time.



Pilot testing	A trail collection of data to detect weaknesses in design and instrumentation and provide proxy data for selection of a probability sample. (see also pre-testing)



Plagiarism	Means you use (parts of the) work of others and claim that it is your own work. It is a serious violation of copy-rights.



Population	The total collection of elements (people, firms, decisions etc.) about which we wish to make some inferences.



Population element	The individual subject on which the measurement is taken; a.k.a. the population unit, case, subject or record.



Population parameters	Summary descriptors of variables (e.g. incidence, mean, variance) of interest in the population.



Population proportion of incidence	The number of elements in the population belonging to the category of interest, divided by the total number of elements in the population.



Portal	Web-sites providing a gateway to a wide array of information or offering you access to further information, such as search engines, directories etc.



Positivism	A research philosophy that builds on the principles that the social world can be viewed objectively, research is value free and the researcher is independent. It assumes that the social world can be observed by collecting objective facts and consists of simple elements to which it can be reduced. See also interpretivism.



Post hoc fallacy	Describes unwarranted conclusions, as causation is difficult to establish with an ex post facto research that does not allow manipulating the independent variable and isolating multiple causes.



Power of the test	1 minus the probability of committing a Type II error, or one minus the probability that we will correctly reject the false null hypothesis.



Practical significance	When a statistically significant difference has real importance to the decision maker.



Practicality	A characteristic of sound measurement concerned with a wide range of factors of economy, convenience, and interpretability.



Pre-coding	Assigning codebook codes to variables in a study and recording them on the questionnaire, thus eliminating a separate coding sheet.



Prediction and confidence bands	Bow-tie shaped confidence interval around a predictor; predictors farther from the mean have larger band widths.



Predictive studies	Builds on theory and attempts to provide answers to the question what (is likely to) happen in the future. It is distinct from pure speculation as the prediction on 'proofed' theoretical explanations, often derived from basic research.



Pre-testing	An established practice for discovering errors in questions, question sequencing, instructions, skip directions, etc. See also pilot test.



Primary data	Original research where the data being collected are designed specifically to answer the research question.



Primary source	These are full text publications of theoretical or empirical studies. Original works of research or raw data without interpretation or pronouncements. Do not confuse it with primary data. (see also secondary sources)



Principal components analysis	The most frequently used method of factor analysis, which transforms a set of variables into a new set of composite variables that are linear and not correlated with each other; see also factor analysis.



Probability sampling	A controlled, randomised procedure that assures that each population element is given a known non-zero chance of selection.



Probing	Techniques for stimulating respondents to answer more fully and relevantly to posed questions.



Probing questions	Similar to follow-up questions, but they address a part of the answer more specifically by asking for additional broader information.



Project management	A master plan revealing how all phases of the research will be brought together and how the research is organized.



Projective techniques	Various techniques (e.g. sentence completion tests, cartoon or balloon tests, word association tests, etc.) used as part of an interview to disguise the study objective and allow the respondent to transfer or project attitudes and behaviour on sensitive subjects to third parties; the data collected via these techniques are often difficult to interpret (nominal, ordinal, or ratio data).



Properties	Characteristics of objects; a person's properties are his weight, height, posture, hair colour, etc.



Proportional reduction in error (PRE)	A type of statistical analysis used with contingency tables (aka cross-tabulations), including lambda, gamma, tau, Somer's d, and Spearman's rho.



Proposal	A work plan, prospectus, outline, statement of intent, or draft plan for a research project, that incorporates all decisions made in the early planning phases of the study including proposed budget. A proposal is usually in a written form, if the researcher needs a sponsor for the research, such as a funding institution or a supervisor for a thesis.



Proposition	A statement about concepts that may be judged as true or false if it refers to observable phenomena. Researchers often call a statement derived purely from reasoning a proposition and distinguish it from a hypothesis, which is a statement asking for empirical testing.



Proximity	An index of perceived similarity or dissimilarity between objects.



Pure research	See basic research.



Purpose	What the author (or in the case of many internet sites, the collective authors in an institution) is trying to accomplish with the secondary source; one of five factors in secondary source evaluation.



Purpose nondisclosure	A type of confidentiality; when the sponsor camouflages the true research objective of the research project, often to mitigate biased answer behaviour of the respondent (see also nondisclosure).



Qualitative techniques	A fundamental approach of exploration and analysis, including in-depth interviews, participant observation, videotaping of subjects, projective techniques and psychological testing, case studies, street ethnography, elite interviewing and document and content analysis.



Quartile deviation	In a normal distribution, the median plus one quartile deviation on either side encompasses 50 percent of the observations, eight covers the full range of data; symbol = Q ; always used with the median for ordinal data.



Quasi-experiment	A research strategy, in which the researcher frames a real situation as an experiment without having the possibility of random assignment and manipulation.



Quota matrix	A means of visualizing and organising the matching process.



Quota sampling	A type of purposive sampling in which relevant characteristics are used to stratify the sample in an attempt to improve the representativeness of the sample. Another motive for using quota sampling is to ensure that elements with a certain rarely occurring characteristic, which is important to the research, are included in the sample.



Random assignment	A process that uses a randomised sample frame for assigning subjects to experimental [experimental group] and control groups in an attempt to assure that the groups are as comparable as possible with respect to the dependent variables; each subject must have an equal chance for exposure to each level of the independent variable. Aka randomisation.



Random dialling procedures	A procedure for bypassing out-of-date phone directories that requires choosing phone exchanges or exchange blocks and then generating random numbers within these blocks for calling. In certain countries law restricts the use of random dialling.



Random error	Error that occurs erratically, without pattern; see also sampling error.



Randomisation	See random assignment.



Range	The difference between the largest and smallest score in the distribution; a very rough measure of spread of a dispersion.



Ranking question	A measurement question that asks the respondent to compare and order two or more objects or properties using a numeric scale. See also ranking scale.



Ranking scale	A measurement approach that asks the respondent to make comparisons among two or more objects or properties in relation to each other using a numeric scale, thus providing a relative order of those factors (ordinal or interval data). See also ranking question.



Rating question	A measurement question that asks the respondent to position each property or object on a companion verbal, numeric, or graphic scale. See also rating scale.



Rating scale	A measurement approach that asks the respondent to score an object or property without making a direct comparison to another object and thus position each factor on a companion scale, either verbal, numeric, or graphic (ordinal or interval data). See also rating question.



Ratio data	Data with order, distance, equal intervals (distance), and unique origin; numbers used as measurements have numerical value; e.g. weight of an object.



Reactivity response	The phenomenon where subjects alter their behaviour due to the presence of the observer.



Readability indexes	Indexes that measure the difficulty level of written material; e.g. Flesch Reading Ease Score, Flesch Kincaid Grade Level, Gunning's Fog Index; most word-processing programs calculate one or several of the indexes.



Realism	A research philosophy, which shares principles of positivism and interpretivism. It accepts the existence of a reality independent of human beliefs, but still concedes that understanding requires acknowledgement of human subjectivity.



Reciprocal relationship	Two variables mutually influence or reinforce each other.



Record	A set of data fields that are related, usually buy subject or respondent; represented by rows in a spreadsheet or statistical data base; a.k.a. data case, data record.



Record analysis	The extraction of data from current or historical records, either private or in the public domain; a technique of data mining.



Refusal rate	Ratio of respondents who decline the interview to all potential/eligible contacts.



Regions of acceptance	Area between the two regions of rejection based on a chosen level of significance (two-tailed test) or the area above/below the region of acceptance on a one tailed test.



Regions of rejection	Area beyond the region of acceptance set by the level of significance.



Regression analysis	Uses simple and multiple predictions to predict Y from X values.



Regression coefficients	The two association measures between X and Y variables, intercept and slope.



Relational hypothesis	Describes the relationship between two variables with respect to some case; relationships are co-relational or explanatory.



Reliability	A characteristic of measurement concerned with accuracy, precision, and consistency; a necessary but not sufficient condition for validity (if the measure is not reliable, it cannot be valid, but reliable measures are not necessarily valid).



Reliability-equivalence	A characteristic of measurement in which an instrument can secure consistent results with repeated measures by the same investigator or by different samples.



Reliability-stability	A characteristic of measurement in which an instrument can secure consistent results with repeated measurements of the same person.



Replication	The process of repeating an experiment with different subject groups and conditions to determine the average effect of the IV across people, situations, and times.



Replication logic	The rationale of case study research. If a case study offers support for a specific theoretical proposition, replicating the research in slightly different setting through a new case study reinforces the theoretical proposition.



Reporting study	Provides an account or summation of some data, perhaps the generation of some statistics, but requires little inference or conclusion drawing.



Request for proposal (RFP)	Bid request for research to be done by an outside supplier of research services.



Research design	The blueprint for fulfilling research objectives and answering the research questions.



Research dilemma	A symptom of a problem or an early indication of an opportunity. A problem or opportunity that requires a decision based on a systematic inquiry. While a management dilemma refers to a problem or opportunity managers face, research dilemma can include issues beyond the interest of managers, e.g. societal issues or academic puzzles.



Research process	Various decision stages involved in a research project, and the relationship between those stages.



Research question	The choice hypothesis that best states the objective of the research; the answer to this question provides the desired information necessary to make a decision with respect to the research dilemma.



Research variable	See variable.



Residual	What remains after the regression line is fit (the difference between the regression line value of Y and the real Y value).



Resistant statistics	A statistical measure that is relatively unaffected by outliers within a data set; e.g. median and quartiles.



Respondent	Another term for a participant in a communication study.



Response error	Error created when the data reported differs from the actual data.



Right to privacy	The respondent's right to refuse to be interviewed or to refuse to answer any questions in an interview. Furthermore, it is closely related to confidentiality. A researcher, who has obtained information from others, is not allowed to use this information in a way that might harm the provider of the information.



Right to quality	The client's right to a research design appropriate for the research question, maximum value for the resources expended, and data handling and reporting techniques appropriate for the data collected.



Right to safety	The right of interviewers, surveyors, experimenters, observers, and subjects to be protected from any threat of physical or psychological harm.



Rotation	In principal components analysis, a technique used to secure less ambiguous relationships between factors and variables by performing a matrix analysis and thus aid in interpretation of the analysis.



Sample	A group of cases, respondents, or records comprised of part of the target population, carefully selected to represent that population (see also pilot testing, data mining).



Sample statistics	Descriptors of the relevant variables computed from sample data.



Sampling	The process of selecting some elements from a population to represent the population as a whole.



Sampling error	A reflection of the influences of chance in drawing the sample from the population; the error not accounted for by systematic variance.



Sampling frame	A list of elements in the population from which the sample is actually drawn, such as a directory.



Scaling	The assignment of numbers or symbols to an indicant of a property of objects to impart some of the characteristics of the numbers to the property.



Scalogram	A procedure for determining whether a set of items forms a unidimensional scale and is therefore appropriate for scaling.



Scatter plots	A visual technique for depicting both the direction and the shape of a relationship between variables.



Scientific method	Disciplined procedures for generating quality research including direct observation of phenomena; clearly defined variables, methods, and procedures; empirically testable hypotheses; the ability to rule out rival hypotheses; and statistical rather than linguistic justification of conclusions.



Scope	The degree of comprehensiveness of coverage of a secondary source (by time frame, topics, geography, etc.); one of the five factors for evaluating the quality of secondary sources.



Screen question	The question(s) asked during a phone interview, to determine whether the person answering the phone is a qualified sample unit (nominal data). See also filter question.



Search query	A statement combining different terms with logical operators, such as "and", "or", "not", "smaller than" etc. to search a database. It is advisable to generate a list of alternative key terms and combine those to various search queries.



Secondary data	Studies done by others and for different purposes than for which the data are being reviewed and reused.



Secondary source	Compilations of information in printed or electronic form and subsequent forms of publications of primary sources. Including interpretations of primary sources. Do not confuse it with secondary data.



Self-administered questionnaire	A survey delivered to the respondent via personal (intercept) or non-personal (computer-delivered, mail delivered) means that is completed by the respondent without additional contact with an interviewer. See also mail survey.



Semantic differential scale	A scale that measures the psychological meanings of an attitude object and produces interval data; uses bipolar nouns, noun phrases, adjectives, or nonverbal stimuli such as visual sketches.



Semi-structured interviews	In a semi-structured interview, the researcher formulates the questions before the interview, but usually does not offer the respondent a choice among answer alternatives, i.e. the questions are open. Moreover, it is possible to deepen an answer to a specific question by asking additional questions.



Sentence outline	One of two types of outlines normally used in the pre-writing phase of report development uses complete sentences rather than key words or phrases to draft each report section.



Sequence-based analysis	A variant of traditional market-basket analysis, used to tie together a series of activities or purchases, taking into account not only the association of items but their order.



Sequential sampling	See double sampling.



Simple category scale	A scale with two response choices; aka dichotomous scale.



Simple cluster sampling	A cluster sampling procedure with only one phase, thus we sample clusters and take all elements within the cluster. If all clusters have the same size, this is analogous to simple random sampling.



Simple observation	Another term for data collection during the exploratory phase of a study. See also observation.



Simple random sample	A probability sample in which each element has a known and equal chance of selection.



Simulations	A study in which the conditions of a system or process are replicated.



Single case study	A research using one case study, e.g., one particular firm or event to investigate a phenomenon.



Skewness	A measure of a distribution's deviation from symmetry; if fully symmetrical the mean, median and mode are in the same location.



Skip pattern	Instructions designed to route or sequence the respondent to another question based on the answer to a branched question.



Slope	One of two regression coefficients, β₁, is the change in Y for a one-unit change in X.



Snowball sampling	A non-probability sampling procedure in which initial sample elements, which may or may not have been chosen by probability techniques, refer to additional sample elements based on similar characteristics. For example you start interviewing people interested in starting a business selected from an address list of a seminar on entrepreneurship and ask them if they know other people who intend to start a business.



Solicited proposal	Proposal developed in response to a RFP.



Somer's d	A measure of association for ordinal data that compensates for "tied" ranks and adjusts for direction of the independent variable. See also ordinal measures.



Source evaluation	The five-factor process for evaluating the quality and value secondary sources and also secondary data (see purpose, scope, authority, audience, and format).



Sources of evidence	Refers to the sources where you get the information from to describe phenomena or measure variables. They include formal and informal interviews, documents, databases etc.



Spatial relationships	The recording of human behaviour and how humans physically relate to each other, including proxemics (the study of how people organize the territory around them and the discrete distances they maintain between themselves and others).



Spearman's rho	One of the most popular ordinal measures of association; correlates ranks between two ordered variables.



Specification error	A bias that overestimates the importance of the variables included in a structural model.



Specifying questions	Questions that ask the respondent to explain an answer more deeply.



Sponsor nondisclosure	A type of confidentiality; when the sponsor of the research disassociates itself from the sponsorship of the research project (see also nondisclosure).



Spreadsheet	A data-entry software that arranges data cases or records as rows, with a separate column for each variable in the study.



Stability	A characteristic of measurement in which an instrument can secure consistent results with repeated measurements of the same person. See also reliability.



Standard deviation	The positive square root of the variance, it is the most frequently used measure of the spread or variability of a data dispersion; symbol = s, or σ, or std. dev.; affected by extreme scores.



Standard error of the mean	A measure of the standard deviation of the distribution of sample means.



Standard normal distribution	The most significant theoretical distribution in statistics which is the standard comparison for describing distributions of sample data and is used with inferential statistics that assume normally distributed variables.



Standard scores (Z scores)	A calculation that conveys distance in standard deviation units; a mean of 0 and a standard deviation of 1; designed to improve compatibility among variables that come from different scales yet require comparison; includes both linear manipulations and non linear transformations.



Standardized coefficients	The X values re-stated in terms of their standard deviations (a measure of the amount that Y varies with each unit change of the associated X variable).



Stapel scale	A numerical scale with up to 10 categories (5 positive, 5 negative) where the central position is an attribute. The higher the positive number, the more accurately the attribute describes the object or its indicant.



Statistical process control	A statistical tool to analyse, monitor, and improve process (a business system that transforms inputs to outputs) performance.



Statistical significance	The quality of the difference between a sample value and its population value; the difference is statistical significance if it is unlikely to have occurred by chance (represent random sampling fluctuations).



Statistical study	A study that attempts to capture a population's characteristics by making inferences from a sample's characteristics; involves hypothesis testing and is more comprehensive than a case study.



Stem-and-leaf display	An EDA technique closely related to a histogram that reveals frequency distribution for each data value, without equal interval grouping.



Stepwise selection	In modelling and regression, the most popular method for sequentially adding or removing variables; combines forward and backward sequential methods.



Stratified random sampling	A probability sampling technique where the sample is constrained to include elements from each of the mutually exclusive segments or strata within a population.



Stratified random sampling-disproportionate	A probability sampling technique in which each stratum's size is not proportionate to the stratum's share of the population; allocation is usually based on variability of measures expected from the stratum, cost of sampling from a given stratum, and size of the various strata.



Stratified random sampling-proportionate	A probability sampling technique in which each stratum is properly represented so the sample drawn from it is proportionate to the stratum's share of the population; higher statistical efficiency than a simple random sample.



Stress index	An index used in multidimensional scaling that ranges from 1 (worst fit) to 0 (perfect fit).



Structured interviews	In a structured interview the questions and the possible answers are defined ex-ante. Self-administered questionnaires are a typical example of structured interviews.



Structured question	A type of measurement question that presents the respondent with a fixed set of choices (nominal, ordinal or interval data). See also structured response, closed question.



Structured Response	A response that is based on a fix pre-defined set of alternative answer options. See also structured question.



Structuring questions	Questions that are asked to allow a smooth transition to a new topic.



Subjects' perceptions	The subtle or major changes that occur in subjects responses when they perceive that a research study is being conducted.



Successive intervals	Infrequently used process for ordering many objects where the respondent groups objects on properties by allocating cards to piles or groups representing a succession of values or importance of properties.



Survey	A means of questioning a respondent via a collection of questions and instructions for both the respondent and the interviewer; aka questionnaire or instrument.



Systematic error	Error that results from a bias; see also systematic variance.



Systematic observation	Data collection through observation that employs standardized procedures, trained observers, schedules for recording and other devices for the observer that mirror the scientific procedures of other primary data methods.



Systematic sampling	A complex probability sampling technique in which the population (N) is divided by the desired sample (n) to obtain a skip pattern (k). Using a random start between 1 – k, each k^thelement is chosen from the sample frame; usually treated as a simple random sample but statistically more efficient.



Systematic variance	The variation in measures due to some known or unknown influences that cause the scores or measurements to skew in one direction or another; see also systematic error.



T distribution	A normal distribution with more tail area than in a Z normal distribution.



Target question	A type of measurement question that addresses the investigative questions (core information questions) of a specific study.



tau	Uses table marginals to reduce prediction errors, with measures from 0 to 1.0 reflecting percentage of error estimates for prediction of one variable based on another variable.



tau b (used with Spearman's rho)	A refinement of gamma for ordinal data that considers "tied' pairs, rather than only discordant and concordant pairings (values from -1.0 to +1.0); used best on square tables (one of the most widely used measures for ordinal data).



tau c (used with Spearman's rho)	A refinement of gamma for ordinal data that considers "tied' pairs, rather than only discordant and concordant pairings (values from -1.0 to +1.0); useful for any sized table (one of the most widely used measures for ordinal data).



Technical report	A report written for an audience of researchers.



Telephone interview	A study conducted wholly by telephone contact between respondent and interviewer.



Test unit	An alternative term for a subject within an experiment; it can be a person, an animal, a machine, a geographic entity, etc.



Theory	A set of systematically interrelated concepts, definitions and propositions that are advanced to explain or predict phenomena (facts); the generalizations we make about variables and the relationships among variables.



Time sampling	The process of selecting some time points or intervals to observe elements, acts, or conditions from a population of observable behaviour or conditions to represent the population as a whole; three types include time-point samples, time-interval samples or continuous real-time samples.



Topic outline	One of two types of outlines normally used in the pre-writing phase of report development uses key words or phrases rather than complete sentences to draft each report section.



Transformation	The re-expression of data on a new scale using a single mathematical function for each data point; designed to improve interpretation when researcher finds alternate ways to understand the data and discover patterns or relationships not revealed by original scales (for interval data; possible for nominal and ordinal data).



Treatment (factor)	The experimental factor to which subjects are exposed.



Treatment level	The arbitrary or natural groupings within the independent variable of an experiment.



Trials (repeated measures)	Repeated measures taken from the same subject.



Triangulation	A process of verifying information through multiple sources to increase the validity of the description of what is observed. Triangulation is especially important in case study research.



Triangulation	A process of verifying information through multiple sources to increase the validity of the description of what is observed. Triangulation is especially important in case study research.



t-test	A parametric test to determine the statistical significance between a sample distribution mean and a population parameter, when the population standard deviation is unknown and the sample standard deviation is used as a proxy.



Two independent samples tests	Parametric and non-parametric tests used when the measurements are taken from two samples that are unrelated (Z test, t-test, chi-square, etc.).



Two related samples tests	Parametric and nonparametric tests used when the measurements are taken from closely matched samples or the phenomena are measured twice from the same sample; t-test, McNemar test, etc).



Two-tailed test	A non-directional test of a null hypothesis that considers two possibilities: that the sample parameter is either greater than the population statistic or less than the population statistic.



Type I error	A type of hypothesis testing error when a true null hypothesis (there is no difference) is rejected; the alpha (α) value called the level of significance is the probability of rejecting the true null hypothesis.



Type II error	A type of hypothesis testing error when a false null hypothesis (there is no difference) is rejected; the beta (β) value is the probability of incorrectly rejecting the false null hypothesis; the power of the test = 1 – β, and is the probability that we will correctly reject the false null hypothesis.



Unidimensional scale	Instrument scale that seeks to measure only one attribute of the respondents or object.



Unit of analysis	It refers to the level at which the phenomenon occurs and at which the research needs to be conducted. For example if you want to investigate internationalisation strategies, you might select firms as unit of analysis, but depending on the explanations you are looking for it might be better to investigate a firm's decision to enter a foreign market.



Unobtrusive measures	A set of observational approaches which encourage creative and imaginative forms of indirect observation, archival searches, and variations on simple and contrived observation, including physical traces observation (erosion and accretion).



Unsolicited proposal	A suggestion by a contract researcher for research that might be done.



Unstructured interviews	In an unstructured interview, the researcher rather defines topic areas that will addressed in the interview and is flexible to change the course of the interview depending on the answers given.



Unstructured question	A type of measurement question in which the respondent provides the answer to without the aid of an interviewer (either in phone, personal interview or self-administered surveys); see also open-ended question, unstructured response or free response question.



Unstructured response	A response strategy where participant's opinions are limited only by space, layout, instructions, or time; usually free-response or 'fill-in' response strategies. See also unstructured question.



Utility scores (part-worths)	Used in conjoint analysis and computed from a subject's rankings or ratings of a set of cards, with each card describing one possible configuration of combined concepts and constructs.



Validity	A characteristic of measurement concerned that a test measures what the researcher actually wishes to measure; that differences found with a measurement tool reflect true differences among respondents drawn from a population. See also construct validity, content validity, criterion-related validity.



Variability	Another term for measures of spread or dispersion within a data set.



Variable	A characteristic, trait, or attribute that is measured; a synonym for a construct or the property being studied; a symbol to which values are assigned; includes several different types: continuous, control, decision, dependent, dichotomous, discrete, dummy, extraneous, independent, intervening, and moderating variables.



Variance	Calculated as the squared deviation scores from the data distribution's mean, it is a measure of score dispersion about the mean; the greater the dispersion of scores, the greater the variance in the data set.



Vector	A quantity that has direction and magnitude commonly represented by a directed line segment whose orientation in space represents the direction and length represents the magnitude.



Vignette research	See factorial survey.



Visual aids	Presentation tools used to facilitate understanding of content (e.g. chalkboards, whiteboards, handouts, flip charts, overhead transparencies, slides, computer-drawn visuals, computer animation).



Voice recognition	Computer systems programmed to record verbal answers to questions.



Web-based questionnaire	Special surveys designed to be delivered via the Internet with data capture and processing a potential part of the process. Two options currently exist: proprietary solutions offered through research firms and off-the-shelf software for researchers who possess the necessary knowledge and skills.



Z distribution	The normal distribution of measurements assumed for comparison.



Z test	The parametric test to determine the statistical significance between a sample distribution mean and a population parameter employs the Z distribution.



Transcribing	The process to transform recordings of interviews into a written account. Very labour intensive.



Emerging coding	A coding technique to develop codes and categories while reading through the data, i.e. the data themselves are an inspiration source to develop codes



Coding (process)	Assigning numbers or other symbols to answers to responses that can be tallied and grouped into a limited number of classes or categories.



Taxonomy	Describes a system in which similar things are grouped together under one new broader term.



Typology	Describes a structure that organizes things (concepts or codes) into distinct categories.



Induction	To draw a conclusion from one or more particular facts or pieces of evidence; the conclusion explains the facts.



Deduction	A form of inference in which the conclusion must necessarily follow from the reasons given. If the reasons (premises) are true and the conclusion follows necessarily from the reasons, the deduction is valid (see also Induction).



Domain analysis	A higher-ordered category that includes other categories and formalizes relationships between categories.



Hermeneutical analysis	The theory of interpretations that addresses in a systematic way challenges faced in interpretation.



Heuristic research	A qualitative analysis approach that starts with a personal question that is answered in a process of self-enquiry and dialogue.



Phenomenology	Qualitative analysis technique where the understanding of reported experiences is central.



Semiotic analysis	A qualitative analysis approach that looks for patterns and structures between visual, linguistic and aural signs.



Discourse analysis	A qualitative analysis approach to look at the flow of communication and especially its relational aspects.



CAQDAS	Stands for computer assisted qualitative data analysis software.

Log In
You must be a registered user to view the premium content in this website. If you already have a username and password, enter it below. If your textbook came with a card and this is your first visit to this site, you can use your registration code to register.
Username:
Password:
	Forgot your password?

Site Preferences	(Log out)

		Send mail as:

TA email:
Other email:

"Floating" navigation?
Drawer speed:

Notes		(What is this?)

Add a note
1.	(optional) Enter a note here:
2.	(optional) Select some text on the page (or do this before you open the "Notes" drawer).
3.	Highlighter Color:
4.

Search
Search for:
Search in:

Home > Glossary

	Course-wide Content
		Statistics Chapters SPSS Guide Research Skills Centre Dataset Exercises Annotated References Appendices Cases Web Links PowerPoint Tutorial Glossary