Glossary – Management Research Methodology: Integration of Principles, Methods and Techniques


Accessible population—A subset of a target population that is reachable by the researcher and from which the sample is drawn.

Adaptation—The process by which an organisation, system, or organism is fitted to its environment.

Ad hoc sample—Sample of participants drawn from an accessible population. Characteristics of the ad hoc sample must be described to define the limits of generalisability.

Alpha level (α level)—Level of type I error (the probability of rejecting the null hypothesis when the null hypothesis is true).

Alternate courses of action—Different ways of finding solutions to a problem in the process of decision-making.

Alternative hypothesis—Opposite of null hypothesis (notation: HA).

Analysis of covariance (ANCOVA)—Statistical procedure, similar to analysis of variance, used to evaluate whether two or more groups have different population means. Analysis of covariance statistically removes the effects of extraneous variables on the dependent variable and, hence, increases the power of the statistical test.

Analysis of variance (ANOVA)—Statistical procedure used to analyse mean differences between two or more groups. ANOVAs compare the variability among groups with the variability within groups. Many variations of analysis of variance are possible, including repeated measure ANOVAs and factorial ANOVAs.

Applied research—Research to provide solutions to practical problems.

Area sampling—A type of cluster sampling applied to a population having well defined boundaries but without a sample frame.

Artifact—In research, any apparent effect of a major conceptual variable that is actually the result of a confounding variable that has not been properly controlled. Artifacts threaten the validity of research conclusions.

Artificial intelligence—Machines that are designed to evaluate and respond to situations in an appropriate manner. Most artificial intelligence machines are computer based and many of them have achieved remarkable levels of performance in specific areas.

Association—Relationship or correlation.

Authority—A way of acquiring knowledge. New ideas are accepted as valid because some respected authority has declared the idea to be true.

Automation—Use of equipment to conduct most or all aspects of presenting stimuli and recording participants’ responses. Automation can minimise the work required of the researcher, increase precision in data gathering, and minimise experimenter bias in gathering and recording data.

Average deviation—The sum of the deviations from the mean divided by the number of scores.

B variance—It is the sum of the squared deviation scores from data mean and is a measure of dispersion.

Bar chart—It is a statistical presentation technique that represents frequency data as horizontal or vertical bars. Bar charts are used to represent time series and quantitative data.

Basic research—Fundamental or pure research that is carried out to add to knowledge but without applied or practical goals. Basic research is often contrasted with applied research.

Behaviour—Any observable act from any living or non-living organism. Behaviour is the subject matter of psychology.

Behaviourism—A philosophical perspective in psychology, which argues that scientific psychology should base its theories on observable events only (such as behaviour). This perspective challenged the introspective methodologies that dominated the early discipline of psychology.

Beta (β)—The probability of making a type II error (see Type II error).

Between-groups variance—Index of variability among group means.

Between-subjects factors—Independent variables in factorial designs in which participants are assigned to conditions so that each participant appears only in one condition.

Between-subjects design—Research design using two or more groups, in which each participant appears in only one of the groups.

Blind—When the researcher and/or participant is not aware of information that would, if available, increase the likelihood of biasing the experimental results (see single-blind procedure).

Box-Plot—An exploratory data analysis technique that gives a visual image of a variable’s distribution, spread, shape, and outliers.

Branching questions—A technique used to direct respondents to different places in a questionnaire, based on the response to the question at hand.

Cannonical correlation—This is correlation between two sets of variables. The first canonical correlation is derived by computing the linear combination of each set of variables that will give the highest possible correlation. Additional canonical correlations can be computed using different linear combinations of the variables in each set. This technique helps scientists to understand complex relationships between constructs that cannot be easily trapped by a single measure.

Carry-over effects—These effects occur when a participant’s involvement in one condition affects his or her performance in all subsequent conditions. Carry-over effects occur only when each participant appears in more than one experimental condition (that is, in within-subjects designs).

Categorical data—This kind of data is synonymous with nominal data.

Categorical variable—Synonymous with discrete variable, a categorical variable can have only a finite number of values.

Causal hypothesis—A form of research hypothesis in experimental research. It states that the independent variable has a causal relationship to the dependent variable. To accept this hypothesis, one must have rejected the null hypothesis and all confounding variable hypotheses.

Causal inference—Conclusion that changes in the independent variable resulted in a change in the dependent variable. It may be drawn only if all potential confounding variables are properly controlled.

Central tendency—Average or typical score in a distribution. Three measures of central tendency are the mean, medium, and mode.

Chi-square (χ2)—A statistical distribution that forms the basis for inferential statistics used with nominal data.

Classification variables—Organismic or subject variables used to classify participants into discrete groups. Classification variables are used for assigning objects to groups in differential research.

Cluster analysis—A technique that identifies homogenous subgroups or clusters of subjects or study objects.

Cluster sampling—A sampling procedure in which the population is divided into clusters or subgroups. The sample is then drawn from each subgroup.

Coding data—Process by which numerical or classification scores are assigned to a set of responses from a respondent. The coded data are usually in a form that can be statistically analysed more easily.

Coefficient alpha—An index of (internal consistency) reliability.

Coefficient of determination—The square of the Pearson product moment correlation. It represents the proportion of variability in one variable that can be predicted on the basis of information about the other variable.

Cognitive mapping—A pictorial representation of how a manager models, thinks about, and perceives his problems and situations. This is developed from group interview.

Cohort groups—People of a given chronological age in a given culture, who behave similarly throughout their lives and differently from people of other ages because of shared life experiences.

Computer modelling—Using a computer to simulate a process where the intention is to have the computer mimic a process as closely as possible to the way in which it actually occurs.

Confidence interval—An interval in which we predict the population parameter to fall with a specified level of confidence (for example, a 95 per cent confidence interval will contain the population parameter 95 per cent of the time).

Confidentiality—An ethical requirement in most research information, particularly sensitive and personal information, provided by participants as part of a research study should be protected and made unavailable to anyone other than the researchers.

Confounded—Two independent variables are said to be confounded if they vary simultaneously during a study, thus, not allowing us to determine which of these variables was responsible for the observed change in the dependent variable.

Confounding variable—Any uncontrolled variable that might affect the outcome of a study. A potential confounding variable exists only if (i) there is a mean difference between the groups on the variable, and (ii) there is a correlation between the variable and the dependent measure.

Constants—Parameters that do not vary are called constants.

Constraints—Restrictions placed on research in an effort to increase the precision of the research and enhance the validity of the conclusions (also, limitations of resource/time).

Construct—An idea constructed by the researcher to explain events observed in a particular situation. Constructs are not necessarily direct representations of reality, they are not facts. They are explanatory fiction because, in most cases, we do not know the real reason for a particular event. Once formulated, constructs are used as if they are true (that is, analogically) to predict relationships between variables in situations that had not previously been observed.

Construct validity—Validity of a theory is also known as construct validity. Most theories in science present broad conceptual explanations of relationship between variables and make many different predictions about the relationships between particular variables in certain situations. Construct validity is established by verifying the accuracy of each possible prediction that might be made from the theory. Because the number of predictions is usually infinite, construct validity can never be fully established. However, the more independent predictions for the theory verified as accurate, the stronger the construct validity of the theory.

Consultancy research—Research carried out for an organisation by specialist consultant.

Content analysis—An analysis tool used for measuring the semantic content of messages or communications like speeches, advertisements, editorials.

Content items—In questionnaires and interviews, content items focus on the issues studied, such as respondents opinions, attitudes, and knowledge, in contrast to factual items, which can be independently verified (for example, age, gender, and so on).

Contingency—A particular relationship between two or more variables, where, given that the first event occurs, the second event is highly probable. The relationship between the variables is a probabilistic one and does not necessarily imply a causal connection between them.

Continuous variable—Any variable that can theoretically take on an infinite number of values. These are often contrasted with discrete or categorical variables.

Control group—A group of participants, used in either differential or experimental research, that serves as a basis for comparison of other (experimental) groups. The ideal control group is similar to the experimental group on all variables except the variable that defines the group (independent variable).

Controlled research—This includes any research that employs adequate control procedures to rule out competing hypotheses. Well-controlled research permits scientists to draw causal conclusions.

Convenience sampling—A non-probability sampling procedure in which the sample is unrestricted but accessible and available.

Correlated t-test (or direct-differences t-test or matched-pairs t-test)—statistical procedure used to test for mean differences between two groups in a within-subjects or matched-subjects design.

Correlation—The degree of association between two or more variables.

Correlation coefficient—A statistic that quantifies the degree of association between two or more variables. There are many kinds of correlation coefficients, depending on the type of data and relationship predicted.

Correlational research—Research that seeks to measure the relationship between variables, without trying to determine causality. The term is sometimes used broadly to include any non-experimental research design.

Criterion—In a regression analysis, the criterion is the variable that one attempts to predict.

Crossover effects—In quasi-experimental research, a finding where two non-equivalent groups show one pattern of scores before the manipulation and the reverse pattern of scores after the manipulation. The name derives from the crossing of lines when such a result is shown on graph.

Cross-sectional design—A design that compares the performance, attitudes, or histories of people of different ages or at different times in history. The groups are defined by the age range of the people in the groups or the historical time in which participants were tested. In a cross-sectional study, participants appear in only one group. This design is often contrasted with longitudinal designs.

Cross-sectional research—Research in which a cross-sectional research design is used.

Cross-tabulation—Procedure for organising frequency data that displays the relationship between two or more nominal variables. A cross-tabulation table contains individual cells, with the number in each cell representing the frequency of participants who show that particular combination of characteristics.

Data—Plural noun that refers to information gathered in research. Research conclusions are drawn on the basis of an evaluation of the data gathered as part of a study.

Data analysis—Research phase in which data gathered from observing participants are analysed, usually with statistical procedures.

Data mining—The process of extracting meaningful knowledge from large volumes of data contained in data warehouses.

Data warehouse—Electronic storehouses where vast amounts of data are arrayed, integrated, categorised, stored, and sold.

Decision tree—An organised pathway of ideas leading to a defined goal, in which at various points, a decision is made about which of two ‘branches’ of ideas to follow to the next decision point.

Deductive reasoning—Reasoning from the general to the particular. In deductive reasoning, specific predictions are made about future events.

Deductive theory—A theory that emphasises constructs and the relationship between constructs and seeks to make predictions, from the theory which can be tested through empirical research.

Degrees of freedom (df)—This is a statistical concept. One degree of freedom is lost each time a population parameter is estimated on the basis of a sample of data from the population. The distribution of most statistics (t, F.., and so on) are tabled by degrees of freedom (df).

Demographic variables—Data that describe the participants in a study (for example, their age, gender, occupation, and so on). This information should be routinely collected and reported in research.

Dependence analysis—Problem in multivariate analysis in which one (or more) of the variables is to be considered separately and to investigate how it depends upon others.

Dependent variable—Variable that is hypothesised to have a relationship with the independent variable.

Depth interview—A type of interview that is unstructured and made in a free environment in which the respondent shares information freely.

Descriptive statistics—Those statistics or statistical procedures that summarise and/or describe the characteristics of a sample of scores.

Discrete variable—A discrete variable can take on only a finite number of values. It is often contrasted with continuous variable.

Dispersion (variability)—This illustrates how spread out the scores are in a sample.

Double-barrelled questions—A question that calls for two responses, thereby creating confusion in the respondent.

Effectiveness function—An equation representing objectives of a model. It indicates in quantitative terms how effective an alternative course of action is in achieving the objectives.

Empirical—Based on observed data. A relationship between variables is empirically established if it has been observed to occur.

Empiricism—System of learning that is based solely on observation of the events around us.

Equivalence—A measure of reliability that is applied to both single instruments and measurement situations.

Experimentation—Process by which a researcher studies the relationship between independent and dependent variables by systematically manipulating the independent variable, assigning participants without bias to each level of the independent variable, and observing the effects of the independent variable on the dependant variable.

Experimenter bias—Any effect that the expectations of the researcher might have on the measurement and recording of the dependent variable. Uncontrolled experimenter expectancies can create powerful experimenter effects.

Experience survey—Interviews with people knowledgeable about the general subject being investigated.

Exploratory data analysis—Process in which data patterns guide the analysis or suggest revisions to the preliminary data analysis plan.

External validity—Extent to which the results of a study accurately indicate the true nature of a relationship between variables in the real world. If a study has external validity, the results are said to be generalisable to the real world.

Extraneous variable—Any variable other than the independent variable that might affect the dependent measure in a study. Extraneous variables are potentially confounding and must be controlled.

Extraneous variance—Variability in scores on the dependent measure that can be accounted for by the effects of extraneous variables.

Factor analysis—Body of techniques concerned with the study of interrelationships among a set of variables, none of which is a criterion variable.

Factorial design—Research designs employing more than one independent variable simultaneously. The major advantage of a factorial design is that it can measure the joint (interactive) effects of two or more independent variables.

Factor—In a factorial design, each of the independent variables is a factor.

Facts—Empirically observed events.

Factual items—In questionnaires and interviews, factual items are questions that can be independently verified, such as the respondent’s age, gender, and occupation. In contrast, content items cannot be factually verified.

Fields (in computer files)—In data files, the field represents a variable that has a score for each participant in the study. Normally, the fields are shown as columns in a data matrix, where the rows represent records.

Focus group—An information collection approach widely used in exploratory research. A panel of subjects are met by a moderator for obtaining feelings, perceptions, and experiences on a specific topic.

Formal science—Science that uses formulation rules (automatic in nature, having theorems, axioms) and pure deductive systems.

Grounded analysis—An analysis approach based on grounded theory, to extract a problem structure from the data in qualitative data analysis. It is considered as a creative process, having as its components reflection, conceptualisation and cataloging.

Heuristics—Rules of thumb derived by experience, intuition, and simple logic.

Hypothesis—A declarative statement or a proposition that describes the relationship of two or more variables or the characteristic of a variable (like mean or variance).

Hypothetico-deductive method—A scientific procedure in which hypothesis are accumulated on specific observations and thereafter generalised by deductively testing them on larger observations.

Ideographic knowledge—Knowledge gained by studying particular cases, social groups, or situations.

Inductive theory—Inductive theories are built on a strong empirical base and tend not to stray far from that empirical base. It is often contrasted with deductive theory and functional theory.

Inference—Any conclusion drawn on the basis of some set of information. In research, we draw inferences on the basis of empirical data we collect and ideas we construct.

Inferential statistics—Statistical procedures that computes the probability of obtaining the particular pattern of data in a study if all participants were actually drawn from the same population. If the probability of obtaining such a pattern of scores is low, we reject the hypothesis that all participants were drawn from the same population (null hypothesis) and conclude that there were meaningful differences between groups or conditions.

Informed consent—Critical principle in the ethical treatment of participants. Participants have the right to know exactly what they are getting into and to refuse to participate.

Instrumentation—Potential confounding variable involving any change in the measuring instrument over time, which causes the instrument to give different readings when no change has occurred in the participant.

Interaction—Combined effect of two or more independent variables on the dependent variable. Interactions can be measured only in factorial designs.

Interdependence analysis—Problem in multivariate analysis to determine the relationship of a set of variables among themselves (no variate is a dependent variable).

Internal consistency reliability—Index of how homogeneous the individual items of a measure are. If the individual items are homogenous, they will tend to correlate strongly with one another, suggestions that all items are measuring the same characteristic.

Internal validity—Accuracy of the research study in determining the relationship between independent and the dependent variables. Internal validity can be assured only if all potential confounding variables have been properly controlled.

Intervening variable—A variable that affects a phenomena or system but cannot be seen, measured, or manipulated. Its effects are determined by its effects on independent or moderator variables.

Interpretation—Results of statistical analyses of data are interpreted in the light of (i) the adequacy of control procedures in the research design, (ii) previous research, and (iii) existing theories about the study.

Interrater reliability—Index of consistency between two or more ratings made by separate raters. It is indexed by the correlation between the ratings of the two raters.

Interrupted time-series design—Type of research design suitable for either single participants or groups in which multiple measures of the dependent variable are taken before and after some experimental manipulation. Time-series designs provide some degree of control for history and maturation.

Interval scale—Scale of measurement in which the distance between any two adjacent scores is the same as the distance between any other two adjacent scores, but zero is not a true zero. An example of an interval scale is temperature measured in either Celsius or Fahrenheit.

Interview schedule—A standardised interview, with each question and procedure spelled out for the interviewer. Interview schedules provide consistency in interviews, which are a part of research projects.

Intuition—A way of acquiring knowledge. In intuition, ideas come to people, supposedly, without intellectual effort or sensory processes.

Latin square design—A procedure used to provide a measure of counterbalancing in a within-subjects design. Instead of using all possible orders of presentation (as in counterbalancing), a Latin square design uses a set of orders that ensure that every experimental condition appears equally often in every position in the order.

Likert scale items—In Likert scales, each item is presented on a continuum, with extreme positions at the end points. For example, the scale may range from “strongly agree” to “strongly disagree”.

Linear programming—An optimisation technique of operations research in which the objective function and constraints are linear functions.

Linear relationship—Relationship between two or more variables that, when plotted in a standard coordinate system, tend to cluster around a straight line. Most correlation coefficients are sensitive only to linear relationships between variables.

Logic—Set of operations that can be applied to statements and conclusions drawn from those statements to determine the internal accuracy of the conclusions.

Longitudinal (panel) design—A research design in which a group of participants is observed over time, with the dependent measures repeated during follow-up testing. Longitudinal designs are frequently used in the study of developmental aspects and historical research. This design is often contrasted with cross-sectional designs.

Magnitude—A characteristic of the abstract number system and of some measurement scales in which the numbers have an inherent order (for example, low to high).

Main effects—In a factorial design, main effects refer to the individual effects of the independent variables. In contrast, interaction effects are the combined effects of two or more independent variables on a dependent variable.

Manipulated factors—Independent variables in a factorial design in which the levels of the factors are determined by active manipulation by the experimenter.

Manipulated independent variable—Type of independent variable found in an experimental research study. When manipulated independent variables are used, participants are assigned to groups or conditions without bias.

Manipulation—Explicit control of the independent variable by the researcher.

Mann-Whitney U-test—A non-parametric inferential statistic used to test the difference between two groups when the dependent measure produces ordinal data.

Matched-pairs t-test—See correlated t-test.

Maturation—Potential confounding factor involving changes in participants on the dependent measure during the course of the study, which results from normal growth processes.

Mean—Arithmetic averages of scores. The mean is the most commonly used measure of central tendency, but should be computed only for score data.

Mean square—In analysis of variance (ANOVA), the mean square is a variance estimate. Several different mean squares are computed in any ANOVA. It is the ratio of mean squares that is the F-ratio and constitutes the inferential statistical test.

Measurement error—Any inaccuracy found in the measurement of a variable. Although it is impossible to determine the precise degree and direction of measurement error for a given participant, it is possible to specify the average error associated with a particular measure.

Median—Middle score in a distribution.

Meta-analysis—A procedure that allows the statistical averaging of results from independent studies of the same phenomena. Meta-analysis essentially combines studies on the same topic into a single large study, providing an index of how strongly the independent variable affected the dependent variable on an average in the set of studies.

Mode—Most frequent score in a distribution.

Metaphors—Metaphors are a way of communicating classification systems of qualitative analysis. They reveal the special properties of an object or event (for example, a teacher may be classified as a “Traffic cop”, in an “Ostrich role”, and so on.)

Models—In science, models are representations of the complex reality of the real world.

Moderator variable—Any variable that has an effect on the observed relationship between two or more other variables. When a moderator variable is operating, it is best to measure the relationship between variables separately in sub-groups defined by the moderator variable. For example, relationships between variables are often evaluated separately in males and females (a commonly used moderator variable).

Monte Carlo study—A procedure that evaluates the effectiveness of statistical tests by simulating with a computer the repeated sampling of participants from a population with known parameters. The characteristics of the populations can be systematically varied to see what effect these variations have on the accuracy of the statistical decision. This process allows one to empirically determine the probability of type I and type II errors, to see the strength of the impact of violations of the assumptions of statistical procedures.

Multidimensional scaling—A group of statistical methods that are used to simplify data by finding a small number of dimensions or factors that collectively account for most of the variability in a group of scores.

Multiple choice items—In a questionnaire, each question or item is presented with several answers, from which the respondent chooses one.

Multiple correlation—A correlation where one variable (say, X) is correlated with a set of variables. The correlation is computed by finding the linear combination of the set that will provide the highest possible correlation with the X variable.

Multiple observers—Control used to evaluate the accuracy of observations made by two or more independent observers.

Multivariate analysis of variance (MANOVA)—Extension of analysis of variance where two or more dependent measures are simultaneously evaluated.

Multivariate correlational designs—Correlational designs that include more than two variables.

Natural sciences—Branches of organised knowledge concerned with the material aspects of existence.

Negative correlation—Relationship between two variables in which an increase in one variable predicts a decrease in the other.

Nominal data—Data produced when a nominal scale of measurement is used. Nominal data are frequencies of participants in each of the specific categories.

Nominal fallacy—The tendency to confuse a label for a behaviour as an explanation for the behaviour. For example, labelling people as kind because they do many kind things for other people is reasonable, but it is unreasonable to say that they do those kind things because they are kind people.

Nominal scale—Scale of measurement in which only categories are produced as scores. Examples are diagnostic classification, sex of the participant, and political affiliation.

Nomothetic knowledge—Knowledge gained by studying and applying general laws and theories.

Nonequivalent control group design—Quasi experimental design used in field settings. In this design, two or more groups, which may not be equivalent at the beginning of the study, are compared on the dependent measure.

Non-experimental design—Any research design that fails to provide adequate controls for typical confounding.

Non-linear relationship—Any relationship between two or more variables that is characterised by a scatter plot where the points tend to cluster around a curved instead of a straight line. Most correlation coefficients are insensitive to non-linear relationships.

Non-manipulated independent variable—Pre-existing variable that determines group membership in a differential research study.

Non-parametric statistics—Inferential statistical procedures that do not rely on estimating population parameters such as the mean and variance.

Non-probability sampling—Any sampling procedure in which some participants have a higher probability of being selected than other participants or where the selection of a given participant changes the probability of selecting other participants. Often contrasted with probability sampling.

Normal distribution—Distribution of scores that are characterised by a bell-shaped curve in which the probability of a score drops off rapidly from the midpoint to the tails of the distribution. A true normal curve is defined by a mathematical equation and is a function of two variables (the mean and variance of the distribution).

Null hypothesis—Theory that States that the samples from each group are drawn from populations with identical population parameters. The null hypothesis is tested by inferential statistics.

Objective measure—Any measure that requires little or no judgment on the part of the person making the measurement. Objective measures are more resistant to experimenter biases than subjective measures.

Observation—Empirical process in which data about the phenomenon of interest is gathered and reported. Careful observation is a central task in all research.

Observational variable—Any variable that is observed and not manipulated in research. The term is usually used in low constraint research where the independent and dependent variable distinction does not apply.

One-way ANOVA—Statistical procedure that evaluates differences in mean scores of two or more groups where the groups are defined by a single independent variable.

Open-ended items—Items or questions on a questionnaire for which the respondent writes the answer (for example, an essay).

Operational definition—Detailed set of procedures used to measure or manipulate the level of a variable.

Ordered data—Data produced by ordinal scales of measurement.

Ordinal scale—Scale of measurement in which scores can be rank-ordered, but the distance between any two adjacent scores will not necessarily be the same as the distance between any other two adjacent scores.

Organismic variable—Any characteristic of the participant that can be used for classification. An organismic variable may be either directly observed (observed organismic variable) or may be inferred on the basis of the responses of the participant (response inferred organismic variable).

Panel design—See longitudinal design.

Paradigm of research—A framework within which all thinking and theories of science are ordered (Kuhn).

Parametric statistics—Inferential statistical procedures that rely on sample statistics to draw inferences about population parameters, such as mean and variance.

Parsimony—A guiding principle in science where a simple theory is preferred to a more complex theory if both theories explain the data equally well.

Partial correlation—A correlation between two variables (say X and Y) where the effects of a third variable (Z) are statistically removed from one of the two original variables before computing the correlation. Conceptually, it is the correlation between X and Y if Z were constant.

Participant observer—Any researcher gathering data in a setting in which the researcher is an active part. Participant observation tends to be less obtrusive than other observational procedures. However, the possibility for experimenter reactivity in participant observation is quite high.

Participant variable—Synonymous with organismic variable.

Partition—In an ANOVA calculation, the total sum of squares is divided (partitioned) into the between-groups sum of squares and the within-groups sum of squares.

Path analysis—A procedure that seeks to unravel causal links between variables from strictly correlational data by hypothesising detailed causal models and factoring the correlation matrix to see how closely the pattern of observed relationships fits the hypothesised causal model.

Pearson product-moment correlation—Index of the degree of linear relationship between two variables where each variable represents score data.

Percentile—Normative score that converts the raw score earned by a participant into a number from 0 to 100, which reflects the percentage of participants who scored lower than the participant in question.

Perfect correlation—Correlation of a +1.00 or a –1.00. When two variables are perfectly correlated, knowing the score on one variable permits perfect prediction of the score on the other. In a scatter plot, a perfect correlation is shown by all points falling on a straight line (but not a horizontal or vertical line).

Phases of research—Every research project develops through phases in which certain types of questions are asked and answered. These phases are idea-generating, problem-definition, procedures-design, observation, data-analysis, interpretation, and communication.

Pie chart—Statistical representation technique that uses sections of a circle to represent 100 per cent of a frequency distribution.

Placebo effect—In a treatment study, any observed improvement in response to a sham treatment. Placebo effects are probably the result of participants’ expectations regarding treatment effectiveness.

Population—Any clearly defined set of objects or events (people, occurrences, animals, and so on). Populations usually represent all events in a particular class (for example, all college students, all boys between the ages of 10 and 12, all headache sufferers).

Population parameters—Any summary statistic computed on the entire population.

Positivism—View that all true knowledge is scientific, in the sense of describing observable phenomena.

Post hoc tests (or comparisons or analyses)—Secondary analyses that evaluate relationships between variables, not specifically hypothesised by the researcher prior to the study.

Power of statistical test—Ability of an inferential statistical procedure to detect differences between groups when such differences actually exist.

Practical significance (substantive)—Often contrasted with statistical significance. Practical significance refers to whether the observed difference between two groups or conditions in a study is large enough to have a meaningful impact on the observed situation and/or subject of study.

Predictor—Variable in regression that is used to predict the scores on the criterion measure. For example, a test score (the predictor measure) might be used to predict future performance in a job.

Pretest—Use of a questionnaire (observation form) on a trial basis in a small pilot study to find out how well the questionnaire works.

Pretest-post-test design—Set of research designs in which participants are tested at two points in time—before and after the administration of the independent variable.

Pretest-post-test, natural control group design—Non-experimental research design in which pre-existing groups are each measured before and after the manipulation of an independent variable. These naturally occurring groups are assigned to different levels of the independent variable.

Probability—The ratio of specific events to the total number of possible events. For example, the probability of rolling snake-eyes (two ones) on one roll of a pair of dice is 1/36.

Probability sampling—A sampling procedure in which all participants have an equal probability of being selected and the selection of any participant does not change the probability of selecting any other participant. Often contrasted with non-probability sampling.

Problem definition (phase of research)—Research phase where vague and general research ideas are converted into precise questions to be studied.

Process of inquiry—Perspective taken by this text, which views research as a dynamic process focussed on formulating questions and systematically answering those questions through carefully controlled observation/studies.

Programme evaluation research—Specific area of field research for evaluating the effectiveness of a programme in meeting its stated goals.

Projective techniques—Various tests used to disguise the study objective and to allow the respondent to transfer or project attitudes and behaviour on sensitive subjects to the researcher.

Protocol analysis—Qualitative research method in which an episode of a manager’s work is analysed as it occurs or immediately after it has occurred. The method uses the trust and confidence of the manager and gets the viewpoint of the manager about the episode.

Pseudoscience—Popular distortions of scientific knowledge and procedures that appear on the surface to be scientific, but in reality lack critical scientific procedures. Some fields such as astrology, extrasensory perception, the study of alien abductions, and medical quackery, have traditionally relied on pseudoscience to make them appear legitimate.

Psychology—Scientific study of the behaviour of organisms.

Pure research—Another term for basic or fundamental research (see basic research).

p-value—The probability of obtaining the statistic (for example, t or F) or a larger statistic, by chance, if the null hypothesis is true. Statistical analysis programmes routinely compute p-values in addition to the test statistic.

Q-Sort technique—General methodology for gathering and processing data. The subjects are asked to sort a number statements. The aim is to determine their relative ranking.

Quasi experiential design—Research design that, although not a true experimental design with all experimental controls built in, provides experiment-like controls to minimise threats to internal validity.

Questionnaire—An instrument that lists questions to be asked of participants.

Quota sampling—Purposive sampling in which relevant characteristics are used to stratify the sample to improve representativeness.

Random-number generator—A computer function that generates an endless sequence of random numbers.

Random sampling—Procedure for the selection of participants to be included in a research study, where each participant in the population has an equal chance of being selected and where the selection of any one participant will not affect the probability of selecting any other participant. In most research, random sampling from the population is not carried out because the procedure is not feasible. Instead, researchers rely on sampling from an accessible population.

Randomisation—Any procedure that assigns a value or order in an unpredictable or random way, such as by use of tables of random numbers. Randomisation procedures may be used for selecting participants, assigning participants to groups or conditions, or assigning the order in which a participant will experience a number of successive conditions.

Range—Distance between the lowest score and the highest score, inclusive of the scores.

Ratio scale—Scale of measurement in which the intervals between scores are equal (as in the interval scale) and the zero point on the scale represents some of the quality being measured (a true zero). Examples of ratio scales are height, weight, and frequency of an event.

Rationalism—One of the approaches to study about the universe, rationalism relies on systematic logic and a set of premises from which logical inferences are made.

Regression—A statistical procedure that produces an equation for predicting a variable (the criterion measure) from one or more other variables (the predictor measures).

Regression equation—Mathematical equation that predicts the value of one variable from one or more other variables. Most regression equations are linear regression equations.

Relationship—Any connection between two or more variables. In research, there are many types of relationships, from simple contingencies to established causal relationships.

Reliability—Index of the consistency of a measuring instrument in repeatedly providing the same score for a given participant. There are many different types of reliability, each referring to a different aspect of consistency. Types of reliability include interrater reliability, test-retest reliability, and internal consistency reliability.

Repeated measures ANOVA—Statistical procedure to evaluate the mean differences between two or more conditions, where the same participants contribute scores in each condition. The repeated measures ANOVA takes into account the fact that the same participants appear in all conditions.

Repeated-measures design—Any research design in which participants are tested more than once. Examples of such designs are pretest-post-test designs within-subjects designs, and time series designs.

Replication—A repeat study with either no changes at all in the procedure exact replication) or carefully planned changes in the procedure (systematic or conceptual replication).

Representative sample—Sample of participants that adequately reflects the characteristics of the population from which the sample is drawn.

Representativeness—Degree to which a sample is representative of the population from which it is drawn.

Research design—A map or a blueprint for achieving research objectives.

Research ethics—Set of guidelines designed to protect human and non-human participants from the risks of participating in research.

Research hypothesis—Precise and formal statement of a research question. The research hypothesis is constructed by adding operational definitions for each of the variables to the statement of the problem.

Research setting—Any characteristic of the situation and/or surroundings in which a research project is carried out. Settings may vary from natural, real-world settings to highly constrained and carefully controlled laboratory situations.

Reversal (ABA) design—Research design often used with single participants, where the effects of an independent variable on a dependent variable are inferred from observations made first without the independent variable being present, then with the independent variable being present, and again without the independent variable present.

Rival hypothesis—Any feasible alternative hypothesis to the causal hypothesis.

Robust—Term used in reference to statistical procedures. A statistical test is said to be robust to the violation of the assumptions on which the test is based if the test consistently leads to accurate conclusions in spite of the assumption violations.

Sample—Any subset drawn from a population. Researchers work with samples of participants and draw inferences about the larger population.

Sample statistic—Descriptive index of some characteristic of the sample of participants. Population parameters are estimated on the basis of statistics.

Sampling—Process of drawing a sample from a population. Many sampling techniques are available, including random sampling, stratified random sampling, and various non-random sampling techniques.

Sampling error—Chance variation among different samples drawn from the same population.

Sampling frame—In survey research, a sampling frame is a list of all participants from an available population. The sampling frame is a subset of a larger population from which a representative sample is drawn.

Scales of measurement—Characteristics of the scores produced by a particular measurement instrument. Scales of measurement vary depending on how closely scores match the real number system. There are four generally recognised scales of measurement: nominal scale, ordinal scale, interval scale, and ratio scale.

Scatter plot—Graphic technique that illustrates the relationship between two or more variables. In a two-variable situation, the scatter plot is constructed by labelling the x-axis with one of the variables and the y-axis with the other variable and plotting each participant’s part of scores in the xy coordinate system.

Science—Way of knowing about the universe around us, which combines rationalism and empiricism to form a system that places great demands or procedures, data, and theories.

Score data—Data produced by interval or ratio scales of measurement.

Selection—Potential confounding variable in any research project. Selection represents any process that may create groups not equivalent at the beginning of the study.

Sensitivity analysis—An analysis used in mathematical modelling, where the sensitivity of model results to variations in a particular variable is studied.

Simple random sampling—See random sampling.

Single-blind procedure—Research procedure in which the researcher is unaware of the condition assigned to each participant. The purpose of the single-blind procedure is to minimise measurement biases.

Single-group, post-test only design—Nonexperimental research design in which the researcher manipulates the independent variable and then takes a postmanipulation measure on the dependent variable. The difference between this design and an ex post facto design is the actual manipulation of the independent variable by the researcher.

Single-group, pro-test-post-test design—Non-experimental design in which a group of participants is measured on a dependent variable. The independent variable is manipulated, and a second measure on the dependent variable is taken.

Skepticism—Unwillingness to accept information as valid knowledge without some documentation to confirm the information. Skepticism is one of the strongest tools available to a scientist.

Snowball sample—Judgement sample that relies on the researcher’s ability to locate an initial set of respondents with desired characteristics, who will help in locating other similar respondents.

Solomon’s four-group design—Sophisticated experimental design that combines the randomised, post-test only, control group design and the randomised pre-test-post-test control group design.

Spearman rank-order correlation—Correlation coefficient that indexes the degree of relationship between two variables, each of which is measured on an ordinal scale of measurement.

Spread—Synonymous with variability.

SPSS for Windows—A computer package (statistical package for the social sciences) for statistical data analysis on a Windows-based computer.

Standard deviation—Square root of variance. The standard deviation is an index of variability in the distribution of scores.

Standard error of the differences between means—In statistics the denominator in a t-test.

Standard score—A score that gives the relative standing in a distribution. A standard score is computed by subtracting the distribution mean from the score and dividing the value by the standard deviation from the distribution.

Statement of the problem—First major refinement of initial research ideas, in which a clear statement of the expected relationship between conceptual variables is made. The statement of the problem is refined into one or more research hypotheses by specifying the operational definitions of each conceptual variable in the statement.

Statistical Analysis Systems (SAS)—Computer package for statistical data analysis.

Statistical hypothesis—Synonymous with null hypothesis.

Statistical Package for the Social Sciences (SPSS)—Computer package for statistical data analysis.

Statistical significance—A finding is said to achieve statistical significance if it is unlikely that such a finding would have occurred by chance alone (see statistically significant differences).

Statistical validity—Accuracy of conclusions drawn from a statistical test. To enhance statistical validity, one must meet the critical assumptions and requirements of a statistical procedure.

Status survey—A simple survey designed to provide a description of the current status of some population characteristic.

Stem and leaf display—It is an explorative data analysis technique to reveal frequency distribution for each data value.

Stimulus variable—Any variable part of the environment to which an organisation reacts. A stimulus variable may be a natural part of the environment and observed by the researcher, or may be actively manipulated by the researcher.

Strata—Sub-populations within populations from which we draw samples based on the base rates in the population of the factor(s) that determine the strata.

Stratified random sampling—Variation of the random sampling procedure in which a population is divided in narrow strata along some critical dimension. Participants are then selected randomly from each of the strata in the same proportion that the strata are represented in the population. Stratified random sampling can increase the representativeness of the sample and is used extensively in sophisticated survey research.

Stress—Measure of ‘badness of fit’ of configuration determined in multidimensional scaling analysis when compared with original input data.

Structuralism—A philosophical perspective in which the scientist seeks to identify the structure of the underlying mechanisms that control behaviour, such as consciousness. This approach was popularised by Wundt. Often contrasted with functionalism.

Structural modelling—Method of scenario generation of a problem space to help develop a mathematical model.

Subject effects—Any response by participants in a study that does not represent the way they would normally behave if not under study. Two powerful participant effects are the placebo effect and a participant’s response to the demand characteristics of the study.

Subject selection—See participant selection.

Subjective measures—Measures based primarily on uncorroborated opinions, feelings, biases, or judgements. Subjective measures, as contrasted with objective measures, are more prone to distortions due to subject or experimenter effects.

Summary statistics—Descriptive statistics that provide some general characteristic of the sample in a single number. Typical summary statistics are the mean, median, variance, and standard deviation.

Sum of squares—Sum of the squared differences from the mean. The sum of squares is the numerator in the variance formula.

Survey—A set of one or more questions posed to a group of participants about their attitudes, beliefs, plans, life-styles, or any other variable of interest. Surveys may be conducted over the phone, in person, or through the use of a written form.

Survey research—Research that seeks to use survey procedures to identify relationships among the variables being surveyed.

Systematic replication (or conceptual replication)—Situation where a study is repeated with small, theory-based changes in the procedures. Systematic replication is more common than exact replication, because it verifies original findings while also expanding knowledge of the phenomena.

System study—Study of decision makers, information systems, decision-making procedures, and the environments of a decision-making system before modelling a particular decision problem.

Table of random numbers—A table containing a long list of randomly generated numbers. Such tables are used frequently in research for random selection and assignment of participants. A table of random numbers is included in Appendix D of this text.

Target population—Population to which we hope to generalise the findings of a research study. In most research, the entire target population is not accessible to the researcher.

Testing—Potential confounding variable in research. Testing represents any change in a participant’s score on a dependent measure, which is a function of the participant that has been tested previously in the research project.

Test-retest reliability—Index of consistency in scores over time. Test-retest reliability is computed by calculating the Pearson’s product-moment correlation between scores from two testings, separated by some specified time interval.

Theoretical concept—Abstraction (thought or idea) that defines the relationship between two or more variables.

Theory—In science, theory is the collection of ideas about how and why variables are related to one another. Theory is usually built on empirical observations and is validated by making predictions deduced from the theory, which are then empirically tested.

Transcription of data—It is the conversion of recording of interviews or conversations into a readable document. It is widely used in qualitative research.

Treatment—See manipulation.

True zero—Characteristic of a measurement scale where zero represents none of the concepts being measured.

t-test—Statistical procedure designed to test for mean differences between two groups of participants.

t-test for independent groups—Statistical procedure designed to test for mean differences between two groups of participants, all participants in the study appear in one and only one group.

Two-group, posttest-only design—A design in which two groups of participants are compared once after some manipulation of the independent variable.

Two-way ANOVA—Statistical procedure for the analysis of a factorial design with two independent variables.

Typologies—Classification systems used in qualitative research for description purposes. They are made up of categories of the world or a phenomenon particularly of the behaviour of people.

Type I error—Probability of rejecting the null hypothesis when the null hypothesis is true.

Type II error—Probability of not rejecting the null hypothesis when the null hypothesis is false.

Univariate—Having to do with one variable. For example, a univariate distribution would provide the distribution for a single variable.

Univariate designs—See single-variable designs.

Univariate Analysis—Statistical analysis of single variables, mainly hypothesis testing.

Unobtrusive measure—Any measure that can be taken of participants without their being aware that they are being measured.

Unobtrusive observer—Anyone who is able to observe the behaviour of participants without their being aware they are being observed.

Validity—Major concept in research that has several specific meanings (internal validity, external validity, construct validity, and statistical validity). In a general sense, validity refers to the methodological and/or conceptual soundness of research (for example, in the case of an experiment, a question regarding the validity of an experiment is, “Does this experiment really test what it is supposed to test?”)

Variability—Differences among participants on any given variable.

Variable—Any characteristic that can take on different values. Variables are sets of events measured in research. Research is aimed at defining the relationships between variables.

Variance—Summary statistic that indicates the degree of variability among participants for a given variable. The variance is essentially the average squared deviation from the mean and is the square of the standard deviation.

Within-group variance—Variability among participants within a particular group or condition. Provides a basis for comparing mean differences between groups in most statistical procedures.