back to top

acquiescence response set  a response bias whereby respondents tend to answer in the direction of agreement, regardless of item content. (chapter 10)

active membership role  a type of participant role in field research in which the researcher assumes a functional role in the group or organization he or she is studying; see also complete and peripheral membership roles. (chapter 11)

aggregate data  information about one set of units that is statistically combined to describe a larger social unit. For example, information about students (gender, race, college board scores, etc.) might be aggregated to describe characteristics of their colleges (sex ratio, ethnic composition, average college board score, etc.). (chapter 4)

analytical history  the use of historical events and evidence to develop a generalized understanding of the social world; see also descriptive history. (chapter 12)

anonymity  an ethical safeguard against invasion of privacy; the condition wherein researchers are unable to identify data with particular research participants. (chapter 3)

antecedent variable  a variable causally antecedent to others in a theoretical model. Two variables may be spuriously associated because both are affected by an antecedent variable. (chapters 4, 16)

applied social research  research undertaken for the explicit purpose of providing information to solve an existing social problem; see also basic social research. (chapter 14)

association  the strength of the observed relationship between two variables. (chapter 4)

attrition  a threat to internal validity; the loss of subjects during the course of a study (also called mortality); see also differential attrition. (chapter 8)

available data  existing sources of information that were not produced directly by the social researcher who uses them; this includes public and private documents, products of the mass media, physical evidence, and data archives; the use of available data is one of the four major approaches to social research. (chapters 1, 12, 13)

back to top

causal relationship  a theoretical notion that change in one variable forces, produces, or brings about a change in another. Although the concept of “cause” is unobservable and philosophically controversial, causal theorizing is commonly and productively used in scientific investigations. (chapters 2, 4)

cell frequency  the number of cases in a cell of a cross-tabulation table. (chapter 15)

chi-square test for independence  a test of statistical significance used to assess the likelihood that an observed bivariate relationship differs significantly from that which easily could have occurred by chance. (chapter 15)

closed-ended questions  survey questions that require respondents to choose responses from those provided; also called “fixed-choice questions.” (chapter 10)

cluster sampling  a probability sampling procedure in which the population is broken down into natural groupings or areas, called clusters, and a random sample of clusters is drawn. Cluster sampling may occur in a series of stages, moving from larger to smaller clusters, with individual cases sampled at the last stage; see multistage cluster sampling. (chapter 6)

codebook  a “dictionary” for a survey study, which lists the answers or categories that correspond to each numerical code, the location of each variable in the data file, and coding and decision rules. (chapter 15)

coding  (1) the sorting of raw data, such as responses to open-ended questions or field observations, into categories; (2) for computer analysis, coding consists of assigning numbers or symbols to variable categories (chapters 10, 11, 15)

cognitive interviewing  verbal reports from respondents, collected during (“thinkalouds”) or after (follow-up probes, paraphrasing requests) responding to pretest questions, and used to diagnose question wording, ordering, and formatting problems. (chapter 10)

cohort  persons (or other units) who experience the same significant life event (e.g., birth, marriage, high school graduation) within a specified period of time; see cohort study. (chapter 9)

cohort study  a longitudinal research design that attempts to assess the relative effects of age (life-course changes), period (current history), and cohort (“generations” who share the same past history) on attitudes and behavior. (chapter 9)

collinear association  a perfect linear relationship between two variables. (chapter 16)

Common Rule informal label for the set of regulations for the treatment of human research subjects which are codified under Title 45 of the Code of Federal Regulations, Part 46 (45 CFR 46). (chapter 3)

complete membership role  a type of participant role in field research in which the researcher assumes full-member status in the organization or group he or she is studying; see also active and peripheral membership roles. (chapter 11)

computer-assisted personal interviewing (CAPI)  a software program, usually on a portable computer, that aids interviewers by providing appropriate instructions, question wording, and data-entry supervision. (chapter 9)

computer-assisted self interviewing (CASI)  an electronic survey in which a questionnaire is transmitted on a computer disk mailed to the respondent or on a laptop computer provided by the researcher. (chapter 9)

computer-assisted telephone interviewing (CATI)  a set of computerized tools that aid telephone interviewers and supervisors by automating various data collection tasks. (chapter 9)

concept  abstractions communicated by words or other signs that refer to common properties among phenomena (e.g., the concept “extroversion” represents a broad range of specific behaviors); concepts developed for scientific purposes are sometimes called “constructs.” (chapters 2, 5)

conceptualization  the development and clarification of concepts. (chapter 5)

conditional a statement (the antecedent) introduced by an “if,” followed by a second statement (the consequent) preceded by a “then.” For example, “If this is a conditional proposition, then it should have an antecedent and a consequent.” (chapter 4)

confidence interval  a range (interval) within which a population value is estimated to lie at a specific level of confidence; used to qualify sample estimates to take into account sampling error. For example, a researcher might report that he or she is 99-percent confident (confidence level) that the mean personal income for a population lies within plus or minus $359 of the sample mean of $18,325 (i.e., confidence interval of $18,684–$17,966). (chapter 6)

confidentiality  an ethical safeguard against invasion of privacy; the assumption that all data on research participants are given to the researcher in strict confidence, not to be divulged to anyone without the participants’ permission. (chapter 3)

consistency checking  a data-cleaning procedure involving checking for unreasonable patterns of responses, such as a 12-year-old who voted in the last presidential election. (chapter 15)

construct validation  measurement validation based on an accumulation of research evidence, which may include evidence that the measure in question (1) relates to other variables in a theoretically expected manner, (2) correlates highly with other measures of the same concept (convergent validity), (3) correlates not too highly with measures of other concepts (discriminant validity), and (4) varies among groups known to differ on the characteristic being measured. (chapter 5)

content analysis  a set of methods for analyzing the symbolic content of communications, which typically entails (1) defining a set of content categories, (2) sampling elements of the communication that are described by the categories, (3) quantifying the categories such as by counting their frequency of occurrence, and (4) relating category frequencies to one another or to other variables. (chapter 12)

content validity  a subjective judgment of whether a measure adequately represents all facets (the domain) of a concept. (chapter 5)

contextual design  a survey design in which information about respondents and their social environments (contexts) is collected for the purpose of studying the separate and joint effects on individuals of personal characteristics and of social contexts. (chapter 9)

context units  larger units in which the recording units of content analysis are embedded, which provide the “context” necessary to classify the recording units. For example, the meaning of words (recording units) might be determined from the sentences (context units) in which they appear; see also recording units. (chapter 12)

contingency question  a survey question intended for a subset of the respondents, addressing only those persons for whom the question is relevant; see also filter question. (chapter 10)

control  a procedure (or procedures) that eliminates, as far as possible, unwanted variation, such as sources of bias and error that may distort study results. A common approach is to control potentially confounding variables by holding them constant or preventing them from varying. (chapters 2, 4)

control variable  a variable that is held constant during the course of observation or statistical analysis. (chapter 4)

convenience sampling  a form of nonprobability sampling in which the researcher simply selects cases that are conveniently available; also called “haphazard,” “fortuitous,” and “accidental” sampling. (chapter 6)

convergent validity  the extent to which independent measures of the same concept correlate with each other; the higher the correlation, the greater the convergent validity. (chapter 5)

correlation coefficient  a measure of association, symbolized as r (Pearson’s), that describes the direction and strength of a linear relationship between two variables measured at the interval or ratio level; the square of Pearson ’s r represents the proportion of variance in one variable that may be predicted from the other using linear regression. (chapters 4, 15).

cost-benefit analysis  (1) an examination of the potential costs (e.g., harm to subjects) and benefits (e.g., knowledge gained, beneficial application) of a study as a way of assessing the ethics of the study; (2) in evaluation research, an assessment of the economic efficiency of a program in terms of its monetary costs and benefits; see also cost-effectiveness analysis. (chapters 3, 14)

cost-effectiveness analysis  in evaluation research, an assessment of the efficiency of a program by measuring intervention outcomes in terms of program costs; see also cost-benefit analysis. (chapter 14)

counterbalancing a technique used to control for order effects in within-subjects experimental designs; the sequence of treatment and control conditions is reversed, so that different groups of subjects experience each sequence. (chapter 8)

cover letter  in a survey, a letter designed to obtain cooperation from persons in the sample. The cover letter is sent before the interviewer calls on the respondent, or, in mail surveys, it accompanies the questionnaire. (chapter 9)

cover story  an introduction provided to experimental subjects to obtain their cooperation while disguising the research hypothesis. (chapter 7)

coverage error in sampling, the error that occurs when the sampling frame does not match the intended target population. (chapters 6, 9)

covert research a type of field research in which the investigator conducts research in private settings while concealing his or her identity as a researcher. (chapter 11)

criterion-related validation  a method of assessing the validity of a measure in terms of the degree to which it correlates with an objective current or future criterion. (chapter 5)

cross-national surveys  studies in which equivalent-sample surveys are conducted in different countries; also called “comparative studies.” (chapter 9)

cross-sectional design  the most common survey design, in which data on a cross section of respondents chosen to represent a larger population of interest are gathered at essentially one point in time; see also longitudinal design. (chapter 9)

cultural relativity  the principle that the cultural standards of a given society must be examined on their own terms and that researchers should be nonjudgmental regarding the society or group that is being studied. (chapter 3)

back to top

data archives  repositories of survey or ethnographic data collected by various agencies and researchers that are accessible to the public. (chapter 12)

data cleaning  the detection and correction of errors that occur during data collection, coding, and data entry. (chapter 14)

data matrix  the form of a computer data file, with rows as cases and columns as variables; each cell represents the value of a particular variable (column) for a particular case (row). (chapter 15)

data modification  the process of transforming raw data for analysis, such as by collapsing the categories of a variable or imputing values for missing cases. (chapter 15)

data processing  the preparation of data for analysis; in survey research it entails five steps: coding, editing, entry, cleaning, and modification. (chapter 15)

debriefing  a session at the end of an experiment in which the experimenter discusses with the subject what has taken place, the real purpose of the study, the need for confidentiality, the subject’s responses and feelings, and so on. (chapters 3, 7)

deductive reasoning  a process of reasoning in which the conclusion necessarily follows if the evidence is true. (chapter 2)

demand characteristics  cues in an experiment that convey to subjects the experimenter’s hypothesis or what is expected of them. (chapter 7)

dependent variable  a variable that the researcher tries to explain or predict; the presumed effect of one or more independent variables. (chapter 4)

descriptive history  type of historical analysis that emphasizes the complete and accurate description of past events; see also analytical history. (chapter 12)

descriptive research  studies undertaken to collect facts about a specified population or sample, for example, a public opinion poll. (chapters 4, 9)

descriptive statistics  procedures for organizing and summarizing data. (chapter 15)

differential attrition  a threat to internal validity; the existence of varying dropout rates among conditions of an experiment, which tends to make the conditions nonequivalent in composition; also called “differential mortality.” (chapter 8)

direct effects  when one variable is hypothesized to affect another directly in a causal model; there may also be “indirect effects” in which the impact of one variable on another is transmitted through one or a series of causally intervening variables. The total impact of one variable on another is the sum of the direct and indirect effects. (chapter 16)

direct questions  survey questions in which there is a direct, clear link between what is asked and what the researcher wants to know; see also indirect questions. (chapter 10)

direction of influence  for a presumed asymmetric causal relationship, the identification of which variable is the cause (independent) and which the effect (dependent). (chapter 4)

discriminant validity  the extent to which a measure of a particular concept differentiates that concept from other concepts from which it is intended to differ. Discriminant validity is lacking if measures of supposedly different concepts correlate too highly. (chapter 5)

disproportionate stratified sampling  a sampling procedure in which strata are sampled disproportionately to population composition. For example, a study of religious leaders might sample female pastors at a much higher rate than male pastors to ensure a sufficient number of the former for analysis purposes. (chapter 6)

distorter variable in elaboration analysis, a test variable that reverses the direction of the original bivariate relationship, as from a positive zero-order association to a negative association in the partial tables. (chapter 16)

double-barreled question  a survey question in which two separate ideas are erroneously presented together in one question. (chapter 10)

double-blind technique  a method of preventing both subjects and research personnel from knowing the subjects’ treatment conditions during the running of an experiment. (chapter 7)

dummy variable  a data-modification procedure that involves recoding the categories of nominal- or ordinal-scale variables for the purpose of regression or other numerical analysis. For example, gender categories may be represented by a single dummy variable having a value of 1 if the respondent is female and a value of 0 if male. (chapters 15, 16)

back to top

ecological fallacy  erroneous use of information pertaining to an aggregate (e.g., organizations) to draw inferences about the units of analysis that comprise the aggregate (e.g., individual members of organizations). (chapter 4)

editing  a quality-control process designed to ensure that survey or other data to be read into the computer are as complete, error-free, and readable as possible. (chapter 15)

effect assessment  the most basic type of evaluation research, undertaken to assess the impact of a social program on program participants; also called "summative" or "outcome" evaluation; see also efficiency assessment. (chapter 14)

efficiency assessment  a type of evaluation research in which a program’s outcomes are weighed against its costs; may take the form of cost-benefit or cost-effectiveness analysis; see also effect assessment. (chapter 14)

elaboration  a traditional technique for the multivariate analysis of contingency tables that “elaborates” the relationship between two variables by introducing a third (and sometimes additional) variable and testing the resultant causal models. (chapter 16)

empirical/empiricism  a way of knowing or understanding the world that relies directly or indirectly on what we experience through our senses—sight, hearing, taste, smell, and touch; admissible evidence in science is limited to empirical phenomena. (chapter 2)

empirical generalization  a generalization or hypothesis inductively derived from observations. (chapter 2)

environmental impact study  a study undertaken, usually as required by law, to assess the impact of a public works project on the social and physical environment; see also social impact assessment. (chapter 14)

ethics  guidelines or standards for moral conduct. In research, ethical codes prescribe principles for upholding the values of science and for resolving conflicts between scientific ideals and societal values. (chapter 3)

evaluation apprehension  subjects’ anxiety about being evaluated by the experimenter, typically a psychologist, which may make them overly concerned with producing “normal” behavior. (chapter 7)

evaluation research  an area of social science research, often linked to policymaking, concerned with analyzing social policies, programs, and interventions. (chapter 14)

exhaustive  the measurement requirement that a measure includes all possible values or attributes of a variable, so that every case can be classified. (chapter 5)

expedited review an IRB review procedure for studies involving no more than minimal risk of harm in which a subset of IRB members, usually the chair, determines whether a study meets federal ethical guidelines; see also institutional review board. (chapter 3)

experiment  a major approach to social research that entails the manipulation of the independent variable and reseacher control over the events to which research participants are exposed. (chapters 1, 7, 13)

experimental realism  when subjects become involved in or are affected by the procedures of an experiment rather than remaining detached. (chapter 7)

experimenter expectancy effect  effect on subject’s behavior due to the experimenter’s expectations about how the experiment will turn out. (chapter 7)

explanation (1) in science, abstract statements that relate changes in one class of events to changes in another; (2) in elaboration analysis, an outcome in which, upon controlling for an antecedent test variable, the original bivariate relationship reduces to zero or near zero in each partial table. (chapters 2, 16)

explanatory research  studies that investigate relationships between two or more variables, attempting to explain them in cause-and-effect terms. (chapters 4, 9)

explanatory variables  those variables that are the focus of the research—the independent and dependent variables in a hypothesized relationship; see also extraneous variables. (chapter 4)

exploratory research  studies undertaken to explore a phenomenon or topic about which very little is known. (chapter 3)

external validity  the extent to which experimental findings are generalizable to other settings, subject populations, and time periods. (chapter 7)

extraneous variables  all variables other than the independent and dependent variables in a hypothesized relationship; see also explanatory variables. (chapter 4)

back to top

face-to-face interview  a type of interview in which the interviewer interacts face-to-face with the respondent; also called "in-person interview"; see also telephone interview. (chapters 9, 13)

face validity  a personal judgment that an operational definition appears, on the face of it, to measure the concept it is intended to measure. (chapter 5)

factorial experimental designs  an extension of the basic experimental design in which two or more independent variables (factors) are manipulated; information is provided about the separate (main) effects and joint (interaction) effects of the independent variables. (chapter 8)

field experiment  a “true” experimental design conducted in a natural setting. (chapter 7)

field interviewing  interviewing in field research; the researcher may question others through either casual conversation or formal, in-depth interviews; see also in-depth interviews. (chapter 11)

field jottings  short phrases, quotes, key words, and the like recorded by field researchers while in the field; see also field notes. (chapter 11)

field notes  thorough account of field observations recorded by field researchers as soon after each period of observation as possible, preferably at the end of each day. (chapter 11)

field pretesting  the evaluation of the survey instrument and personnel under realistic field conditions with respondents similar to those for whom the survey is designed. (chapter 10)

field research  a major approach to social research that involves directly observing and often interacting with others, usually for an extended period, in a natural setting. (chapters 1, 11, 13)

filter question  a type of survey question, the responses to which determine which subjects are to answer which subsequent contingency questions. (chapter 10)

first-order relationship  a relationship between two variables in which a third variable is controlled; see also zero-order relationship. (chapter 16)

fixed-effects model  a statistical framework that assumes, in meta-analysis, that between-study variability arises from sampling and other chance processes; see also random-effects model. (chapter 13)

focus group a small group of participants who engage in unstructured discussions, focused on a general topic and guided by a skilled interviewer. (chapter 10)

formative evaluation  research undertaken in the developmental stages of a social program to guide program design; see evaluation research. (chapter 14)

frequency distribution  a tabulation of the number of cases falling into each category of a variable. (chapter 15)

frequency polygon  the graphic presentation of an interval/ratio-scale variable in which the X-axis consists of the values of the variable and the Y-axis consists of the number of cases (or frequency); a line connects the frequency for each value of the variable; in a percentage polygon, the Y-axis consists of percentages. (chapter 15)

funnel sequence  a sequence of survey questions that progresses from a very general question to gradually more specific questions. (chapter 10)

back to top

gatekeepers  authorities whose permission is needed to conduct research in their setting. (chapter 11)

grounded theory  theory developed inductively from firsthand observations, in contrast to theories generated by other means. (chapter 11)

Guttman scaling  a scaling procedure that attempts to ensure a unidimensional scale by selecting a set of items that, when ordered in terms of their “strength,” will order individuals or other relevant units on the concept being measured. (chapter 13)

back to top

heterogeneity  the degree of dissimilarity among cases with respect to a particular characteristic. For example, the gender composition of groups might vary from 100 percent female (maximum homogeneity) to 50 percent female (maximum heterogeneity). (chapter 6)

history  a threat to internal validity; events in the subjects’ environment, other than the intended experimental manipulation, that occur during the course of an experiment and that may affect the outcome. (chapter 8)

hypothesis  an expected but unconfirmed relationship among two or more variables. (chapters 2, 4)

back to top

imputation  a procedure for handling missing data in which missing values are imputed from other information, such as the sample mean or known values of other variables. (chapter 15)

independent variable  a presumed cause of a dependent variable. (chapter 4)

in-depth interviews  intensive interviews that in field research are much less structured than in survey research and are much longer, often requiring several sessions. (chapter 11)

index  a composite measure of a concept constructed by adding or averaging the scores of separate indicators; differs from a scale, which uses less arbitrary procedures for combining indicators. (chapters 5, 13)

indicator  an empirical manifestation of a concept. For example, the indicator “years of schooling” often represents the concept “education.” (chapters 5, 13)

indirect effects  in a causal model, when one variable is hypothesized to affect another indirectly through one or a series of intervening variables; see also direct effects. (chapter 16)

indirect questions  questions in which the relationship between the researcher’s objectives and the questions asked is not obvious; a technique usually based on the psychological concept of "projection"; see also direct questions. (chapter 10)

inductive reasoning  a reasoning process in which the conclusion goes beyond the evidence; unlike in a deductive argument, the evidence may be true and the conclusion false. (chapter 2)

inferential statistics  procedures for determining the extent to which one may generalize beyond the data at hand. (chapter 15)

informant a person from whom field researchers acquire information in the field. (chapter 11)

informed consent  an ethical practice of providing research participants with enough information about a study, especially its potential risks, to enable them to make an informed decision about whether to participate. (chapter 3)

institutional review board (IRB)  a committee formed at nearly all research institutions (e.g., universities) that is responsible for reviewing research proposals in order to assess provisions for the ethical treatment of human (and animal) subjects; IRB approval is required for federally funded research. (chapter 3)

instrumentation  a threat to internal validity; unwanted changes in characteristics of the measuring instrument or measurement procedure. (chapter 8)

interaction effect  an outcome in which the effect of one independent variable on the dependent variable varies according to the value or level of another independent variable. That is, the effects of the variables together differ from the effects of either alone. (chapter 8)

intercoder reliability  an “equivalence” method for assessing reliability that examines the extent to which different interviewers, observers, or coders get equivalent results using the same instrument or measure. (chapter 5)

internal-consistency reliability  an “equivalence” method of assessing reliability in which a statistical procedure is used to examine the consistency of “scores” across all the items of a composite measure. (chapter 5)

internal validity  sound evidence in an experiment that rules out the possibility that extraneous variables, rather than the manipulated independent variable, are responsible for the observed outcome. (chapters 7, 8)

interpretation an outcome in elaboration analysis in which, upon controlling for an intervening test variable, the original bivariate relationship reduces to zero or near zero in each partial table. (chapter 16)

interrupted time-series design  a quasi-experimental design resembling the one-group pretest-posttest design but with a series of observations (measurements) before and after the treatment manipulation. (chapter 8)

intersubjective testability  a condition wherein two or more scientists can agree on the results of observations. (chapter 2)

interval measurement  a level of measurement that has the qualities of the ordinal level plus the requirement that equal distances (intervals) between assigned numbers represent equal distances in the variable being measured; consequently, with interval measurement it is possible to perform basic mathematical operations such as addition and subtraction. (chapter 5)

intervening variable  a variable that is intermediate between two other variables in a causal chain. For example, if the model specifies that X affects W, which in turn affects Y, then W is an intervening variable that interprets the causal process by which X affects Y. (chapters 4, 16)

interview guide  an outline of topics and questions used in in-depth interviewing; unlike the interview schedule in a structured inteview, the exact wording and order of the questions may vary. (chapter 11)

interview schedule  a survey form used by interviewers that consists of instructions, the questions to be asked, and, if they are used, response options. (chapter 9)

interviewer debriefings  a field pretest technique, usually involving focus group discussions, used to identify instrument problems from the interviewer’s perspective. (chapter 10)

inverted-funnel sequence  a sequence of survey questions that begins with the most specific questions on a topic and ends with the most general. (chapter 10)

back to top

key informant  a contact who helps a field researcher gain entry to, acceptance within, and information about the research setting. (chapter 11)

back to top

law  a proposition that has been repeatedly verified scientifically and is widely accepted. (chapter 2)

leading question  a survey question that suggests a possible answer or makes some responses seem more acceptable than others. (chapter 10)

Likert response format  the ordered responses (usually “strongly agree,” “agree,” “neutral,” “disagree,” and “strongly disagree”) to individual items; see Likert scaling. (chapter 10)

Likert scaling  a scaling approach commonly used to measure attitudes in which respondents choose from an ordered series of responses (e.g., ranging from “strongly approve” to “strongly disapprove”) to indicate their reaction to each of a set of statements. (chapter 13)

listwise deletion  a common procedure for handling missing values in multivariate analysis that excludes cases which have missing values on any of the variables in the analysis. (chapter 15)

longitudinal design  survey design in which data are collected at more than one point in time; see trend study, panel study, and cohort study. (chapter 9)

back to top

main effect  in a factorial experimental design, the effect of a single independent variable (factor) by itself. (chapter 8)

manipulation check  evidence collected in an experiment that the manipulation of the independent variable was experienced or interpreted by the subject in the way intended. (chapter 7)

manuscript census  the original schedules on which census enumerators record their data, which are released to the public after a period of 72 years. (chapter 12)

marginal frequencies  row and column totals in a contingency table (cross-tabulation) that represent the univariate frequency distributions for the row and column variables. (chapter 15)

matching  a technique for assigning subjects to experimental groups so that the composition of each group matches that of the others on one or more characteristics thought to be related to the dependent variable; should be used in conjunction with, but not as a substitute for, randomization. (chapters 7, 14)

maturation  a threat to internal validity; any psychological or physiological changes taking place within subjects that occur over time, regardless of experimental manipulations. (chapter 8)

mean  a measure of central tendency that indicates the average value of a univariate distribution of interval- or ratio-scale data; calculated by adding up the individual values and dividing by the total number of cases. (chapter 15)

measures of association  descriptive statistics used to measure the strength and direction of an observed bivariate relationship. (chapters 4, 15)

median  a measure of central tendency indicating the midpoint in a univariate distribution of interval- or ratio-scale data; indicates the point below and above which 50 percent of the values fall. (chapter 15)

membership role  the participant role adopted by the researcher in participant observation studies; see active, complete, and peripheral membership roles. (chapter 11)

memos  an adjunct to field notes in field research that consists of brief, recorded analyses that come to mind in going over notes and observations. (chapter 11)

meta-analysis  systematic procedures for synthesizing and summarizing the results from previous, related studies. (chapter 13)

methodological empathy  a fundamental approach to field research that attempts to understand behavior as it is perceived and interpreted by those under study. (chapter 11)

mixed-mode survey  a survey which uses more than one mode (questionnaires, personal interviews, telephone interviews) to sample and/or collect the data. (chapter 9)

modal instance sampling  a sampling procedure recommended in evaluation research in which sites for program assessment are selected for their similarity to the setting in which the program will be implemented. (chapter 14)

mode  a measure of central tendency representing the value or category of a frequency distribution having the highest frequency; the most typical value. (chapter 15)

modeling  in statistical analysis, formal representations of hypothesized, or theoretical, relations among two or more variables. Usually alternative models are compared to determine their fit to the observed data. (chapter 16)

multicollinearity  a problem that arises in multiple regression when combinations of two or more independent variables are highly correlated with each other and that renders regression results (estimates of the coefficients) difficult to interpret. (chapter 16)

multiple regression  a statistical method for studying the simultaneous effects of several independent variables on a dependent variable. (chapter 16)

multiple time-series design  a quasi-experimental design in which a series of pretreatment and posttreatment observations (measurements) are made on a treatment group as well as on nonequivalent control groups. (chapters 8, 14)

multistage cluster sampling  a sampling procedure in which sampling occurs at two or more steps or stages (e.g., a sample of school districts, then a sample of schools from the selected school districts, and then a sample of pupils from the selected schools). (chapter 6)

multivariate analysis  statistical analysis of the simultaneous relationships among three or more variables. (chapter 16)

mundane realism  when the events in an experiment are similar to everyday experiences. (chapter 7)

mutually exclusive  the measurement requirement that each case can be placed in only one category of a a variable; see also exhaustive. (chapter 5)

back to top

N   an abbreviation representing the number of observations on which a statistic is based (e.g., N =279). (chapter 15)

needs assessment  the systematic assessment of the needs and problems of a community or agency and the ability of a program to meet appropriate needs. (chapter 14)

negative (inverse) relationship  a relationship in which an increase in the value of one variable is accompanied by a decrease in the value of the other; that is, changes in one variable are opposite in direction to changes in the other. (chapter 4)

network sampling  a sampling procedure in which respondents initially contacted in screening a probability sample are asked to identify other members of the target population who are socially linked to the respondent (e.g., relative, neighbor, co-worker). (chapter 6)

nominal measurement  the lowest level of measurement, in which numbers serve only to label category membership; categories are not ranked but must be exhaustive and mutually exclusive. (chapter 5)

noncausal association  when two variables are allowed to be statistically associated in a model, but a causal explanation of the association is not part of the model. (chapter 16)

nonequivalent control group design  a quasi-experimental design that resembles a true experiment (the pretest-posttest control group design), except that random assignment of subjects to treatment and control groups is lacking. (chapter 8)

nonparticipant observation  an approach to field research in which the researcher attempts to observe people without interacting with them and, typically, without their knowing that they are being observed. (chapter 11)

nonprobability sampling  processes of case selection other than random selection. (chapter 6)

nonreactive measurement  any process of measurement that by itself does not bring about changes in what is being measured. In contrast, reactive measures may produce changes in behavior because of people’s awareness that they are being studied or observed. (chapter 12)

nonresponse bias  in survey sampling, when nonrespondents (sampled individuals who subsequently do not respond or cannot be contacted) differ in important ways from respondents. (chapters 6, 9)

nonspuriousness  a criterion for inferring causality that requires that an association or correlation between two variables cannot be explained away by the action of extraneous variables. (chapter 4)

normal distribution  a bell-shaped distribution of data that characterizes many variables and statistics, such as the sampling distribution of the mean. (chapter 15)

null hypothesis  the hypothesis that an observed relationship is due to chance—the opposite of the research hypothesis—which is associated with tests of statistical significance; a test that is significant rejects the null hypothesis at a specified level of probability. (chapter 4)

back to top

objectivity  in science, the methological condition that makes it possible for two or more people to agree on the results of an observation; see also intersubjective testability. (chapter 2)

one-group pretest-posttest design  a pre-experimental design in which a group of subjects is observed or measured (the pretest), a treatment is introduced, and the subjects are measured again (the posttest); threats to internal validity include maturation, history, testing, instrumentation, and sometimes statistical regression. (chapter 8)

one-shot case study  a pre-experimental design in which a treatment is administered to a group, after which the group is observed or tested to determine the treatment effects; threats to internal validity include attrition, maturation, and history. (chapter 8)

open-ended questions  survey questions that require respondents to answer in their own words; also called “free-response questions.” (chapter 10)

operationalization  the detailed description of the research operations or procedures necessary to assign units of analysis to the categories of a variable in order to represent conceptual properties. (chapter 5)

ordinal measurement  a level of measurement in which different numbers indicate the rank order of cases on some variable. (chapter 5)

outliers  unusual or suspicious values that are far removed from the preponderance of observations for a variable. (chapters 15, 16)

back to top

panel study  a longitudinal design in which the same individuals are surveyed more than once, permitting the study of individual and group change. (chapters 9)

paper-and-pencil questionnaire  a survey form filled out by respondents. When an interviewer records survey responses, the form usually is called an “interview schedule.” (chapter 9)

parameter  a characteristic of a population, such as the percentage of women or average age. (chapter 6)

partial-regression coefficients/partial slopes  coefficients in a multiple-regression equation that estimate the effects of each independent variable on the dependent variable when all other variables in the equation are held constant. (chapter 16)

partial tables  a control procedure used in elaboration that involves recomputing the original two-variable relationship separately for each category of the control variable. Each partial table displays the association between the two original variables when the control variable is held constant. (chapter 16)

participant observation  an approach to field research in which the researcher actively participates, for an extended period of time, in the daily lives of the people and situations under study. (chapter 11)

path analysis  a form of causal modeling utilizing standardized regression coefficients that provides, among other things, quantitative estimates of the total direct and indirect effects of one variable on another. (chapter 16)

percentage distribution  a norming operation that facilitates interpreting and comparing frequency distributions by transforming each to a common yardstick of 100 units (percentage points) in length; the number of cases in each category is divided by the total and multiplied by 100. (chapter 15)

percentage polygon  see frequency polygon. (chapter 15)

peripheral membership role  a type of participant role in field research in which the researcher is only marginally involved in the setting or group under study; see also active and complete membership roles. (chapter 11)

population  the total membership of a defined class of people, objects, or events; also called “universe.” (chapter 6)

position response set  a tendency of some respondents to mark options located in a certain position, such as the first choice in a series. (chapter 10)

positive (direct) relationship  a relationship in which an increase in the value of one variable is accompanied by an increase in the value of the other, or a decrease in one is accompanied by a decrease in the other; that is, the two variables consistently change in the same direction. (chapter 4)

posttest-only control group design  the simplest of the true experimental designs, which incorporates these features: random assignment to treatment and control groups, introduction of the independent variable to the treatment group, and a posttreatment measure of both groups. (chapter 8)

pre-experimental designs  designs that lack one or more features of true experiments, such as a comparison group or random assignment. (chapter 8)

pretesting  a trial run of an experiment or survey instrument with a small number of preliminary subjects or respondents to evaluate and rehearse the study procedures and personnel; see also field pretesting. (chapters 7, 9, 10)

pretest-posttest control group design  a true experimental design in which subjects are randomly assigned to a treatment group measured before (the pretest) and after (the posttest) the experimental treatment and a no-treatment control group measured at the same times. (chapter 8)

primary sampling unit  the sampling units in the first stage of a multistage sample. (chapter 6)

primary source  a historical document, such as a diary or autobiography, that contains the testimony of an eyewitness; see also secondary source. (chapter 12)

probability proportionate to size sampling  the selection of cases in cluster sampling so that the probability of selection is proportionate to the size of (i.e., the number of cases in) the cluster. (chapter 6)

probability sampling  sampling based on a process of random selection that gives each case in the population an equal or known chance of being included in the sample; see random selection. (chapter 6)

program monitoring  a type of evaluation research undertaken to determine whether a program is being carried out as designed and is reaching its intended target population; see evaluation research. (chapter 14)

Public Use Microdata Sample  a computer-based sample of individual census returns with certain information excluded to ensure confidentiality. (chapter 12)

purposive sampling  a form of nonprobability sampling that involves the careful selection of typical cases or of cases that represent relevant dimensions of the population. (chapter 6)

back to top

qualitative variable  a variable that has discrete categories, usually designated by words or labels, and nonnumerical differences between categories (i.e., nominal level of measurement); see also quantitative variable. (chapter 4)

quantitative variable  a variable that has categories that express numerical distinctions (ratio, interval, and ordinal levels of measurement); see also qualitative variable. (chapter 4)

quasi-experimental designs  designs that lack some features (usually randomization) of true experiments, but permit stronger inferences about cause and effect than do pre–experimental designs, by means of special design features and supplementary data testing. (chapters 8)

quota sampling  a form of nonprobability sampling that involves the allocation of quotas of cases for various strata (usually proportionate to representation in the population) and the nonrandom selection of cases to fill the quotas. (chapter 6)

back to top

R ²   a measure of fit in multiple regression that indicates approximately the proportion of the variance in the dependent variable (spread of observations about the mean) predicted or “explained” by the independent variables. (chapter 16)

random assignment  the assignment of subjects to experimental conditions by means of a random device, such as a coin toss or use of a table of random numbers, thus ensuring that each subject has an equal chance of being in any of the treatment and control conditions; also called “randomization.” (chapters 7, 14)

random-digit dialing  a sampling-frame technique for resolving the problem in telephone surveys of missing those with unlisted numbers: dialable telephone numbers are generated (sampled) from a table of random numbers or from a computer random-number program. (chapter 9)

random-effects model  a statistical framework that assumes, in meta-analysis, that results may be shaped by between-study differences in methods (inherent biases), conditions and settings (external validity), as well as sampling (size and design); see also fixed-effects model. (chapter 13)

random measurement error  an error unrelated to the concept being measured that is the result of temporary, chance factors. Random errors are inconsistent across measurements (unpredictably varying in extent and direction) and affect reliability; see also systematic measurement error. (chapter 5)

random selection  a process that gives each case in the sampling frame an equal chance of being included in the sample. (chapter 6)

range  a univariate measure of variability or dispersion indicating the difference between the lowest and highest values, which is usually reported by identifying these two extreme values. (chapter 15)

ratio measurement  the highest level of measurement, which has the features of the other levels plus an absolute (nonarbitrary) zero point; consequently, it is possible to form ratios of the numbers assigned to categories. (chapter 5)

reactive measurement effect  an effect whereby the process of measurement itself, due to people’s awareness of being studied, produces changes in what is being measured. (chapters 5, 7)

reason analysis  the development of an accounting scheme outlining the categories of reasons for decisions, which serves as a model for formulating a series of questions in survey instrumentation. (chapter 10)

reciprocal causation  when two variables are hypothesized to be causally linked in terms of their mutual influence on each other; also called “mutual causation.” (chapter 16)

recording units  units of analysis in content analysis, such as words, sentences, paragraphs, or plots. (chapter 12)

referral sampling  see network sampling and snowball sampling. (chapter 6)

regression analysis  a statistical method for studying bivariate (simple-regression) and multivariate (multiple-regression) relationships among interval- or ratio-scale variables. (chapters 15, 16)

regression line  a geometric representation of a bivariate regression equation that provides the best linear fit to the observed data by virtue of minimizing the sum of the squared deviations from the line; also called the “least squares line.” (chapter 15)

reliability  the stability or consistency of an operational definition. (chapter 5)

replication  (1) a repetition of a previous study, using a different sample of cases and often different settings and methods, for the purpose of exploring the possibility that the original findings were an artifact of particular research conditions and procedures; (2) an outcome in elaboration analysis in which the original relationship is essentially repeated in each partial table. (chapters 7, 13, 16)

research design the overall plan of an empirical study including the basic approach, sampling design, and measurement of key variables. (chapter 4)

research ethics  see ethics.

research misconduct  as defined by federal policy, unethical actions in proposing, conducting, or reporting research involving fabrication, falsification, or plagiarism. (chapter 3)

residuals  the difference between observed values of the dependent variable and those predicted by a regression equation. (chapters 15, 16)

respondent debriefings  a field pretest technique in which structured follow-up questions at the end of pretest interviews are used to identify instrument problems from the respondent’s perspective. (chapter 10)

response analysis  a field pretest technique in which the responses of pretest respondents are tabulated and examined for problematic response patterns. (chapter 10)

response bias tendency  a tendency of a respondent to answer in a certain biased direction (such as in the direction of social desirability) as a function of the content or form of survey questions. (chapter 10)

response effect  in survey research, a general term for systematic errors due to such factors as biased or confusing questions, response bias tendencies (e.g., social desirability effects), the effects of interviewer’s physical characteristics, and so on. (chapter 9)

response format the form of the response categories in closed-ended questions. (chapter 10)

response rate  in a survey, the proportion of people in the sample from whom completed interviews or questionnaires are obtained. (chapter 6, 9)

back to top

sample  a subset of cases selected from a population. (chapter 6)

sample bias  systematic error or bias in sample results due to problems in executing the sampling plan, such as incomplete sampling frames and incomplete data collection (not-at-home respondents, refusals, etc.). (chapter 6)

sampling design  that part of the research plan that specifies how as well as how many cases are to be selected. (chapter 6)

sampling distribution  a theoretical distribution of sample results (means, proportions, etc.) that would result from drawing all possible samples of a fixed size from a particular population. (chapter 6)

sampling error  the difference between an actual population value (e.g., a mean) and the population value estimated from a sample. (chapter 6)

sampling fraction  the proportion of the population included in the sample. (chapter 6)

sampling frame  an operational definition of the population that provides the basis for drawing a sample; a sampling frame is constructed by either (1) listing all cases from which a sample may be selected or (2) defining population membership by a rule that provides a basis for case selection. (chapter 6)

sampling interval  the ratio of the number of cases in the population to the desired sample size, which is used to select every Kth (the interval) case in systematic sampling. (chapter 6)

scale  a composite measure of a concept constructed by combining separate indicators according to procedures designed to ensure unidimensionality or other desirable qualities. (chapters 5, 13)

scatterplot  a graph plotting the values of two variables for each observation. (chapter 15)

secondary analysis  analysis of survey or other data originally collected by another researcher, ordinarily for a different purpose. (chapters 9, 12)

secondary source  a document, such as a written history, that contains indirect evidence of past events; that is, an account of anyone who was not an eyewitness; see also primary source. (chapter 12)

selection  a threat to internal validity; systematic differences in the composition of control and experimental groups. (chapter 8)

selective deposit  systematic biases in the content of available historical data due to actions such as selective destruction or editing of written records. (chapter 12)

selective survival  incompleteness of available historical data due to the fact that some objects survive longer than others. (chapter 12)

semistructured interview  a type of interview that, while having specific objectives, permits the interviewer some freedom in meeting them; see also structured interview and unstructured interview. (chapter 9)

separate-sample pretest-posttest design  a quasi-experimental design having two groups that receive the treatment, one group randomly selected for pretreatment measurement and the other randomly selected for posttreatment measurement. (chapter 8)

serendipity pattern  unanticipated findings that cannot be interpreted meaningfully in terms of prevailing theories and that give rise to new theories. (chapter 2)

simple random sampling  a probability sampling procedure in which every possible combination of cases has an equal chance of being included in the sample. (chapter 6)

slope/regression coefficient  a bivariate regression statistic indicating how much the dependent variable increases (or decreases) for every unit change in the independent variable; the slope of a regression line. (chapter 15)

snowball sampling  a sampling procedure that uses a process of chain referral, whereby each contact is asked to identify additional members of the target population, who are asked to name others, and so on. (chapters 6, 11)

social bookkeeping  documents, such as court records and vital statistics, that contain recorded information produced by groups or organizations. (chapter 12)

social desirability bias/effect  a tendency of some respondents to bias their answers in the direction of socially desirable traits or attitudes, thereby endeavoring to enhance their self-esteem or make a favorable impression on the interviewer or researcher. (chapters 5, 10)

social impact assessment  the systematic assessment of the consequences of a policy or project on individuals, groups, and communities; see also environmental impact assessment. (chapter 14)

social indicators  broad measures of important social conditions, such as measures of crime, family life, health, schools, or job satisfaction, applied periodically to track trends. (chapter 14)

social network design  a survey design in which data are collected on relationships or connections among social actors (people, organizations, countries, etc.) and the transaction flows (processes) occurring along the connecting links. (chapter 9)

Solomon four-group design  a true experimental design requiring four groups: a treatment and a control group that are pretested as well as a treatment and a control group that are not pretested. This design provides information regarding the effect of the treatment, the effect of pretesting alone, the possible interaction of pretesting and treatment, and the effectiveness of the randomization procedure. (chapter 8)

specification  (1) an outcome in elaboration analysis involving the increase, reduction, or disappearance of the original relationship in some partial tables but not in others (i.e., an interaction effect); (2) the formulation of a formal theoretical model. (chapter 16)

specification error  the fitting of a false model to data, such as omission of an important variable, which may produce misleading results. (chapter 16)

split-ballot tests  experimental manipulations of question ordering, wording, or formats, used occasionally in field pretests to compare instrument versions. (chapter 10)

split-half reliability  an “equivalence” method for assessing reliability in which scores on half of the items of a scale or index are correlated with scores on the other half. (chapter 5)

spurious relationship  a statistical association between two variables produced by extraneous variables rather than by a causal link between the original variables. (chapter 4)

stakeholders  of particular concern to evaluation researchers, the individuals or groups affected by the implementation and outcome of a social program; see evaluation research. (chapter 14)

standard deviation  a univariate measure of variability or dispersion that indicates the average “spread” of observations about the mean; the square root of the variance, which is calculated by subtracting each value from the mean and squaring the result, then taking the arithmetic average of the squared differences. (chapters 6, 15)

standard error  a statistical measure of the “average” sampling error for a particular sampling distribution; the standard deviation of a sampling distribution and thus a measure of how much sample results will vary from sample to sample. (chapter 6)

standardized regression coefficients  coefficients obtained from a norming operation that puts the various partial-regression coefficients on common footing by standardizing them to the same metric of standard deviation units; consequently, indicates the number of standard deviation units the dependent variable changes for every unit change of 1 standard deviation in an independent variable. (chapter 16)

static group comparison  a pre-experimental design in which a treatment group and a no-treatment group are both measured following the treatment; threats to internal validity include selection, differential attrition, and sometimes maturation. (chapter 8)

statistic  a characteristic of a sample; statistics estimate population parameters; see also parameter. (chapter 6)

statistical regression  a threat to internal validity; the tendency for extreme scorers on one measurement to move (regress) closer to the mean score on a later measurement; also known as “regression toward the mean.” (chapter 8)

statistical significance  see test of statistical significance.

stochastic component  the variability in a statistical model that is attributable to random or chance factors; see also systematic component. (see chapter 16)

stratified random sampling  a probability sampling procedure in which the population is divided into strata and independent random samples are drawn from each stratum. (chapter 6)

stratum the variable categories (e.g., male and female) into which a population is divided in selecting a stratified random sample; see also stratified random sampling. (chapter 6)

structural equation modeling  a statistical technique for testing different theorized models, including “structures” of relationships among observed indicators and their underlying unobserved concepts. (chapter 13)

structured interview  a type of standardized interview in which the objectives are highly specific, all questions are written beforehand and asked in the same order for all respondents, and the interviewer’s remarks are standardized; the preferred interviewing approach when the research purpose is to test hypotheses; see also unstructured interview. (chapter 9)

structured observation  a type of direct observation with explicit and preset procedures for deciding when, where, how, and what to observe; see also unstructured observation. (chapter 11)

subjective validation  validity assessment based on the researcher’s judgments; see face validity and content validity. (chapter 5)

summative evaluation  a type of evaluation research that assesses the overall outcome of a social program in terms of meeting program goals and/or efficiency; see effect assessment and efficiency assessment. (chapter 14)

suppressor variable in elaboration analysis, a test variable that weakens or conceals a relationship so that when it is uncontrolled there may be little or no zero-order association. (chapter 16)

survey  a major approach to social research that involves asking a relatively large sample of people direct questions through interviews or questionnaires. (chapters 1, 9, 13)

systematic component  the variability in a statistical model that is attributable to hypothesized relations among variables; see also stochastic component. (chapter 16)

systematic measurement error  error from factors that systematically influence (bias) either the process of measurement or the concept being measured. Systematic errors are consistent across measurements taken at different times or are systematically related to characteristics of the cases being measured (e.g., the cultural bias of IQ tests) and thereby affect validity; see also random measurement error. (chapter 5)

systematic sampling  a probability sampling procedure in which cases are selected from an available list at a fixed interval after a random start. (chapter 6)

back to top

target population  in sampling, the population to which the researcher would like to generalize his or her results. (chapter 6)

telephone interview  a type of interview in which interviewers interact with respondents by telephone; see also face-to-face interview. (chapter 9)

testimony  documents, such as depositions and private letters, containing individual eyewitness accounts of events. (chapter 12)

testing  a threat to internal validity; changes in what is being measured that are brought about by reactions to the process of measurement; the effects of being measured once on being measured a second time. (chapter 8)

test of statistical significance  a statistical procedure used to assess the likelihood that the results of an experiment or other study could have occurred by chance. (chapters 4, 7, 15, 16)

test-retest reliability  a “stability” method for assessing reliability that involves calculating the correlation between outcomes of repeated applications of a measure. (chapter 5)

theoretical sampling  a field-research sampling strategy in which broad analytical categories are sampled in order to facilitate the development of theoretical insights. (chapter 11)

theory a logically interconnected set of propositions that show how or why a relationship occurs. (chapter 2)

treatment  the manipulated conditions comprising an independent variable in an experiment. (chapter 7)

treatment diffusion of information  a threat to internal validity, most problematic in evaluation studies, in which subjects’ actions are affected by their knowledge of conditions in other treatment conditions. (chapter 14)

trend study  a longitudinal design in which a research question is investigated by repeated surveys of independently selected samples of the same population. (chapter 9)

triangulation  the addressing of a social research question with multiple methods or measures that do not share the same methodological weaknesses; if different approaches produce similar findings, confidence in the results increases. (chapter 13)

true experimental designs  an experimental design that affords strong evidence of cause and effect by means of meeting five requirements: (1) manipulation of the independent variable, (2) measurement of the dependent variable, (3) randomization, (4) at least one comparison group, and (5) constancy of conditions. (chapters 7, 8)

t -test in multiple regression a test of significance that tests the hypothesis that a regression coefficient is significantly different from zero when the other independent variables are taken into account. (chapter 16)

back to top

unidimensionality  evidence that a scale or index is measuring only a single dimension of a concept. (chapter 13)

unit of analysis  the entity about whom or which the researcher gathers information; the unit may be people, social roles and relationships, groups, organizations, communities, nations, and social artifacts. (chapter 4)

univariate analysis  statistical analysis of one variable at a time; see also bivariate and multivariate analysis. (chapter 15)

unstructured interview  a type of nonstandardized interview in which the objectives may be very general and the questions developed as the interview proceeds; preferred when the purpose is to acquire preliminary data or understandings; see also structured interview. (chapter 9)

back to top

validity  the congruence or “goodness of fit” between an operational definition and the concept it is purported to measure. (chapter 5)

variance  see standard deviation.

verbal (self-) report  the form of measurement typical of survey research consisting of people’s replies to direct questions. (chapter 5)

back to top

wild-code checking  a data-cleaning procedure involving checking for out-of-range and other “illegal” codes among the values recorded for each variable. (chapter 15)

within-subjects designs experimental designs in which each subject experiences all experimental conditions, such as both treatment and control; see also between-subject designs. (chapter 8)

back to top

Y -intercept the predicted value of the dependent variable in bivariate regression when the independent variable has a value of zero; graphically, the point at which the regression line crosses the Y-axis. (chapter 15)

back to top

zero-order relationship  a bivariate correlation or the relationship in a bivariate table; "zero-order' means that there are no (“zero”) control variables; see also first-order relationship. (chapter 16)

Copyright © 2009 Royce A. Singleton, Jr. and Bruce C. Straits. All rights reserved.
Reproduction in any form or by any means is strictly prohibited without prior written permission.
This site was designed by Katherine Blehar and Kevin Hyland.