TERMINOLOGY

A
Absolute risk (AR) The proportion of people in a group who experienced an undesirable
outcome.

Absolute risk reduction (ARR) The difference between the absolute risk in one group
( e.g., those exposed to an intervention ) and the absolute risk in another group ( e.g., those
not exposed); sometimes called the risk difference or RD .

Alpha ( α ) 1.In tests of statistical significance, the significance criterion-the risk the
researcher is willing to accept of making a type I error; 2.in measurement, an index of
internal consistency, that is, Cronbach’s alpha.

Analysis The organization and synthesis of data so as to answer research questions and test
hypotheses.

Analysis of covariance (ANCOVA)

A statistical procedure used to test mean group difference on a dependent variable while
controlling for one or more covariate.

Analysis of variance (ANOVA) a statistical procedure for testing mean difference among
three or more groups by contrasting variability between group to variability within group,
yielding an F-ratio statistic.

Analysis triangulation The use of two or more analytic approaches to analyze the same set
of data.

Asymmetric distribution A distribution of data values that is skewed, with two halves that
are not mirror images of each other.

B
Beta ( β ) 1. In multiple regression, the standardized coefficients indicating the relative
weights of the predictor variable in the equation; 2.in statistical testing the probability of a
type II error.
Bimodal distribution A distribution of data values with two peaks (high frequencies).

Binomial distribution A statistical distribution with known properties describing the

number of occurrences of event in a series of observation; forms the basis for analyzing
dichotomous data.

Bivariate statistics statistical analysis of two variable to assess the empirical relationship
between them.

Bland-altman plot A graphic depiction of the degree of agreement between two sets of
scores, for people who have been measured twice on the same continuous measurement scale;
the plot highlights random differences between the two measurements through the
construction of a parameter called the limits of agreement.

Bonferroni correction An adjustment made to establish a more conservative alpha level

when multiple statistical tests are being run from the same data set; the correction is
computed by dividing the desired α by the number of tests—for example, .05/3 = 0.017.

Bricolage The tendency in qualitative research to derive a complex array of data from a
variety of sources, using a variety of methods.

C
Casual modeling The development and statistical testing of an explanatory model of
hypothesized causal relationships among phenomena.

Census A survey covering an entire population.

Central category The main category or pattern of behavior in grounded theory analysis;
sometimes referred to as the core category.

Central limit theorem A statistical principle stipulating that the larger the sample, the more
closely the sampling distribution of the mean will approximate a normal distribution, and that
the mean of a sampling distribution equals the population mean.

Central tendency A statistical index of what is “typical” in set of scores, derived from the
center of the score distribution; indices of central tendency include the mode, median, and
mean.
change score A person’s score difference between two measurements on the same measure,
calculated by subtracting the value at one point in time from the value at the second point.

Chi-square test A statistical test used in various contexts, most often to assess differences in
proportions; symbolized as X2.

Classical test theory (CTT) A measurement theory that has traditionally been used in the
development of multi-item scales; in CTT, any score on a measure is conceptualized as
having a “true score” component and an error component, and the goal is to approximate the
true score.

Coefficient alpha The most widely used index of internal consistency that indicates the
degree to which the items on a multi-item scale are measuring the same underlying construct;
also referred to as Cronbach’s alpha.

Cohen’s d An effect size index for comparing two group means, computed by subtracting
one mean from the other and dividing by the pooled standard deviation; also called
standardized mean difference or SMD.

Confidence interval (CI) The range of values within which a population parameter is
estimated to lie, at a specified probability (e.g., 95% CI)

Contingency table A two-dimensional table in which the frequencies of two categorical

variables are cross-tabulated; also called a crosstabs table.

Correlation An association or bond between variables, with variation in one variable

systematically related to variation in another.

Correlation coefficient An index summarizing the degree of relationship between variables,

typically ranging from +1.00 ( for a perfect positive relationship) through 0.0 ( for no
relationship) to -1.00 (for a perfect negative relationship).

Correlation matrix A two-dimensional display showing the correlation coefficients

between all pairs of variables in set of several variables.

Cost-benefit analysis An economic analysis in which both costs and outcomes of a

program or intervention are expressed in monetary terms and compared.
Cost-effectiveness analysis An economic analysis in which costs of an intervention are
measured in monetary terms, but outcomes are expressed in natural units ( e.g. the costs per
added year of life).

Cost-utility analysis An economic analysis in which costs of an intervention as overall

health improvement and described costs for some additional utility gain—usually in relation
to gains in quality-adjusted life years ( QALY ).

Covariate A variable that is statistically controlled ( held constant ) in ANCOVA, typically

a confounding influence on, or a pre-intervention measure of, the dependent variable.

Covert data collection The collection of information in a study without participants’

knowledge.

Cramer’s V An index describing the magnitude of relationship between nominal-level

data, used when the contingency table to which it is applied is larger than 2 x 2.

Cronbach’s alpha A widely used index that estimates the internal consistency of a
composite measure composed of several subparts;also called coefficient alpha.

Crosstabulation A calculation of frequencies for two variable considered simultaneously—

for example, gender (male/female) crosstabulated with smoking status (smoker/non-smoker)

cutpoint (cutoff point) the point in a distribution of scores used to classify or divine people
into different groups, such as cases and noncases for a disease or health problem (e.g., the
cutpoint for classifying newborns as being low birth weight is 5.5 pounds [2,500])

D
data analysis The systematic organization and synthesis of research data and, in quantitative
studies, the testing of hypotheses using those data.

Data cleaning The preparation of data for analysis by performing check to ensure that the
data are consistent and accurate.

Data collection the gathering of information to address a research problem.

Data entry the process of entering data onto an input medium for computer analysis.

Data set The total collection of data on all variables for all study participants.
Data transformation A step often undertaken before data analysis to put the data in a form
that can be meaningfully analyzed ( e.g., recording of values ).

Data triangulation The use of multiple data sources for the purpose of validating
conclusions.

Degree of freedom (df) A statistical concept referring to the number of sample values free
to vary (e.g., with a given sample mean, all but one value would be free vary ).

De-identified data Data or records from which identifying information is removed to

protect the privacy of individuals.

Delay of treatment design A design for an intervention study that involves putting control
group members on waiting list for the intervention until follow-up data are collected; also
called a wait-list design.

Delphi survey A technique for obtaining judgements from an expert panel about an issue of
concern; experts are questioned individually in several rounds, with a summary of the panel’s
views circulated between rounds, to achieve some consensus.

Descriptive statistics statistics used to describe and summarize data (e.g., means,
percentage ).

Detection bias Systematic difference between groups being compared in how outcome
variables are measured, verified, or recorded; a bias that can result when there is no blinding
of data collectors.

Deviation score A score computed by subtracting an individual score from the mean of all
scores.

Dummy variables dichotomous variables created for use in many multivariate statistical
analyses, typically using codes of 0 and 1 (e.g., female = 1, male = 0).

E
Effect size(ES) In quantitative research, an index summarizing, in standardized units, the
magnitude of change in a group or the amount of difference in two group on a measure; for
mean-difference situations, calculated by dividing the mean difference in two scores by an
index of variability, usually the baseline SD; sometimes referred to as cohen’s d or the
standardized mean difference; in metasynthesis, an index used to characterize the salience of
a theme or category.

Eigenvalue The value equal to the sum of the squared weight for a linear composite, such as
a factor in a factor analysis, indicating how much variance in the solution is accounted for.

Error of measurement The deviation between hypothetical true scores and obtained scores
of a measured characteristic.

Error term The mathematic expression (e.g., in a regression analysis) that represents all
unknown or un-measurable attributes that can affect the dependent variable.

Estimation procedures Statistical procedures that estimate population parameters based on

sample statistics.

Eta squared In ANOVA, a statistic calculated to indicate the proportion of variance in the
dependent variable explained by the independent variables, analogous to R 2 in multiple
regression.

Event history calendar A data collection matrix that plots time on one dimension and
events or activities of interest on the other.

Expectation maximization ( EM) imputation A sophisticated single-imputation process

that generates an estimated value for missing data in two steps (an expectation or E-step and a
maximization or M-step), using maximum likelihood estimation.

Exploratory factor analysis (EFA) A factor analysis undertaken to explore the

underlying dimensionality of a set of variables.

F
F-ratio The statistic obtained in several statistical tests (e.g., ANOVA) in which variation
attributable to different sources (e.g., between-group variation ) is contrasted.

Factor analysis A statistical procedure for disentangling complex interrelationships among

items and identifying the items that “go together” as a unified dimension.
Factor extraction The first phase of a factor analysis, which involves the extraction of as
much variance as possible through the successive creation of linear combinations of the
variables or items in the data set.

Factor loading In factor analysis, the weight associated with a variable or item on a given
factor

Factor matrix In factor analysis of scale items, a matrix with items on one dimension and
factors on the other, with matrix entries being factor loadings of the items on the factors;
factor matrices can be either rotated or un-rotated.

Factor rotation The second phase of factor analysis, during which the reference axes for the
factors are pivoted to more clearly align items variable with a single factor.

Factorial design An experimental design in which two or more independent variable are
simultaneously manipulated, permitting a separate analysis of the main effects of the
independent variables and their interaction.

Fail-safe number In meta-analysis, an estimate of the number of studies with nonsignificant

results that would be needed to reserve the conclusion of a significant effect.

Feasibility study Research completed prior to a main intervention study to test specific
aspects of an emerging intervention or the anticipated trial (e.g., the intervention’s
acceptability)

Fisher’s exact test A statistical procedure used to test the significance of differences in
proportions, used when the sample size is small or cells in the contingency table have no
observations.

Frequency distribution A systematic array of numeric values from the lowest to the
highest, together with a count of the number of times each value was obtained.

Frequency effect size In a qualitative metasummary, the percentage of reports that contain a
given thematic finding.

Frequency polygon Graphic display of a frequency distribution, in which dots connected by

a straight line indicate the number of time score values occur in a data set.
Friedman test A nonparametric analog of ANOVA, used with paired-groups or repeated
measures situations.

Funnel plot A graphical display that plots a measure of study precision (e.g., sample size)
against effect size to explore the possibility of publication bias.

G
General linear model (GLM) A large class of statistical techniques (including regression
analysis, ANOVA, and correlational analysis) that describe the relationship between a
dependent variable and one or more independent variables.

Global rating scale (GRS) A single item designed to provide a summary measurement of a
person’s status on a construct, or his or her perception of change on a construct over a
specified interval; also referred to as a health transition rating.

H
Hosmer-Lemeshow test A test used in logistic regression to evaluate the degree to which
observed frequencies of predicted probabilities correspond to expected frequencies in an ideal
model over the range of probability values; a good fit is indicated by lack of statistical
significance.

I
Inferential statistics Statistics that permits inferences about whether resultsobserved in a
sample are likely to be reliable , that is, found in the population.

Interquartile range (IQR) A measure of variaboility ,indicating the difference between

between Q3 (the third quartileor 75 the percentile) and Q1 (the first quartile or 25 th
percentile)

Intraclass correlation co efficient The statistical index used to assess the reliability of a
measure
K
Kappa A statistical index of chance corrected agreement or consistency between two
nominal measurement s,often used to assess interratyer or intrarater reliability

Kendall’s tau A correlation coefficient used to indicate the magnitude of a relationship

between ordinal level variables.

L
Least squares estimation A method of statistical estimation in which the solution minimizes
the sums of squares of error terms also called ordinary least squares.

Level of significance The risk of making a type I error in a statistical analysis,with the
criterion established by the researcher beforehand

M
Mann-whitney U test A non parametric statistic used to test the difference between two
independent groups,based on ranked scores.

Mc nemar test A statistical test for comparing differences in proportions when values are
derived from paired groups

Mean A measure of central tendency computed by summing all scores and dividing by the
total number of cases

Median The point in a distribution of scores above and below which 50% of the values fall

Mode A measure of central tendency the value that occurs most frequently in a distribution
of scores multiple comparison procedures

Statistical tests normally applied after an ANOVA indicates statistically significant group
differences that compare different pairs of groups

Multiple regression analysis A statistical procedure for understanding the effects of two or
more independent variables on a dependent variable
Multivariate analysis of variance (MANOVA) A statistical procedure used to test the
significance of differences between the means of two or more groups on two or more
dependent variables, considered simultaneously.

Multivariate statistics. Statistical procedures designed to analyze the relationship among

three or more variables

N
Nonparametric statistics A class of statistical tests that do not involve stringent
assumptions about the distribution of variables

O
One tailed test Astatistical test in which only values in one tail of a distribution are
considered in determining significance sometimes used when the researcher states a
directional hypothesis

P
P value In statistical testing the probability that the obtained results are due to chance alone

Parametric statistics A class of statistical tests that involve assumptions about the
distribution of the variables and the estimation of a parameter

Pearson’s r A corerelation coeffienct designating the magnitude of relationship between two

variables measured on at least an interval scale also called the product moment correlation

Phi coefficient A statiscical index describing the magnitude of relationship between two
dichotomous variables point estimation

A statistical procedure in which information from a sample is used to estimate the single
value that best represents the population parameter

Product moment correlation coefficient A correlation coeffienct designating the

magnitude of relationship between two variables measured on at least an interval scale also
called pearson’s r
Q
Quasi statistics An accounting system used to assess the validity of conclusions derived
from qualitative analysis

R
Reliability coefficient A quantitative index usually ranging in value from 0.00 to 1.00,that
provides an estimate of how reliable an instrument is (e.g the intraclass correlation
coefficient)

S
Spearman’s rank order correlation A correlation coefficient indicating the magnitude of a
relationship between variables measured on the ordinal scale

Standard deviation The most frequently used statistic for measuring the degree of variability
in a set of scores

Statistic An estimate of a parameter calculated from sample data

Statistical analysis The organization and analysis of quantitative data using statistical
procedures including both descriptive and inferential statistics. conclusion validity

The degree to which inferences about relationship from a statistical analysis of the data are
correct

Statistical inference An inference about the population based on information from a sample
using laws of probability

Statistical test An analytic tool that estimates the probability that results obtained from
asample reflect true population values.

Structural equations modeling A statistical modeling procedures that involves equations

representing the magnitude of hypothesized relations among sets of variables typically used
to test a model or theory in a path analysis and most often relying on maximum likelihood
estimation.
Survival analysis A statistical procedure used when the dependent variable represents a time
interval between an initial event and an end event

T
T test A prametric statistical test for analyzing the difference between two means

Two tailed tests Statistical tests in which both ends of the sampling distribution are used to
determine improbable values

Type 1 error an error created by rejecting the null hypothesis when it is true

Type 11 error an error created by accepting the null hypothesis when it is false

U
univariate statistics statistical analysis of a single variable for purposes of description

W
wald statistic a statistic distributed as a chisquare used to evaluate the significance of
individual predictors in a logistic in regression equation

wilcoxon signed ranks test a nonparametric statistical test for comparing two paired
groups,based on the relative ranking of values between the pairs

Z
z score A standard score expressed in terms of standard deviation from the mean