Archives   Calendar    Coordinating Center    Multisite Activities    
Project Description  COSP References  Study Sites    Search Our Site   
Table of Contents    Upcoming Meeting Agendas

______________________________________________

Research Glossary:

SAMHSA - Variance

SAMHSA

– See “Substance Abuse and Mental Health Services Administration.” 

Sample

– A part of a larger group of people.  The sample may or may not be chosen by chance.  This term can also be a verb, meaning to choose this smaller group. 

Sample frame

– The methods for choosing the group of people to be studied; or the larger group from which that group is chosen. 

Scaling

– Giving numbers, in order, to information which was in words or ideas; for example, showing a person’s opinion by a number from this list: 1) strongly agree; 2) agree; 3) disagree; 4) strongly disagree.  Scaling always uses numbers. 

Scale

– A test; a group of linked questions that can be added together to form a measure of one thing. 

Scale respondent

– The person filling out the survey or taking the test; for example, a parent, spouse, relative, or teacher of the participant. 

Secondary analysis

– A way of studying information that has been found or written about by someone else. 

Self-administered questionnaire (self-report)

– A set of written questions which the person being studied fills out and returns to the researcher. 

Self-selection

– A way of choosing the people for a study by letting them set themselves apart from a larger group in some way; for example, by responding to a questionnaire or by going to a program. 

Sensitivity

  A measure of how well a scale shows differences among people. 

SEP

– See “socioeconomic status.” 

Services research

– The study of places or groups, like a mental health center, that offer services to people.  The research usually focuses on how well the services work. 

Service utilization

– The use (or rate of use) of health services or social services. 

SES

– See “socioeconomic status.” 

Snowball sampling

– A way of finding people to study by asking them about each other; for example, choosing one person who was born in Germany, then asking that person for the name of a second person born in Germany, and so on, in a chain. 

Socioeconomic status (SES)

– Also called “socioeconomic position” (SEP).  A measure that combines a person’s education, work history, income, etc. into a single rating that tries to show where that person is placed in his or her society, and what larger group (for example, the “middle class”) that person is part of. 

Special populations

– Groups of people that can’t be studied in the same way and by the same rules as other groups, for some reason. 

Specificity

– The measure of how well a scale shows whether a certain person is a member of a certain group; for example, how well it shows whether a client has a phobia of books, rather than showing only whether that client has a high level of general anxiety. 

Split-half reliability

– A measure of how well the different parts of a scale are working together; found by comparing half the items with the other half (for example, the odd-numbered items with the even-numbered items). 

Spurious correlation

– What looks like a link between two things, when the “link” is really caused by a third thing.  For example, doctors often own houses, but that is not just because doctors are doctors; it is also because doctors earn a lot of money. 

SSD

– Social Security Disability.  Monthly payments from the U.S. government to disabled citizens who have not worked and have not paid into the Social Security system. 

SSI

– Social Security Insurance.  Monthly payments from the U.S. government to disabled citizens who have worked in the past and have paid into the Social Security system. 

Stability

– A kind of “reliability”; it shows how alike measurements are at different times with the same test or scale. 

Stakeholders

– People who have a share or an interest in something; for example, people who receive some of the profits of a hospital because they have helped to set up the hospital or have given money to it in the past.  Stakeholders can be clients, relatives, professionals, community leaders, agency administrators, volunteers, etc. 

Standard deviation

– A measure of how widely the values (amounts, numbers, scores) in a group of values are spread around the “mean” (midpoint).  For example, all the scores may be very close to the midpoint, or many of them may be much higher or lower. 

Standardized

– A term that describes a way of giving, scoring, or reading tests or surveys; if a test is standardized, it is like other tests and the information taken from all of them can be compared. 

Statistics

– The study (usually mathematical analysis) of information that is in the form of numbers or can be given numbers; the term can also mean that information itself.  

Significance

– A mathematical test of whether a study’s results could be caused by chance or whether they really show what they seem to show. 

Stratification

– A way of ordering individual people within a social system.  The different rungs of the ladder depend on, for example, income, education, work, or power.  The term can also mean ranking anything on different levels, by group or category. 

Stratified sample

– A group of people or things chosen so that certain levels or smaller groupings (like certain ages, incomes, or diagnoses) are kept together and a set number of people or things from each level are in the larger group.  A stratified “random sample” chooses the people or things within each grouping or level by chance, but decides ahead of time how many people or things there will be within each level and what the levels will be. 

Structured interviews

– Interviews that use a set list of questions the interviewer asks every person.  The interviewer writes down each person’s answers on the form with the questions. 

Substance abuse

– Addiction to alcohol or drugs.  Sometimes this term is separated from “chemical dependence,” which is less severe. 

Substance Abuse and Mental Health Services Administration (SAMHSA)  -- 

-An organization of the federal government, within U.S. Health and Human Services, which focuses on substance abuse (issues related to drug or alcohol dependence) and on mental health. 

Survey research

– A type of study that uses phone questions, mailed questions, interviews, or self-completed forms, and that does not use the “experimental method.”   

Back To Top

TA

– See “technical assistance.” 

T-test: 

-A statistical test of the difference between two “means.” 

Technical assistance

– Manuals, instructions, consultations, etc. that give information or support for learning any technical task (such as how to use email or the internet). 

Trend

- A steady change in one direction over time; for example, more and more parents letting their children have later and later bedtimes over several years would be a trend. 

Test-retest method

– A way of measuring something by first giving a test and then, after waiting a set time, giving the same test again to the same group of people. 

Theory

– A way of explaining or trying to explain a set of facts.   

Time series design

– A way of studying what researchers have noticed at set times; for example, studying how many cavities a group of children have every 6 months. 

Tools

– Ways of testing or measuring; for example, questionnaires, rating scales, etc. 

Type I error

– A mistake based on saying there is a difference when there is not. 

Type II error

– A mistake based on saying there isn’t a difference when there is. 

Typology

– A system that groups information into different types. 

Back To Top

Unit of analysis

– What size or number is being counted as separate within a larger group in a study; for example, an individual person, a family, a city, or a school.  This unit will be different for different studies. 

Univariate analysis

– The study of only one thing that might change, not a group of different things that might each change. 

Validity

– The measure of how well a scale or test shows what it’s supposed to show.  There are several different types of validity and each type must be tested separately: 

Construct validity

– The measure of how well the test fits the ideas behind the study and the way the topic has been set out. Usually such a test separates 2 groups that are known to be opposite extremes. 

Content validity – The measure of how fully the whole topic of the study is covered by the test .  For a test to have content validity, every piece of the topic should also be part of the test.

 This is sometimes called “Face validity.” 

Convergent validity

– The measure of how well the test matches up with other tests of the same thing. 

Criterion validity

– The measure of how well the test matches an accepted test (“gold standard”) outside the study.  There are two types of criterion validity: 

(a)    Concurrent validity – The measure of how well the test being studied and the “gold standard” test measure the same thing at the same time.

(b)   Predictive validity  -- The measure of how well the test being studied predicts some practical result that the “gold standard” will find later. 

Value (statistical)

– An amount written in numbers, not in words, pictures, etc. 

Variable

– Anything that can have different values (be different sizes or amounts) at different times; what is being measured in a study. 

Variance

– The measure of how wide the range of values (amounts, sizes, or scores written in numbers) is; of how far apart these numbers are.  It is a number found by multiplying the “standard deviation” by itself.  

Back To Top