CH 3, 2, 1

Every variable must have two qualities:

1. Attributes composing a variable must be mutually exclusive. 2. Attributes composing a variable must be exhaustive.

Measurement Types:

1. Direct observables (includingmanifest concepts / variables). Physical characteristics of a person directly in front of an interviewer (sex, weight, height, eye color). 2. Indirect observables(includinglatent concepts / variables). Characteristics of a person as indicated by answers given in a self- administered questionnaire. 3. Constructs (theoreticalcreationsbasedon observations that cannot be observed directly or indirectly). Level of self-esteem, as measured by a scale that combines several direct and/or indirect observables.

Four types of validity are important to consider:

1. Face validity 2. Criterion-relatedvalidity 3. Constructvalidity 4. Content validity

Four levels (or scales) of measurement that define all variables

1. Nominal 2. Ordinal 3. Interval 4. Ratio Measures have greater use in data analysis as they move from the nominal to the ratio level.

Three elements are important to consider regarding measurement quality:

1. Precision and accuracy 2. Reliability 3. Validity

3 kinds of definitions

1. Real definitions 2. Nominaldefinitions 3. Operational definitions

The following methods can be used to ensure one has reliable measures:

1. Test-retest method 2. Split-half method 3. Using established measures 4. Having reliable research workers

Four steps in the construction of an index

1.Item selection 2.Examination of empirical relationships 3.Index scoring 4.Index validation

Rhuearon Scales

A Thurstone scale is a type of composite measure constructed in accord with the weights assigned by "judges" to various indicators of some variables. E.g. one could determine the strength of 5 indicators of "aggression" by having many people rank the indicators from 1 to 5 in terms of which indicator they feel are the best and worst indicators of aggression. Thurstone scaling is rare, because their creation is labor intensive.

Dimensions

A dimension is a specifiable aspect of a concept. (E.g. one's religiosity might be specified in terms of the following dimensions): ▪ Beliefs ▪ Rituals ▪ Devotion ▪ Faith

Validity

A term describing a measure that accurately reflects the concept it is intended to measure. E.g. a measure of "social class" should not measure "religiosity" instead. Validity means that one indeed is measuring what they intend.

Mutually Exhausted

A variable's attributes or values are exhaustive when every case can be classified into one of the variable's categories. Example of a variable whose attributes are NOT exhaustive: ▪ Race ▪ White ▪ Black ▪ Mexican ▪ Native American

Mutually Exclusive

A variable's attributes or values are mutually exclusive if every case can have only one attribute.

Precision and Accuracy

Accuracy regards the degree of truth, correctness, or exactness of a variable's attributes. Precise measures are superior to imprecise ones. Precision is not the same as accuracy.

Index Scoring

After choosing the best items for an index, one assigns scores for particular responses, following this two step process: 1.Determination of the desirable range of the index scores. ▪There is a conflicting desire for a range of measurement in the index and an adequate number of cases at each point in the index, so one must determine the desirable range. Two step process to choose the best items for an index (cont.): 2.Determination of whether to give each item in the index equal or different weights. ▪Standard items should be weighted equally unless there are compelling reasons for differential weighting.

Examination of Emperical Relationships

An empirical relationship is established when respondents' answers to one question help one predict how they will answer other questions. Two types of relationships exist: 1.Bivariate relationships describe relationships between two variables. 2.Multivariate relationships describe relationships between more than two variables.

Index & Scales

An index is a type of composite measure that summarizes and rank-orders several specific observations; it represents general dimensions. A scale is a type of composite measure composed of several items that have a logical or empirical structure among them.

Nominal definitions

Are assigned to a term without any claim that the definition represents a "real" entity. Nominal definitions are arbitrary. Nominal definitions represent a consensus or convention about the meaning of something.

Reliability; SPLIT- half

By using the split-half method, one uses multiple sets of randomly assigned variables in order to produce the same classifications. E.g. the Rosenberg self-esteem scale has 10 items that together measure "self-esteem." If one split the 10 items into two groups of 5, both groups should still represent one's level of self-esteem. If the result for each group is different, the measure of self-esteem would likely be unreliable.

Reliability; Test-retest method

By using the test-retest method, one makes the same measurement more than once. If one measures twice and gets the same result, a measurement is more likely to be reliable. If a second measure reveals different results, the measurement is likely to be unreliable.

Operational Choices

Conceptualization is the refinement and specification of abstract concepts. Operationalization is the development of specific research procedures that will result in empirical observations representing those concepts in the real world.

Indicators

Conceptualization provides definite meaning to a concept by specifying one or more indicators of what one has in mind. An indicator is an observation one chooses to consider as a reflection or representation of a variable they wish to study. Indicators are proxies of something; they do not represent the exact concept/variable they are associated with. (E.g. one may use the number of religious services one attends over a period of time as an indicator of religiosity)

Criterion-relatedvalidity

Criterion-related validity represents the degree to which a measure relates to some external criterion. E.g. the validity of SAT tests is based on their ability to predict college success. College success is the criterion by which the SAT test is determined to be valid.

Generality Vs. Specificity

Depending on the nature of one's desired index, one may select general or specific items. E.g. an index measuring general aspects of religiosity might include ritual participation, belief, etc. An index measuring specific aspects of religiosity (such as ritual participation) might select attendanceat church, confessions, bar mitzvah's etc.

Concepts are constructs

Derived by mutual agreement from mental images. Conceptions summarize collections of seemingly related observations and experiences.

Descriptive & Explanatory Research

Descriptive research requires detail and precision in its definitions. Explanatory research often is less concerned with subtle nuances of a definition, and more with general patterns (so multiple definitions for the same phenomenon might be acceptable).

Reliability- Established Measure

Established measures are measures that others have already proved reliable in previous research. E.g. if one has a unique measure for "prejudice" they can compare their results with established measures of prejudice to be confident their measure is a reliable measure of prejudice.

Face Validity

Face validity represents whether the quality of an indicator makes it a reasonable measure of some variable. Face validity means that a measure "makes sense" on the face. It is the lowest level of validity assurance. E.g. one's voting frequency seems to be a good indicator of community involvement.

Variance

In selecting items, one must address the amount of variance they provide. These two procedures can guarantee variance within an item: 1.Select several items will divide people into equal groups concerning the variable. 2.Select items differing in variance.

Scale Construction

Indexes may fail to take into account that not all indicators of a variable are equally important. Scales offer more assurance of ordinality by tapping the intensity structures among indicators.

The interval scale

Interval scales represent an ordered series of equal- sized categories. Interval measurements identify the direction and magnitude of a difference. The zero point is located arbitrarily on an interval scale. Examples: Fahrenheit temperature scale, IQ scores, dates (i.e. March 12 or April 2).

Measurement

Is the careful, deliberate observation of the real world for the purpose of describing objects and events in terms of the attributes composing a variable. Public research is only as good as the instruments one uses to measure elements of the social word.

Index Construction: Handing Missing Data

Missing data can be problematic. One may do the following five things regarding missing data: 1.If there are few cases with missing data, one may decide to exclude them from the construction of the index and analyses. 2.Treat missing data as one of the available responses. 3. Analyze the missing data to interpret their meaning. 4.Assign missing data the middle value, or the mean value. 5.Assign values to the proportion of variables scored.

The nominal scale

Nominal level variables (also called categorical variables) represent unordered categories identified only by name. Nominal measurements only permit one to determine whether two individuals are the same or different. Examples: religion, race, or countries.

Reliability; Reliability & Research Worker

One can determine the reliability of measurements and results by checking the reliability of research assistants. E.g. multiple coders can be used for the same data. E.g. in a study of 1000 people, a principle investigator could randomly contact 50 of those 1000 and personally interview them. If the P.I.'s research assistants gathered reliable data, answers from those 50 should correspond to the answers from the greater 1000.

The ordinal scale

Ordinal variables represent an ordered set of categories. Ordinal measurements tell one the direction of difference between two individuals. Examples: the alphabet, Likert scales, any scale that measures something according to low, medium, and high.

The ratio scale

Ratio scale measures are interval scales that contain an absolute zero at one point along the spectrum of the scale (i.e. zero indicates none of the variable). Ratio measurements identify the direction and magnitude of differences and allow ratio comparisons of measurements. Examples: income, height, 40 yard dash time.

Reliability

Refers to the quality of a measurement method that suggests the same data would have been collected each time in repeated observations of the same phenomenon. Reliability is not the same as accuracy.

Content validity

Represents the degree to which a measure covers the range of meanings included within a concept. E.g. a measure of mathematical ability does not have content validity if it only includes "addition." By including "addition, subtraction, division, and multiplication" one ensures their measure of mathematical ability is more valid.

Construct validity

Represents the degree to which a measure relates to other variables as expected within a system of theoretical relationships. E.g. the variable marital satisfaction is likely to correlate with the variable marital fidelity. By comparing these variables one can better determine whether one has a valid measure.

Operational definitions

Specify precisely how a concept will be measured, (i.e. what operations will be performed). Operational definitions are nominal, but they attempt to achieve clarity about the meaning of a concept within the context of a study.

Real definitions

Statements about the essential nature of some entity. Real definitions assume a construct is a real entity (which it is not: it is a proxy). Since real definitions are often vague they are not useful for the purpose of rigorous inquiry.

Bogardus Social Distance Scales

The Bogardus social distance scale (BSDS) is a measurement technique for determining the willingness of people to participate in social relations - of varying degrees of closeness - with other kinds of people. Differences in intensity suggest a structure among items. Logically, once a person changes their answer on a BSDS scale, their answers will remain changed (i.e. when one changes from "yes" to "no" they will not answer "yes" to any additional items on the BSDS).

Face Validity

The first criterion for selecting index items is face validity. E.g. if one was to create an index of morality, items such as compassion, justice, honesty, and caring would be logical representations of "morality." These items have face validity: it makes sense that they represent the meanings of morality.

Item Selection

The first step in creating an index is to select items for a composite index. This is created to measure some variable. When selecting items, one must consider the following: a.Face validity b.Unidimensionality c.Generality vs. specificity d.Variance

Types of Scales

The following are common examples of scales used in social research: 1.Bogardus social distance scales 2.Thurstone scales 3.Likert scales 4.Semantic differential scales 5.Guttman scales

Conceptualization

The mental process whereby fuzzy and imprecise notions (concepts) are made more specific and precise. Conceptualization processes a specific, agreed- upon meaning for a concept for the purposes of research. (E.g. the concept "prejudice" invokes feelings toward what it means. How would one measure prejudice?

Conceptual Order

The order of conceiving a research question often is as follows: 1. Conceptualization; What are the different meanings and dimensions of the concept "aggression?") 2. Nominal definition; For our study, we will define aggression as representing physical harm, specifically, how often one hits another. 3. Operational definition; We will measure physical harm via responses to the survey question "How many times have you hit someone in the past year?" 4. Real world measurement; The interviewer will ask, "How many times have you hit someone in the past year?"

Unidimensinality

The second criterion for selecting index items is unidimensionality. A composite measure should only represent one dimension of a concept. E.g. items representing religious fundamentalism should not be included in a measure of political conservatism.

Index Validation

There are two general ways to validate an index: 1.Item analysis ▪An item analysis is an assessment of whether each of the items included in a composite measure makes an independent contribution or merely duplicates the contribution of other items in the measure. Two general ways to validate an index (cont.): 2.External validation ▪External validation is the process of testing the validity of a measure, such as an index or score, by examining its relationship to other presumed indicators of the same variable.