Chapter 2 Flashcards
construct
A construct is a concept. It may or may not be directly observable
indicators
observable measures representing the actual construct
operatizing
The process of selecting an indicator(s) of a construct. Indicators should mimic a construct’s real values as closely as possible
operational definition
How an indicator is defined for measurement
multi-dimensional
Concepts that have several distinct subdimensions or subconstructs.
variables
characteristics or measures that can be several possible values
Constructs and indicators in statistical work are variables.
Variables have definitions or descriptions
range
A variable’s set of possible values
single value variable
variables that have only one value
Most research involves single-valued variables
example: someone has one weight, one body temperature, and one sex
multi-valued variable
variables that can have multiple values
example: Undergraduate major is multi-valued as most colleges allow multiple majors; Race
Continuous variables
any value in their range
example: temperature can be 70, 72.4, 31.2 Celsius and so forth depending on how precisely it is measured
Discrete variables
only be specific values in a variable’s range
example: A student’s year in secondary school can range between 1st to 12th grade but can only be an integer (a number without a decimal)
indicator variables (aka dichotomous or dummy variables)
Variables that can only be one of two possible values
Quantitative variables
values that can be ranked and/or their differences calculated
example: grade level and temperature
Qualitative variables
values that can only be categorized. Values can be classified but there is no rank ordering or mathematical difference between the categories
examples: Race, gender, and state of residence
latent variable
abstract constructs which are not directly observable or measurable
level of measurement
There are four levels of measurement. In rank order from lowest amount of information to the highest they are:
- nominal
- ordinal
- interval
- ratio
Nominal variables
can only be classified. They are discrete and qualitative
ordinal variable
value can classified, but can also be rank ordered. the exact differences between values cannot be determined
interval variables
can be classified, ranked, the difference between values can be calculated, and the differences between values are consistent but it does not have a true, absolute zero
Ratio variables
interval variables (can be classified, ranked, the difference between values can be calculated) but have a true, absolute, non-arbitrary zero
composite measure
indicators combined into a single measure for the construct
Indexes
Composite measure that presume indicators are separate components that together form the concept. As the indicators add up to make the concept, index indicators are often simply summed to form the composite measure in an index.
Example: Dow Jones Industrial Average
formative constructs
the indicators are seen as causing the construct (typical for indexes)
Scales
Composite measure that presume indicators are reflections of a concept. Indicators do not combine to form a construct but rather the construct’s value is tapped by their values
reflective constructs
the construct is seen as causing the construct and the indicators reflect its variation. Indicators in scales must covary
Descriptive statistics
provide summary information organizing and describing the data whether a sample or population
Inferential statistics
use sample data to infer the value of an attribute or relationship in a full population. testing whether some measure, feature, or relationship found in a sample can be generalized to the full population
Univariate
Single variable statistic (descriptive and inferential)
bivariate
two variable statistic (descriptive and inferential) Most common technique is a contingency table
multivariate
3 or more variable statistics (descriptive and inferential)
contingency table
a table providing a crosstabulation of the frequencies for two variables simultaneously
independent variables
variables that are seen as affecting or causing others (inferential statistics)
dependent variables
Variables that are being affected
reciprocal relationships
Relationships where two variables affect each other