-
methodology
objective systematic rule governed observations (surveys and experiments). Evaluated by criteria of validity, reliability and generalizability.
posi: Use quantitative methods to remain detached from respondents
-
What are concepts and constructs?
- Concept: An idea that represents a class of objects (eg cats) or events
- (walking) or their properties (fast). Construct: a complex idea or
- concept formed from a synthesis of simpler ideas or concepts (concepts
- invented by scientists)
-
Quant and qual variables
Qualitative (or categorical) variables: variables whose values are discrete and designated by words. Quantitative variables: values come from a numerical continuum.
-
Predictor variables
variables that predict or explain other variables.
-
outcome variables
variables to be predicted or explained
-
Antecedent
presumed to casually precede predictor variables e.g. church predicts morality.
-
Intervening
(or mediating variables): variables that channel the effect of the predictor variable on the outcome variable (e.g. maybe church attendance -> bible reading -> morality).
-
Moderators
a variable that by itself may or may not predict and outcome variable- changes the strength or direction of an existing correlation or effect
-
which axis predictor/outcome? what is the moderator
- predictor x axis, outcome y axis
- moderator major
-
linear relationship
the relationship b/w two variables.
-
Spurious
Extraneous third factor and neither variable involved in the correlation influenced the other.
-
How are items related to a scale, index, or composite measure of a variable?
Questions create a scale which is an index of composite measure of the construct, based on self-reports. The questions are items/indicators.
-
Nominal
classifying cases into categories. Tool or activity to measure an object which will be a variable, must be exhaustive and mutually exclusive (e.g. Sex: male/female).
-
Ordinal
classify cases into categories and then rank them (e.g. education level 1= high school, 2 bachelors, etc).
-
Interval
Ordinal measures + equal intervals (e.g. attitude toward marijuana use).
-
Ratio
interval measures + a meaningful (e.g. we know how much more $4 is to $1). Each level adds more info than the previous.
-
Reliability
whether or not we’re measuring something consistently (interchangeable w consistency).
-
Validity
whether or not we are measuring the variable that we aim to measure (accuracy)
-
three types of reliability
test-retest, inter-item and inter-coder
-
test-retest
Given the same group of respondents a test at two time points. And see whether scores at Time 1 correlate with scores at Time 2. Strong correlation indicates high reliability
-
Inter-item reliability (internal consistency)
Whether the items in a scale correlate with each other?- Cronbach’s alpha>= .6 acceptable, >= .7 satisfactory, >=.9 will be excellent.
-
Inter-coder
Whether or not different human judges (coders) produce similar results by using the same procedure and instrument?-Content analysis and nonverbal studies.
|
|