1. Early
1.1. Any Field Context
1.1.1. Operationalization
1.1.1.1. Untitled
1.1.1.1.1. Untitled
1.1.1.2. Untitled
1.1.1.2.1. Untitled
1.1.1.3. How to do it
1.1.1.3.1. Unit of Analysis
1.1.1.3.2. Level of Measurement
1.1.1.3.3. Dimensionality
1.1.1.4. Definition
1.1.1.4.1. An exact set of instructions on what you need to do and how to measure a variable that has been conceptually defined
1.1.2. Validity Testing
1.1.2.1. Face Validity
1.1.2.1.1. Based on consensus, until a valid instrument is proven otherwise
1.1.2.1.2. "It seems reasonable"
1.1.2.2. Content Validity
1.1.2.2.1. Instrument has appropriate content to capture a complex concept
1.1.2.2.2. Does the measure reflect the construct?
1.1.2.2.3. Logical validity
1.1.2.2.4. Are all of the relevant domains of knowledge covered on a test?
1.1.2.2.5. Are all symptoms of a syndrome present?
1.1.2.3. Construct Validity
1.1.2.3.1. Close fit between construct measured and actual observations made with the instrument
1.1.2.3.2. Can you make inferences about the construct based on your measurement?
1.1.2.3.3. Does it measure what you think it does?
1.1.2.4. Criterion Validity
1.1.2.4.1. Close fit between the measure it produces and other instruments already established as valid
1.1.2.4.2. "Gold Standard"
1.1.2.4.3. Known group comparison - you know how members of certain groups should score
1.1.2.4.4. Predictive validity - should predict what you know it ought to know
1.1.3. Determining Cause and Effect
1.1.3.1. Do they covary?
1.1.3.1.1. Correlation or association
1.1.3.1.2. Determined statistically or qualitatively
1.1.3.2. Is covariation spurious?
1.1.3.2.1. Control for other variables
1.1.3.2.2. Check literature
1.1.3.2.3. Check logic
1.1.3.2.4. Check with colleagues
1.1.3.3. Is time order logical?
1.1.3.3.1. Can't occur at the same time
1.1.3.3.2. Difficult if relationship appears circular
1.1.3.4. Do we have a theory that explains this covariation?
1.1.3.4.1. Start with primative axiom
1.1.3.4.2. Untitled
1.1.3.4.3. What is a theory?
1.1.3.5. Independent Variables
1.1.3.5.1. Not affected by the values of the dependent variables
1.1.3.5.2. Logically precedes dependent variables
1.1.3.6. Dependent Variables
1.1.3.6.1. Some way related to independent variables