PO217 Lecture Notes - Lecture 5: Observational Error, Predictive Validity, Face Validity
Document Summary
Broad knowledge: full pictures complimented by adding the details. The level of measurement that can be achieved depends on: The choice of data collection procedures - how you go about collect it. The general rule is to aim for the highest possible level of measurement because higher levels of measurements enable us to person more powerful and more varied tests. Sometimes you can have tmi: likely little difference between 22 and a 23 year-old and their views on education. In that case, we may want to re- code our variables. We can always bring the interval-ratio level down to the ordinal level but we can"t bring it up. we can"t bring it up. Involves classifying a variable into two or more categories and then sorting our observations into the appropriate category, but there are no orders to these categories. You can assign females #1 and males #2, but these numbers do not mean anything.
