Causation and Correlations Assume that you have found an interesting - - PowerPoint PPT Presentation

Causation and Correlations

 Assume that you have found an interesting (new?)

correlation between X & Y

 What should you do?  1) schedule the Ss to return (or start a new study that will

have measures at two different times)

why? to be able to do the cross-lagged analysis

 2) during the delay between measures, think about

possible third variables; add these to second test why? to be able to do the covariance analysis

Partials & Spurious Correls

 theory-driven approach

if the X-Y correlation is spurious via Z then prXY•Z will be zero if the X-Y correlation does not involve Z then prXY•Z will be the same as rXY

 data-driven approach

if prXY•Z = rXY

then the X-Y correlation is not spurious via Z if prXY•Z is zero then the X-Y involves Z (or a correlate of Z)

Sampling

how do you choose a method?

ask yourself how important it is to have a sample that accurately represents the target population if “not very”: convenience if “sort of”: simple random sampling if “very”: stratified random sampling then make sure that the method you selected won’t run into any statistical issues

Choosing a Correlational Method

Surveys Observation What are you trying to measure?

attitudes, values, beliefs, and other unobservable attributes behavior

Is reactivity a serious problem? Is realism important?

no yes

Are you willing to invest time/effort?

no yes

Aging Research

 Hybrid Design 1

younger comparison #2 younger

• lder

comparison #1 if comparison #2 (w/ cohort problem) finds the same as comparison #1 (w/ time-frame problem), then OK i.e., if the two “younger” data same, all is well

Definitions (and more)

 Naturalistic Observation – studying behavior in

everyday environments without getting involved

key threat: reactivity (secondary: observer bias)

 Participant Observation – studying behavior from

within the target group

key threats: reactivity + std. exptr bias (secdry: obsr bias) note: Partic.Obs. is not often possible, since no-consent

• bservation can only occur when and where there is no

reasonable expectation of privacy

 Observer Bias – when the beliefs or expectancies of

the observer (consciously or otherwise) influence what is recorded – note: inter-coder reliability must be .90+

Last-minute Questions

 10 pm on Wed evening:

http://www.justin.tv/directory/science_tech look for “Uipsymeth” stream if it asks for password: “exam3”