Research Methods in Psychology: Quasi-Experimental Designs and Interrupted Time Series - P, Study notes of Psychology

Download Research Methods in Psychology: Quasi-Experimental Designs and Interrupted Time Series - P and more Study notes Psychology in PDF only on Docsity!

031:010 - Research Methods in Psychology – Spring, 2009 – April 2nd

1

Outline

 P x E Designs (revisited)

 Interrupted Time Series Designs

P x E Designs (revisited)

 Two ways to run these quasi-experiments:

1. take one sample and split the data – after the fact – in terms of the person factor

• label: ex-post-facto quasi-experiment

2. take two samples, one inside each level of the person factor

• label: planned quasi-experiment

P x E Designs (revisited)

 Why does the distinction between ex-post-facto and planned quasi-experiments matter? answer depends on how many person factors are going to be included If there is only one person factor, then the only important difference is: planned quasi-expts have equal-sized groups ex-post-facto quasi-expts often do not therefore, planned usually have more power

P x E Designs (revisited)

 Example with two person factors: manipulated IV: stimulus-response compatibility PF#1: handedness (note: left-handers often show smaller SRC effects) PF#2: athleticism (note: athletes often show smaller SRC effects) Problem for the ex-post-facto approach: left-handers are more likely to be athletic (there’s a correlation between handedness and athleticism)

P x E Designs (revisited)

 Example with two person factors: PF#1: handedness PF#2: athleticism

table of Ns (out of 200 people sampled) athletes non-athlete right-handed 50 130 left-handed 10 10

Interrupted Time Series Design

 Many real-world “interventions” involve applying some treatment (to a group of people) and looking for a change in their behavior e.g., try a new way of teaching subtraction introduce a new anti-binge-drinking program  The problem is the automatic confound between before-vs-after treatment and time-of-testing  One solution to this problem is to use a non- equivalent control group in some other location  Another solution is to use an interrupted time series

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

is the drop in the DV due to the treatment or due to (anything related to) time?

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

Interrupted Time Series Design

treatment applied

whatever you’re interested in^ time

Interrupted Time Series Design

 If you have access to a second group (with little or no possible “cross-talk” between groups): add a nonequivalent control group apply the treatment to both groups, but at different times ( staggered time series )