Contents - Index - Previous - Next

Study Design

The experimental design used to study the effects of party cues on opinions in this research is the "Posttest Only Control Group Design," * which with four measurement events looks like this:

Treatment Group                   S       T         M1         M2        M3        M4      S      
Control Group                        S      (  )        M1         M2        M3        M4      S

In this diagram, the "S's" stand for Support events (Welcome and End screens), the "T" stands for the Treatment (cues embedded in the questions), the closed parentheses represent a null event - where nothing happens but needs to be specified to hold the time slot, and the "M's" stand for measurements (questions in this case).

Specifying this application a bit further, we have two measurements that are Support/Oppose Questions and two measurements are paired Intensity measurements (as in "How strong is your opinion?").  See below.

Treatment Group                   S       T         Q1         I1        Q2        I2      S      
Control Group                        S      (  )        Q1         I1        Q2        I2      S

This design is unlike many if not most experimental designs because two of the measurement events (Q1 & Q2) in the control group are different from the measurement events (Q1 & Q2) in the treatment group.  The reason is that party cues (the treatment) are used in the questions of the treatment group, but they are not used in the control group.
 
*  For a popular reference on experimental designs, see Donald T. Campbell & Julian C. Stanley, Experimental and Quasi-Experimental Designs for Research, Houghton Mifflin Co., Boston, 1963.