The method of paired comparison was introduced nearly 150 years ago and is perhaps the most straightforward way of presenting items for comparative judgment. PAIRCOMP administers the paired comparison experiment that presents the pairs of items on the monitor in a unique random order to control for order effects.
With the paired-comparison method, items are presented in pairs to one or more judges. For each pair the judge selects the item that best satisfies the specified judgment criterion. The items can be of almost any type including, importantly, goods, services, and environmental conditions.
As demonstrated by Thurstone in the 1920s, the method can yield an interval-scale ordering of items along a dimension such as preference or importance. For example, if the items are alternative improvements in forest condition and the judges are members of the public, the method can reveal the public’s collective judgment of the relative importance of the improvements. When the items to be judged include both goods and monetary amounts, the method can yield estimates of monetary value of the goods. More information and publications on the method of paired comparison are available from RMRS.
The PAIRCOMP program administers the paired comparison experiment on a personal computer that presents the pairs of items on the monitor in a unique random order to control for order effects.
To download and install PAIRCOMP for Windows 95/2000, save this install file (paircomp.exe, 340 KB) to your hard drive. Then double-click on the file to extract the application, user guide, and sample files.
The items entered into the program appear side-by-side on the monitor, with their position (right versus left) also randomized. The respondent enters a choice by pressing the right or left arrow key and can correct a mistake by pressing the backspace key and then selecting the other item.
At the end of the paired comparisons, the program repeats in random order those pairs for which the individual's choice was not consistent with the dominant preference order as defined by the preference scores. The program also randomly selects ten consistent pairs for retesting. The individual pairs in these two sets of repeated choices are randomly intermixed, and there is nothing to indicate to the respondent when the original pairs end and the repeats begin. The program also records the time taken to enter each choice.
PAIRCOMP runs on a personal computer with Microsoft Windows 95/NT.