Conjoint Online

Which Conjoint Method Should I Use?

A brief history of ACA - Adaptive Conjoint Analysis

Conjoint analysis began with research psychologists the 1960s and 1970s. First, there was a theoretical framework provided by Luce & Tukey (1964). This established that it was possible to measure psychological attitudes, abilities or preferences as quantities, like distance or weight. In other words, numbers can be meaningfully assigned to psychological constructs.

Further advances were made at Bell Labs, particularly in computational methods and empirical validation. There, psychometricians were collecting data on individuals' choices (or preferences) made in natural, non-metric judgments, such as "I prefer This to That." The researchers found ways to represent the underlying structure as a map in 1, 2 or more dimensions, even though the original data were just pairwise comparisons (i.e., This vs That). When U Penn researchers (Green & Carmone) began collaborating with market researchers at National Analysts to apply these findings, conjoint analysis was born.

In 1985 Rich Johnson of MarketFacts identified a need for computer-based interviews to facilitate conjoint data collection. Further, he found that his marketing colleagues "always" wanted to know about more attributes, with more precision, than was economical for a survey. Rich formed Sawtooth Software Inc. and developed Adaptive Conjoint Analysis (ACA). ACA efficiently combines users' self-explicated (direct) ratings of importance with pairwise tradeoffs. ACA is able to handle as many as 30 attributes, although such designs may be unwieldy for respondents. The initial self-explicated ratings allow the program to reduce the number of attributes and levels used for pairwise tradeoffs, which produces feasible tasks for respondents.

By the mid-1990s, ACA was by far the most popular conjoint method used by market researchers. In 1996, Sawtooth granted a license to port ACA to the web, in return for providing the technical solutions to Sawtooth Software for their future use.

Where are we today?

Conjoint and Choice-based analyses have been successfully used in market research for over 30 years. Choice-based conjoint (CBC) methods have eclipsed ACA as the most popular approach. Why? CBC more accurately represents the impact of price on choice. ACA can under-represent the impact of price. And CBC can easily include an option for the respondent to say "I wouldn't buy any of these alternatives," thus giving the researcher information about absolute purchase interest.

However, ACA remains a useful tool for researchers. With, researchers have a quick, easy and inexpensive way to product alternatives.

For a full discussion of the strengths and weaknesses of various methods, please visit Sawtooth Software Inc. and read Which Conjoint Method Should I Use?
Please contact if you have further questions:


Home | Archives | Design An ACA Survey || Contact Us

Copyright © 1998 - 2024 Action Research NY & Sawtooth Software Inc. All rights reserved.