|
Learn
more about Adaptive Conjoint Analysis
What
is Conjoint?
Conjoint
data collection and analysis was introduced to the market
research community in the mid-1960s. It was based on academic
work in psychometrics conducted at Bell Labs, Princeton, Univ.
of North Carolina and other institutions.
The premise
of conjoint data collection is that by asking respondents
to state their preferences among alternative products that
systematically vary in their features, researchers can "discover"
each respondent's underlying "utilities" for product features.
A typical conjoint data collection procedure in the '70s and
'80s would be to design 30 or more cards, each representing
various collections of product features. Respondents would
sort the cards from most to least desirable. From the rankings,
each respondent's utilities can be calculated.
By forcing
respondents to "trade off" product features for each other,
conjoint data collection provides more realistic guidance
for product development than simply asking for importance
ratings.
What
is Adaptive Conjoint?
Adaptive
Conjoint Analysis (ACA) is a "hybrid" data collection technique
that combines self-explicated importance ratings with pair-wise
trade-off tasks. The ACA algorithm was programmed by Rich
Johnson of Sawtooth Software, Inc. (SSI) in 1985. It has worked
quite satisfactorily on DOS computers ever since.
ACA has
the virtue of allowing for relatively larger conjoint designs
than could be handled by other conjoint methods. It does this
by first asking respondents for explicit importance ratings
(the "priors" section), followed by trade-off tasks that only
include those attributes and levels each respondent rates
as most important. Thus, the interview is tailored to each
respondent.
What's
the relationship between Conjoint Online and Sawtooth Software,
Inc.?
In 1998,
Conjoint Online re-programmed the interviewing module of ACA
so that it would work in an Internet environment. The conversion
was based on the C source code for ACA's interviewing module.
Conjoint Online doesn't sell software, but provides
the conjoint application to users who don't want to purchase
the ACA software or who don't want to install and maintain
it on their web site. Conjoint Online also provides support
services, including free pretesting, so that users can enjoy
the benefits of conjoint quickly and at low cost.
Is
the Conjoint Online implementation of ACA the same as SSI's?
Conjoint
Online produces identical results -- because it uses the SSI
algorithm. However, SSI's ACA has many features that are infrequently
used. As users need some of the esoteric features, Conjoint
Online has added them.
Also,
Conjoint Online has introduced an option to the ACA algorithm
that makes it more suitable for situations where individual
feedback is important. In the original ACA algorithm, reversals
could occur in unimportant attributes. For example, if you
had a slight preference for a blue over a black car, during
the pairs section, ACA can sometimes become confused and think
you prefer a black car. ACA might then ask you whether you'd
prefer a blue car at a lower price to a black car at a higher
price. Since you prefer blue to black anyway, you might think
the computer slipped a cog. As this can be confusing to respondents,
Conjoint Online has an option to prevent reversals from occuring.
In practice,
reversals don't matter much if you are doing aggregate analysis
(grouping respondents). However, if you want to present a
respondent's results back to him/her, prohibiting reversals
can be useful.
Example
of Individual Level Analysis
Matching
Candidates with Job Openings
(Man Jit Singh and Sam Kingsley, Sawtooth Software Conference,
February 1999)
The authors
demonstrated how ACA is being used on the Web in matching
potential job candidates with job openings. Sing and Kingsley
are the first to develop a Web-enabled version of ACA based
on Sawtooth Software's underlying code. To date, hundreds
of thousands of applicants have completed their ACA survey
on job preferences on the Web.
According
to the authors, "the job search and recruiting business has
traditionally relied on the 'telephone tree'", followed by
a detailed review by senior associates, follow-up interviews,
offers and counter-offers. By collecting much information
up-front in a standard survey, followed by a more detailed
assessment of the tradeoffs (ACA) each applicant makes regarding
compensation and other job-related elements, a large number
of applicants can be sorted through efficiently.
Because
the level of analysis is at the individual, part worth reversals
can be problematic. While ACA generally results in fewer reversals
than other conjoint methods, they still can occur. Simply
averaging and "tieing" procedures are one suggested remedy.
[Read
more >>]
|
