Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

ACBC choice tasks vs CBC

We are beginning to do a beta test of ACBC to see if that method works well for how we use conjoint methodology. As I read, I am debating the utility (pun intended!) of the BYO, Screener, and Unacceptables/Must Haves for our purposes. That leaves the choice tasks.

What is the practical difference and/or advantage of ACBC choice task questions versus just CBC, if any?

asked Feb 17, 2015 by WTW_cal Bronze (1,535 points)

1 Answer

0 votes
This is mostly covered in the ACBC Technical Paper on our website, but let me hit some high points:

BYO Section:  a) allows the respondent to tell us directly what product is of most interest to them (empowers the respondent), b) allows us to customize the rest of the interview via the "near-neighbor" design approach so that the tradeoffs and products seen are more relevant to the respondent than CBC, c) provides an excellent education/training task to introduce respondents to each attribute and the levels of the attributes.

Screener Section: a) allows us to identify Must-Have and Unacceptable levels...and customize the final Tournament section to avoid or include those requirements--this gets at discovering non-compensatory aspects to respondents' decision-making.

CBC-looking Tournament: a) provides the compensatory trade-off portion of the interview that we believe buyers often do when choosing among the final few products within their consideration set, b) mimics closely the CBC tasks that have proven so useful in practice.

The neat thing about ACBC is that the researcher can choose to mix and match, shorten or extend different sections of the ACBC survey, to create a trade-off questionnaire that elicits the kinds of tradeoffs and choice context that most make sense for modeling a particular product category.   Running the "Test Design" option within the software will report absolute standard errors (from aggregate logit) so you can look at how making different changes to the ACBC questionnaire affects the overall statistical precision of the utility estimates.

Now, if you skip BYO and skip the Screener section for ACBC, your ACBC survey will now appear a lot to the respondent like a CBC survey, except:

1.  Losing products go away, winners are carried forward
2.  You cannot include a "None" concept
3.  You are limited to showing either 2 products at a time or 3 (not great for FMCG categories like beverages and yogurts for example)
4.  You cannot use the "shelf display" or the Custom Free Format displays as offered in CBC
5.  Any "tied" attributes across products are "greyed out"
answered Feb 18, 2015 by Bryan Orme Platinum Sawtooth Software, Inc. (154,105 points)
Recommendation on number of Tournament questions