Most popular tags

cbc lighthouse-9 acbc ssi-web maxdiff ssi-web-8 javascript cbc-hb grid-question constructed-list unverified-perl ssi-web-7 free-format skip-logic alternative-specific-design select-question willingness-to-pay choice-simulator cbc-latent-class quota-control acbc-hb mbc pass-in-fields conditional-pricing smrt

Have an idea?

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

Wrong variable displayed as respondent weight in MBC "utility runs" tab

When reviewing previous runs I found that the variable displayed as applied weight was some other variable than my "WEIGHT_final". In fact, it mentioned some var that wasn't in the Demographics file at all (anymore), e.g.:

Filters & Weights
Included Tasks: Task3,Task4,Task5,Task6,Task7,Task8,Task9,Task10,Task11,Task12
Respondent filters: None
Respondent weighting variable: Att2_payin.2_2

At one point it actually said the applied weight was "CHOICE", i.e., my dependent variable. This is just a bug in displaying I suppose?
asked Jan 8, 2018 by alex.wendland Bronze (2,545 points)

1 Answer

+1 vote
I would say it is probably just a display bug in the report, but you can send me the files and I will try to determine whether that is the case.
answered Jan 8, 2018 by Walter Williams Gold Sawtooth Software, Inc. (23,405 points)
I sent the MBC, data and demographic file to walt@sawtoothsoftware.com
Thanks!
commented Jan 9, 2018 by alex.wendland Bronze (2,545 points)
This is a bug which can happen when the locale used in the Preferences window is set to something other than the machine's locale, and respondent weighting is being applied to a Logit estimation.  In this particular case, the machine was set to German (Germany), and the software was set to use English (US).  An updated version of MBC is being released to fix the issue.
commented Jan 10, 2018 by Walter Williams Gold Sawtooth Software, Inc. (23,405 points)
Powered by Question2Answer
...