How to Create a Parameter of Standard Errors Table for a Choice Modeling Experimental Design

This article describes how to create a table of parameter fits and their standard errors for a Choice Modeling - Experimental Design output.

The values are generally compared between designs created by different algorithms for the same specification, rather than assessed on an individual basis.

For the examples above, the Random design has higher standard errors and D-errors compared to the Complete enumeration design.

The parameter estimates from the logit model can generally be ignored and are non-zero because of randomness.

Requirements

A Choice Model Experimental Design output

Please note these steps require a Displayr license.

Method

1. Select the Experimental Design output.

2. Select Anything () > Advanced Analysis > Choice Modeling > Diagnostic > Experimental Design > Parameter Standard Errors of Design.

As an alternative:

From the Choice Modeling Experimental Design dialog Object Inspector, select DIAGNOSTICS > Parameters Standard Errors of Design.

Technical Details

Random responses are created for 300 respondents for the experimental design and the resulting data is fit with a multinomial logit model. The D-error is also calculated. If priors were supplied when creating the design, they will be used in generating the responses and hence be reflected in the standard errors (even if the design algorithm did not use the priors). D-error also accounts for priors if supplied.

References

Hoare, J. (2018, July 20). How Good is Your Choice Model Experimental Design? [Blog post]. Accessed from https://www.displayr.com/how-good-is-your-choice-model-experimental-design/.

See also Choice Modeling - Experimental Design.

For details of the D-error calculation see

1. Yap, J. (2018, August 20). What is D-Error? [Blog post]. Accessed from https://www.displayr.com/what-is-d-error/.
2. Yap, J. (2018, August 21). How to Compute D-error for a Choice Experiment [Blog post]. Accessed from https://www.displayr.com/how-to-compute-d-error-for-a-choice-experiment/.
3. Huber, J., & Zwerina, K. (1996). The importance of utility balance in efficient choice designs. Journal of Marketing research, 307-317. Accessed from https://people.duke.edu/~jch8/bio/publications.htm.

Next

How to Create an Experimental Design for Choice-Based Conjoint