Page 1 of 1

Choice experiment design and analysis of pure scenario CE

PostPosted: Wed Jul 13, 2022 9:59 pm
by semduyn
Dear Michiel,

I am conducting a choice experiment where respondents are provided with a future scenario and have to decide whether they want to relocate away or stay in their neighborhood under the scenario.

As this is a pure scenario CE it is quite peculiar and I just want to be sure that I have applied the design correctly in NGENE. I have generated a bayesian efficient design based on priors from a pilot study MNL:

design
;alts = relocate, stay
;rows = 24
;block = 3
;eff = (mnl,d,mean)
;model:
U(relocate) = asc_r [(n,-0.18119,0.25367)]
+ flood_r[(n,-0.01774,0.00366)] * FLOOD[5,10,20,50,100]
+ beach_r[(n,-0.00507,0.00338)] * BEACH[0,-25,-50,-75,-100]
+ economy_r[(n,-0.00949,0.00420)] * ECONOMY[-40,-20,0,20,40]
+ insurance_r[(n,0.00857,0.00180)] * INSURANCE[0,10,25,50,100,200]
$

Is this design appropriate?
Also, I now use 24 rows and 3 blocks. Where should I base the decision on number of rows/blocks on (all respondents will receive 8 cards).

I also was wondering if I can just use an MNL regression even though I have just two choice alternatives (relocate or stay)? Does an MNL automatically take into account the binary nature of the choices?

Many thanks in advance for your help!

Best regards,
Sem

Re: Choice experiment design and analysis of pure scenario C

PostPosted: Thu Jul 14, 2022 9:51 am
by Michiel Bliemer
Yes that model and syntax looks good. For another example of a pure scenario choice experiment, see this publication:

De Bekker-Grob, E.W., A.D. Bergstra, M.C.J. Bliemer, I.J.M. Trijssenaar-Buhre, A. Burdorf (2015) Protective behaviour of citizens to transport accidents involving hazardous materials: a discrete choice experiment applied to populated areas nearby waterways. PLoS ONE, Vol. 10, Issue 11, pp. 1-14.

Your choice of rows and blocks is fine.

If you want attribute level balance, you could choose 30 rows (you have attributes with 5 and 6 levels) and block the design in 3.

MNL is appropriate for any number of alternatives, including 2 alternatives (binary choice), it works the same.

Michiel

Re: Choice experiment design and analysis of pure scenario C

PostPosted: Thu Jul 14, 2022 5:07 pm
by semduyn
Dear Michiel,

Many thanks for your swift and very helpful suggestions!
Good to know that I am on the right track.

All the best,
Sem