topic Re: Designing a discrete choice experiment in Statistical Procedures

Designing a discrete choice experiment

lvanwijk — Thu, 30 Apr 2020 16:40:21 GMT

Hi everybody,

I have a question about programming a DCE, and I really hope somebody can help me since I'm not used to work in SAS (different coding..).
I want to create a DCE with two options (vaginal delivery vs caesarean delivery) and different scenario alternatives consisting of 5 attributes, and 2-3 levels (see below)

Attributes	Level 1	level 2	level 3
Age mother	22 year	32 year	42 year
BMI mother	23	32
Parity	Primiparity	Multiparity – in obstetric history normal vaginal birth	Multiparity – in obstetric history previous caesarean for dystocia
Underlying pathology	No – uncomplicated pregnancy	Yes – mild preeclampsia
Estimated birth weight	2400g	3300g	4000g

So 108 possibilities and I think it should be a factorial design.

I tried looking in the community and searching in PROC BCHOICE but it won't work. See below for my attempt. What am I doing wrong and what is next in order to create the question like the added file?

Thank you so much!

%mktruns (2 ** 2 3 ** 3);

%mktex (2 ** 2 3 ** 3, n=72, seed= ?);

Proc format;
value Age 1 = 22 2 = 32 3 = 42;
value BMI 1 = 23 2 = 32 ;
value Parity 1 = primiparity 2= multiparity in obstetric history normal vaginal birth 3 = multiparity in obstetric history previous caesarean for abnormal CTG;
value Underlying_pathology 1 = no uncomplicated pregnancy 2 = yes mild preeclampsia;
value Estimated_birth_weight 1 = 2400 2 = 3300 3 = 4000;
run;

%mktlab(data=design, /* input data set */

Vars= Age BMI Parity Underlying_pathology Estimated_birth_weight, /* new attribute names */

int=f1-f2, /* create 2 columns of 1’s in f1-f2 */

out=final, /* output design add a format statement for the attributes */

stms=format Age Age. BMI BMI. Parity Parity. Underlying_pathology Underlying_pathology. Estimated_birth_weight Estimated_birth_weight.)
proc print; run;

Re: Designing a discrete choice experiment

WarrenKuhfeld — Thu, 30 Apr 2020 18:45:08 GMT

https://support.sas.com/techsup/technote/mr2010c.pdf

There is an extensive discussion in the link. Search for "candidate set of alternatives" examples. There are many other possibilities as well.

Re: Designing a discrete choice experiment

lvanwijk — Fri, 01 May 2020 14:35:51 GMT

Thanks for your response.

According to example 6 I tried programming my DCE. However it doesn't work when I'm running it.

What am I doing wrong?

Thanks again!

Title DCE 1-5 #2
%mktruns(3 2 3 2 3)
%mktex(3 2 3 2 3, /* factor level list for one alternative */
n=72) /* number of candidate alternatives */

proc format;
value age 1 = 22y 2 = 32 y 3 = 42 y;
value BMI 1 = 25 2 = 32;
value Parity 1= Primiparity 2 = ’Multiparity – in obstetric history normal vaginal birth’ 3 = ’Multiparity – in obstetric history previous caesarean for abnormal CTG’;
value 'Underlying pathology' 1 = ’No – uncomplicated pregnancy’ 2 = ’Yes – mild preeclampsia’;
value 'Estimated birth weight' 1 = 2400g 2 = 3300g 3 = 4000g;
run;

%mktlab(data=design, /* input data set */
vars=Age BMI Parity 'Underlying pathology' 'Estimated birth weight', /* new attribute names */
int=f1-f4, /* create 4 columns of 1’s in f1-f4 */
out=final, /* output design */
/* add a format statement for the attributes */
stmts=format age age. bmi bmi. parity parity. 'Underlying pathology' path. 'Estimated birth weight' weight.)

proc print; run

Re: Designing a discrete choice experiment

WarrenKuhfeld — Fri, 01 May 2020 14:40:03 GMT

I have not run your code, but you need to get the quotes right in you value statements for starts.