[Hadji Jalotjot]

Hello! I'm trying generate a choice experiment design. O'm fine with the initial commands but when it is time to use the dcreate command, the result is always an error. Any idea on where I made a mistake?

[Hadji Jalotjot]

I kinda figured it out by revising the matrix b into J(1,8,0)....Is this correct?

[Hadji Jalotjot]

I finally got until the end but I have another question. I used the follwing commands:

dcreate i.x1 i.x2 i.x3 i.x4 i.x5, nalt(2) nset(8) bmat(b)
blockdes block, nblock(2)
list

I was expecting to get 8 choice sets = 4 choice sets for each Block 1 and Block 2.
Instead, I got a list of 16 choices sets --> 8 choice sets for Block 1, 8 choice sets for Block 2. Is this corrrect?

[Sara Maria Barbani]

Dear all,

we need to develop a DCE design for two types of different cancer screenings, we call them type1 and type2.
Even if they have the same purpose, these two cancer screenings differ in some attributes and level of these attributes. What we mean is: Cancer Screening Type 1 has the attribute X whose levels could be A, B or C whereas Cancer Screening Type 2 has the same attribute X whose levels could be M, N or O. Here, the levels of the attributes vary in the two alternatives according to different mechanisms. Moreover, in this case, they could not be swapped places each other, I mean, the levels that differ of screening Type 1 from Type 2 will never be the same.
Note that not all the attributes may necessarily provide levels that vary across alternatives, we also include some attributes where levels remain stable.
This is an example:

	Cancer Screening 1	Cancer Screening 2
Attributes
Reliability	95%, 90%, 85%	95%, 90%, 85%
Price	25 €, 120 €, 150 €	20 €, 30 €, 40 €
how often to repeat the test	6 months, 1 year, 2 years	6 months, 1 year, 2 years
where	specialist	Laboratory, home, pharmacy
results	immediately	7 days, 10 days, 14 days

We have employed dcreate however we have some troubles in creating these different levels and how to impose that nalt 1 corresponds to levels of type 1 whereas nalt2 to type 2.
Can you help us? Is it possible to create this design on Stata?