Calculation of "Sampling Weight"

Nilu Babu

unread,

Jul 27, 2022, 4:26:41 AM7/27/22

to Biogeme

Dear Concern

Dear Prof. Bierlaire,

I hope this email finds you well.

I have recently started course on Discrete Choice Modelling and facing some difficulties on survey data. For simulation "Switzerland mode choice" data named "Optima" is used. In that survey data there is a column named "Weight" and I am not understanding how to calculate this data for every individual especially when the variable is more than one. I have attached the data file in excel format which is already in the website for your kind reference.

I would be grateful to you if someone could explain this calculation.

Regards

Nilu

optima.xlsx

Bierlaire Michel

unread,

Jul 27, 2022, 4:30:33 AM7/27/22

to niliub...@gmail.com, Bierlaire Michel, Biogeme

THe population is segmented based on 6 socio-economic characteristics, as described in the document.

https://transp-or.epfl.ch/documents/technicalReports/CS_OptimaDescription.pdf

The weight is the ratio between the share of persons in the corresponding segment in the population, and the share of persons in the corresponding segment in the sample.

--
You received this message because you are subscribed to the Google Groups "Biogeme" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biogeme+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/biogeme/bbf188b6-c1f2-4f32-b7ef-8e6d6d174b8bn%40googlegroups.com.
<optima.xlsx>

Nilu Babu

unread,

Jul 27, 2022, 11:47:03 AM7/27/22

to Bierlaire Michel, Biogeme

Thank you professor for your kind response. I got the point that "The weight is the ratio between the share of persons in the corresponding segment in the population, and the share of persons in the corresponding segment in the sample". But I couldn't understand why the summation of the "Weight" is 1 (in the Optima file) . Is it the probability of an individual?

Bierlaire Michel

unread,

Jul 27, 2022, 11:47:46 AM7/27/22

to niliub...@gmail.com, Bierlaire Michel, Biogeme

On 27 Jul 2022, at 11:22, niliub...@gmail.com wrote:

Thank you professor for your kind response. I got the point that "The weight is the ratio between the share of persons in the corresponding segment in the population, and the share of persons in the corresponding segment in the sample". But I couldn't understand why the summation of the "Weight" is 1 (in the Optima file) . Is it the probability of an individual?

It is just a normalization.

Nilu Babu

unread,

Jul 28, 2022, 2:39:33 AM7/28/22

to Bierlaire Michel, Biogeme

Thank you for the clarification.

Reply all

Reply to author

Forward