Re: st: reshaping data?

[Nick Cox <njcoxstata@gmail.com>]

In that case I really don't understand what you are seeking. See also
Yuval's comments.

I've never come across a model in which there are a different number
of predictors for different observations, although I am always happy
to be educated.

It seems to me that

1. Either you are applying a standard model, in which case you can
give literature references.

2. Or this is a new model, in which case you need to explain how it
would be set-up and estimated.

Note that it's easy to get variables such as the mean of the other
prices in the same family. See

FAQ     . . Creating variables recording prop. of the other members of a group
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        4/05    How do I create variables summarizing for each
                individual properties of the other members of a
                group?
                http://www.stata.com/support/faqs/data/members.html


On Thu, Oct 25, 2012 at 3:39 PM,  <agostino@unical.it> wrote:
> dear Nick,
> thanks for your prompt reply. Unfortunately, the number of items is not
> fixed
> best regards
> Maria
>
>
> Quoting Nick Cox <njcoxstata@gmail.com>:
>
>> This model formulation makes me feel a bit queasy, but I think what
>> you want is something like this.
>>
>> Suppose for concreteness the number of items is fixed at 8. (I don't
>> see how this will work if the number is not fixed.) So, "for any value
>> of 8"
>>
>> sort code_family code_item
>> forval i = 1/8 {
>>        by code_family : gen P`i' = P[`i']
>> }
>>
>> Note that, contrary to your title, there is no -reshape- here as you
>> want your observations to remain observations; at least that's my
>> understanding.
>>
>> Nick
>>
>> On Thu, Oct 25, 2012 at 3:08 PM,  <agostino@unical.it> wrote:
>>
>>> I have to estimate the equation
>>>
>>> Q1=a+b1P1+b2P2+...bnPn+e
>>>
>>>
>>> Where Q1 is the quantity of item 1 sold by a supermarket during a week ,
>>> P1
>>> is the price of item 1 in that week, the other prices are those of the n
>>> items belonging to the same family of items. My data set is organized as
>>> follows:
>>>
>>>
>>> Code_item       week    Q       P       Code_family
>>> 789             1       8       2       1
>>> 790             1       25      4       1
>>> 791             1       9       1.3     1
>>>
>>> 792             1       12      2       1
>>>
>>> 800             1       7       2       2
>>> 801             1       20      1.2     2
>>> 802             1       11      1.6     2
>>>
>>> 803             1       12      2       2
>>>
>>> And so forth for the other weeks and families...
>>>
>>>
>>> For each item, how can I include in my regression the prices of the other
>>> (n-1) items of the same family, ignoring the prices of the items
>>> belonging
>>> to other families?
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/