Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: weight option
From
tashi lama <[email protected]>
To
<[email protected]>
Subject
RE: st: weight option
Date
Thu, 23 Aug 2012 20:53:05 +0000
True. I could have done that. As a matter of fact, I found
graph pie hits [pw=1/doc_published], over(analyst) gives me the same pie chart as
graph pie hitsperdocument, over(analyst) where hitsperdocument=hits/doc_published
But I wanted to get used to knowing weight option which I haven't used yet.
Thanks
----------------------------------------
> From: [email protected]
> To: [email protected]
> Subject: RE: st: weight option
> Date: Thu, 23 Aug 2012 13:39:46 -0700
>
> Tashi,
> It sounds like you want something like hits per document, right? So why not
> just create that measure directly by dividing hits by doc_published?
> -Sarah
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of tashi lama
> Sent: Thursday, August 23, 2012 12:50 PM
> To: [email protected]
> Subject: RE: st: weight option
>
> In my opinion, that would be opposite of what I would like to have. If I use
> frequency weights the way you suggested, would it not give more weight to
> analysts with higher doc_published(since higher doc_published means higher
> frequency) which is opposite of what I want? I want analysts with more
> doc_published to have less weight. I would think may be reciprocal of
> fweight. May be something like [fweight=1/doc_published] but stata doesn't
> allow non-integer for fweight. Looks like pweight is closer to what I want.
> Any ideas?
>
> Thanks
> ----------------------------------------
> > Subject: Re: st: weight option
> > To: [email protected]
> > From: [email protected]
> > Date: Thu, 23 Aug 2012 19:38:52 +0000
> >
> > I would use frequency weights so Stata will understand that your obs are
> replicated as many times as the weights says to be.
> >
> > graph pie hits [fweight=doc_published], over(analyst) ,
> >
> > Best,
> >
> > Pablo.
> > Sent via BlackBerry from T-Mobile
> >
> > -----Original Message-----
> > From: tashi lama <[email protected]>
> > Sender: [email protected]
> > Date: Thu, 23 Aug 2012 19:32:20
> > To: <[email protected]>
> > Reply-To: [email protected]: st: weight option
> >
> > Hello all,
> >
> > The problem is creating a market share of analysts based on the
> > readership views but also that take into account how many documents
> > were published by the analysts. For instance the guy with 20 views on
> > just 2 documents published should have higher market share than the
> > guy who has 25 views on 5 documents published. Say for instance my
> > dataset would look something like this
> >
> > analyst doc_published hits
> > |---------------------------|
> > | A 5 25 |
> > | B 3 20 |
> > | C 1 18 |
> > | D 2 20 |
> > | E 4 12 |
> >
> > If I were to draw a pie chart, I would simply do graph pie hits,
> > over(analyst) This will indeed give me a nice looking pie chart but
> > this doesn't take document published (which is in column 2) into account.
> However, I want to be able to create a market share of analysts' based on
> readership hits but also takes document published into account. I have a
> feeling it has to do something with [weight] option. Could someone please
> explain which weight should I use and may be why it would be better than
> others?
> > it could be something like this
> > graph pie hits [weight=doc_published], over(analyst)
> >
> > Thanks,
> > Tashi
> > *
> > * For searches and help try:
> > * http://www.stata.com/help.cgi?search
> > * http://www.stata.com/support/statalist/faq
> > * http://www.ats.ucla.edu/stat/stata/
> >
> > *
> > * For searches and help try:
> > * http://www.stata.com/help.cgi?search
> > * http://www.stata.com/support/statalist/faq
> > * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/