Hi All,
Can someone give me some languages or point me to a paper explaining how CEM upweight or downweight a strata?
I know the weight for each case is 1 and the weights of all controls adding up to the control's sample size. Some reviewers are asking us to give some language to explain the CEM weights for medical audiences.
Thanks a lot!
Fang
Hi Meysam. see Section 5.1 of this article
<https://gking.harvard.edu/files/political_analysis-2011-iacus-pan_mpr013.pdf>.
CEM works fine with more than 2 treatment categories. However, matching
across 6 categories will require a lot of data or luck. alternatively you
can dichotomize, perhaps repeatedly, although beware of changing quantities
of interest.
Gary
--
*Gary King* - Albert J. Weatherhead III University Professor - Director,
IQSS <http://iq.harvard.edu/> - Harvard University
GaryKing.org - King(a)Harvard.edu - @KingGary <https://twitter.com/kinggary> -
617-500-7570 - Assistant <king-assist(a)iq.harvard.edu>: 617-495-9271
On Mon, Apr 9, 2018 at 4:55 PM, Meysam Alizadeh <alizadeh.meysam(a)gmail.com>
wrote:
> Hi,
>
> I'm comparing the average values of a variable between 6 different groups.
> Before comparing, I need to control for some covariates. I decided to use
> matching for this purpose. However, I don't know what is the best way to
> match these 6 samples since CEM only works for matching 2 samples.
>
> Any help is appreciated.
>
> Best,
> Meysam
>
> --
> *Meysam Alizadeh*
> Postdoctoral Fellow | Center for Complex Networks and Systems Research
> School of Informatics and Computing
> Indiana University Bloomington
> Web <http://malizad.github.io> | Twitter
> <https://twitter.com/M_Aliizadeh> | LinkedIn
> <https://qa.linkedin.com/in/malizadeh>
>
Hi,
I'm comparing the average values of a variable between 6 different groups.
Before comparing, I need to control for some covariates. I decided to use
matching for this purpose. However, I don't know what is the best way to
match these 6 samples since CEM only works for matching 2 samples.
Any help is appreciated.
Best,
Meysam
--
*Meysam Alizadeh*
Postdoctoral Fellow | Center for Complex Networks and Systems Research
School of Informatics and Computing
Indiana University Bloomington
Web <http://malizad.github.io> | Twitter <https://twitter.com/M_Aliizadeh>
| LinkedIn <https://qa.linkedin.com/in/malizadeh>
Hi everyone,
I have problems in running cem which I believe is related to an error in
the installation procedure.
I installed cem using:
sysdir set PLUS "U:[folder location]ado\plus"
ssc install cem
When running cem i get the error message:
file U:[folder location]ado\plus/c/cem-mata.do not found
(error occurred while loading cem.ado)
But I checked: the file "cem-mata.do" is exactly at the location where
Stata is looking for it. Folder and file are named correctly and I can open
it. Everything seems fine with the file. I have no idea how to fix this.
Any help?
Thank you,
Anika
Dear Stefano and Gary,
Many thanks for your prompt responses. They have been very helpful.
All the best,
Cristian
Cristian Vaccari
Reader in Political Communication, Loughborough University
<http://www.lboro.ac.uk/departments/socialsciences/staff/cristian-vaccari/>
Brockington Building U.3.19
Epinal Way, Loughborough, LE11 3TU
Personal website: https://cristianvaccari.com/
Book reviews editor, International Journal of Press/Politics
<http://hij.sagepub.com/>
Research project: Social media and politics in comparative perspective
<http://www.webpoleu.net/>
Latest book: Digital Politics in Western Democracies: A Comparative Study
<http://jhupbooks.press.jhu.edu/ecom/MasterServlet/GetItemDetailsHandler?iN=…>
Google Scholar profile
<http://scholar.google.com/citations?user=3_TethEAAAAJ&hl=en>
Twitter: @25lettori
<https://twitter.com/25lettori>
On Mon, Mar 19, 2018 at 2:17 PM, Gary King <king(a)harvard.edu> wrote:
> Cristian, have a look at http://j.mp/CEMweights which explains the CEM
> weights. Like Stefano, I don't think it would make sense to run an analysis
> without the weights.
>
> Gary
> --
> *Gary King* - Albert J. Weatherhead III University Professor - Director,
> IQSS <http://iq.harvard.edu/> - Harvard University
> GaryKing.org - King(a)Harvard.edu - @KingGary <https://twitter.com/kinggary> -
> 617-500-7570 - Assistant <king-assist(a)iq.harvard.edu>: 617-495-9271
>
> On Fri, Mar 16, 2018 at 11:59 AM, Cristian Vaccari <C.Vaccari(a)lboro.ac.uk>
> wrote:
>
>> Dear all,
>>
>> I am using CEM in a paper based on survey data. The goal is to estimate
>> the relationship between exposure to political incivility on social media
>> and an index of political participation (a count of eight different
>> activities). After matching units based on a set of covariates, I run a
>> negative binomial regression using the weight calculated by CEM that
>> equalizes the number of treated and control units within each stratum.
>>
>> Among the feedback I have received on this paper has been the suggestion
>> that I should run the analysis on the CEM-matched units, but without using
>> the CEM-generated weight.
>>
>> I have tried running the analysis only on the matched units without using
>> the weight and the results are consistent with the analysis where I used
>> the weight, but it seems to me using the weights would be more appropriate.
>> Would anyone be able to comment on whether using the matched units but not
>> the weight is a sensible/acceptable idea or not?
>>
>> Many thanks for your help and for developing this great tool!
>>
>> Cristian
>>
>
>
Cristian, have a look at http://j.mp/CEMweights which explains the CEM
weights. Like Stefano, I don't think it would make sense to run an analysis
without the weights.
Gary
--
*Gary King* - Albert J. Weatherhead III University Professor - Director,
IQSS <http://iq.harvard.edu/> - Harvard University
GaryKing.org - King(a)Harvard.edu - @KingGary <https://twitter.com/kinggary> -
617-500-7570 - Assistant <king-assist(a)iq.harvard.edu>: 617-495-9271
On Fri, Mar 16, 2018 at 11:59 AM, Cristian Vaccari <C.Vaccari(a)lboro.ac.uk>
wrote:
> Dear all,
>
> I am using CEM in a paper based on survey data. The goal is to estimate
> the relationship between exposure to political incivility on social media
> and an index of political participation (a count of eight different
> activities). After matching units based on a set of covariates, I run a
> negative binomial regression using the weight calculated by CEM that
> equalizes the number of treated and control units within each stratum.
>
> Among the feedback I have received on this paper has been the suggestion
> that I should run the analysis on the CEM-matched units, but without using
> the CEM-generated weight.
>
> I have tried running the analysis only on the matched units without using
> the weight and the results are consistent with the analysis where I used
> the weight, but it seems to me using the weights would be more appropriate.
> Would anyone be able to comment on whether using the matched units but not
> the weight is a sensible/acceptable idea or not?
>
> Many thanks for your help and for developing this great tool!
>
> Cristian
>
Dear all,
If you have survey data I suggest you considering also the sampling weights. Preliminary results of a Monte Carlo simulations (unpublished) suggests that they may matter.
Cesare Riillo
On Saturday, 17 March 2018, 17:00:20 CET, cem-request(a)lists.gking.harvard.edu <cem-request(a)lists.gking.harvard.edu> wrote:
Send Cem mailing list submissions to
cem(a)lists.gking.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.gking.harvard.edu/mailman/listinfo/cem
or, via email, send a message with subject or body 'help' to
cem-request(a)lists.gking.harvard.edu
You can reach the person managing the list at
cem-owner(a)lists.gking.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Cem digest..."
Today's Topics:
1. Using CEM without weights (Cristian Vaccari)
2. Re: Using CEM without weights (stefano iacus)
----------------------------------------------------------------------
Message: 1
Date: Fri, 16 Mar 2018 15:59:32 +0000
From: Cristian Vaccari <C.Vaccari(a)lboro.ac.uk>
To: <cem(a)lists.gking.harvard.edu>
Subject: [cem] Using CEM without weights
Message-ID:
<CANbKZGt_fds3CuaWb2BfU5xhY2GXFc3stfM-+MixZMY_pzcXzw(a)mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Dear all,
I am using CEM in a paper based on survey data. The goal is to estimate the
relationship between exposure to political incivility on social media and
an index of political participation (a count of eight different
activities). After matching units based on a set of covariates, I run a
negative binomial regression using the weight calculated by CEM that
equalizes the number of treated and control units within each stratum.
Among the feedback I have received on this paper has been the suggestion
that I should run the analysis on the CEM-matched units, but without using
the CEM-generated weight.
I have tried running the analysis only on the matched units without using
the weight and the results are consistent with the analysis where I used
the weight, but it seems to me using the weights would be more appropriate.
Would anyone be able to comment on whether using the matched units but not
the weight is a sensible/acceptable idea or not?
Many thanks for your help and for developing this great tool!
Cristian
Dear all,
I am using CEM in a paper based on survey data. The goal is to estimate the
relationship between exposure to political incivility on social media and
an index of political participation (a count of eight different
activities). After matching units based on a set of covariates, I run a
negative binomial regression using the weight calculated by CEM that
equalizes the number of treated and control units within each stratum.
Among the feedback I have received on this paper has been the suggestion
that I should run the analysis on the CEM-matched units, but without using
the CEM-generated weight.
I have tried running the analysis only on the matched units without using
the weight and the results are consistent with the analysis where I used
the weight, but it seems to me using the weights would be more appropriate.
Would anyone be able to comment on whether using the matched units but not
the weight is a sensible/acceptable idea or not?
Many thanks for your help and for developing this great tool!
Cristian
Dear all,
Many dataset are results of complex survey design (e.g. uneven probability of selection). In this dataset each observation is associated with a different sampling weights. These weights are different than CEM weights desxribed in j.mp/CEMweights. Sample weights are used to compute sample (weighted) statistics that are representative of population statistic.
My question is, in case of complex survey design, does CEM allow to compute ATT for the sample only or also for the population ? is any adjustment needed ?
my tentative answer is yes and CEM weight should be combined with sample weighting but I would appreciate your advise.
kind regards
Cesare
Cesare A. F. Riillo, PhD
STATEC Research
National Institute of statistics and economic studies of the Grand Duchy of Luxembourg.
13, rue Erasme L-1468 Luxembourg www.statec.lu
Postal address: B.P. 304 L-2013 Luxembourg
Dear all, I have a question partially related to the topic.Many dataset are results of complex survey design (e.g. uneven probability of selection). In this dataset each observation is associated with a different sampling weights. These weights are different than CEM weights desxribed in j.mp/CEMweights. Sample weights are used to compute sample (weighted) statistics that are representative of population statistic.
My question is, in case of complex survey design, does CEM allow to compute ATT for the sample only or also for the population ? is any adjustment needed ? my tentative answer is yes and CEM weight should be combined with sample weighting but I would appreciate your opinion.
kind regards
Cesare
Cesare A. F. Riillo, PhD
STATEC Research
National Institute of statisticsand economic studies of the Grand Duchy of Luxembourg.
Email :cesare.riillo@statec.etat.lu
Direct line: +352-247-84387
Centre Administratif PierreWerner (CAPW)
13, rue Erasme L-1468Luxembourg www.statec.lu
Postal address: B.P. 304 L-2013 Luxembourg
On Monday, 12 February 2018, 14:02:22 CET, cem-request(a)lists.gking.harvard.edu <cem-request(a)lists.gking.harvard.edu> wrote:
Send Cem mailing list submissions to
cem(a)lists.gking.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.gking.harvard.edu/mailman/listinfo/cem
or, via email, send a message with subject or body 'help' to
cem-request(a)lists.gking.harvard.edu
You can reach the person managing the list at
cem-owner(a)lists.gking.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Cem digest..."
Today's Topics:
1. Re: Cem Digest, Vol 99, Issue 1 Combining matching procedures
(cesare riillo)
2. Re: Combining matching procedures (Gary King)
----------------------------------------------------------------------
Message: 1
Date: Mon, 12 Feb 2018 11:04:27 +0000 (UTC)
From: cesare riillo <cesare_riillo(a)yahoo.com>
To: cem(a)lists.gking.harvard.edu, "cem-request(a)lists.gking.harvard.edu"
<cem-request(a)lists.gking.harvard.edu>
Subject: Re: [cem] Cem Digest, Vol 99, Issue 1 Combining matching
procedures
Message-ID: <1442928780.1045722.1518433467855(a)mail.yahoo.com>
Content-Type: text/plain; charset="utf-8"
My feeling is that if you apply Entropy Balancing on CEM matched observations you should combine CEM and Entropy balancing weights.? -your? solution ( A)-.
At least I did so in the case of CEM and complex survey weights in Beyond the question ?Does it pay to be green??: How much green? and when? - ScienceDirect?
|
|
| |
Beyond the question ?Does it pay to be green??: How much green? and when...
|
|
|
please note that there is a typo in page 631, ( the formula for weighted controls after CEM is not reporting "*swi".
But I am also really interested to the topic and would appreciate some more formal reference.
In any case, if Entropy Balancing? balances covariates with respect to the first, second moment and? possibly higher moments? while BEM bounds all centered absolute moments. In this sense, I expect that applying the two methods separately will provide consistent results.
Hope it can help
Cesare
On Friday, 9 February 2018, 18:00:07 CET, cem-request(a)lists.gking.harvard.edu <cem-request(a)lists.gking.harvard.edu> wrote:
Send Cem mailing list submissions to
??? cem(a)lists.gking.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
??? https://lists.gking.harvard.edu/mailman/listinfo/cem
or, via email, send a message with subject or body 'help' to
??? cem-request(a)lists.gking.harvard.edu
You can reach the person managing the list at
??? cem-owner(a)lists.gking.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Cem digest..."
Today's Topics:
? 1. Combining matching procedures (Haakon Gjerl?w)
----------------------------------------------------------------------
Message: 1
Date: Thu, 8 Feb 2018 21:46:55 +0000
From: Haakon Gjerl?w <haakon.gjerlow(a)stv.uio.no>
To: "cem(a)lists.gking.harvard.edu" <cem(a)lists.gking.harvard.edu>
Subject: [cem] Combining matching procedures
Message-ID: <26d789a592874d86a9c300711e316aa4(a)mail-ex02.exprod.uio.no>
Content-Type: text/plain; charset="iso-8859-1"
Dear all,
I have a question concerning the correct way to combine CEM with other matching procedures. Specifically, I am trying to match a data set with CEM, and then apply Entropy balancing to the remaining sample (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1904869). It seems such two-step balancing is hinted at in both Iacus, King, Porro (2011) and Hainmuller (2012)
My questions concerns the correct way to us the weights in regressions after both procedures are done.
My intuition says that this is basically a two-step sampling procedure, and that the correct way to use the weights is to multiply the weights from CEM with the weights from Entropy (Solution A).
However, it might also be that the Entropy Balancing is overriding the weights from CEM, and the observations should only be weighted by the weights from Entropy (Solution B)
Have any of you investigated this in a more systematic/formal fashion?
All the best,
Haakon Gjerl?w | Phd fellow
Department of Political Science | University of Oslo