From shirshobiswas at gmail.com Thu Nov 1 08:23:16 2018 From: shirshobiswas at gmail.com (Shirsho Biswas) Date: Wed, 31 Oct 2018 16:23:16 -0500 Subject: [Limdep Nlogit List] Partial Effects for Tobit Fixed Effects Model Message-ID: Hi, I have estimated a Tobit Fixed Effects model. The model includes a couple of binary dummy variables. Hence, instead of computing partial effects at the means of all Xs (which doesn't make much sense for binary variables), I wish to compute partial effects averaged over all data observations. By specifying the model as: TOBIT ; ...; FEM; Cluster; Partials$ PARTIALS; Effects: ...., Summary$ I do get the partial effects from the Tobit command as expected (computed at the means of all Xs) The second PARTIALS command gives me the following error: "Previous model did not store model setup for PARTIALS. Check model setup and output for errors. Setup does not occur if model command contains Expand(variable)" Is this problem caused because of the Fixed Effects? Is there any way, that I can compute the partials that I need? Best, Shirsho From gutamajale at gmail.com Tue Nov 6 05:40:14 2018 From: gutamajale at gmail.com (S Tarfasa) Date: Mon, 5 Nov 2018 10:40:14 -0800 Subject: [Limdep Nlogit List] Multiple line data set up in NLOGIT 5 to a single line ? Message-ID: I have one question on the use of NLOGIT,i.e., is it possible to convert a multiple line data set up to a single line format using NLOGIT 5? Could you please help me how to do the conversion, if possible?See below,please Thank you so much in advance for your cooperation. Solomon Dummy variable/attribute:X1 with levels 0,1 Dummy vaiiable /Attribute :X2 with levels 0,1 Attribute:X3 with levels 20,30,40,50 Alternatives: 3 One respondents makes 4 sequences of choice tasks. From liberini at kof.ethz.ch Mon Nov 19 21:47:31 2018 From: liberini at kof.ethz.ch (Liberini Federica) Date: Mon, 19 Nov 2018 10:47:31 +0000 Subject: [Limdep Nlogit List] clustering on a RP nested logit model Message-ID: <213F594826F3B9428143CEB7C51517E04D557FC0@MBX13.d.ethz.ch> Hello, I am trying to estimate a random parameter nested logit model, on a pooled database, where there are 19 alternatives to choose from. When using the cluster options (with command "; cluster=groupname "), I get the error message : Error: 1081: Mismatch of # indivs. and number implied by groups Error: 1081: Mismatch of # indivs. and number implied by groups I should point out that the data is sorted by group first, then by individuals, then alternatives. I also tried sorting with different orders and it still did not work. Additionally, I checked the data for any bug on other softwares (e.g. stata) and there I had no problems with the clusters. Any idea of what could be triggering the error code? Many thanks Federica From wgreene at stern.nyu.edu Wed Nov 28 00:50:33 2018 From: wgreene at stern.nyu.edu (William Greene) Date: Tue, 27 Nov 2018 08:50:33 -0500 Subject: [Limdep Nlogit List] clustering on a RP nested logit model In-Reply-To: <213F594826F3B9428143CEB7C51517E04D557FC0@MBX13.d.ethz.ch> References: <213F594826F3B9428143CEB7C51517E04D557FC0@MBX13.d.ethz.ch> Message-ID: Federica. It sounds like there are 2 levels of grouping here - "groups" and "alternatives." For purposes of the RPLogit models, the cluster, such as it is - more on that later - is the number of choice tasks. But, the group will contain 19*groupsize rows of data in the group. You would use ;PDS = 19 for your model to set up the RP model. I do note, the conventional "cluster" idea doesn't make sense in a random parameters multinomial choice model. The RPs are creating the clustering effect, and that is already accounted for in the estimates and the asymptotic covariance matrix. The latter is already hugely complicated, and already accounts for the common random effect across group members. Regards Bill Greene On Mon, Nov 19, 2018 at 5:48 AM Liberini Federica wrote: > Hello, > > I am trying to estimate a random parameter nested logit model, on a pooled > database, where there are 19 alternatives to choose from. When using the > cluster options (with command "; cluster=groupname "), I get the error > message : > > Error: 1081: Mismatch of # indivs. and number implied by groups > Error: 1081: Mismatch of # indivs. and number implied by groups > > I should point out that the data is sorted by group first, then by > individuals, then alternatives. I also tried sorting with different orders > and it still did not work. > Additionally, I checked the data for any bug on other softwares (e.g. > stata) and there I had no problems with the clusters. > > Any idea of what could be triggering the error code? > > Many thanks > Federica > _______________________________________________ > Limdep site list > Limdep at mailman.sydney.edu.au > http://limdep.itls.usyd.edu.au > > -- William Greene Department of Economics Stern School of Business, New York University 44 West 4 St., 7-90 New York, NY, 10012 URL: https://protect-au.mimecast.com/s/2SvmCk8vAZtPJAV1h2Wjrv?domain=people.stern.nyu.edu Email: wgreene at stern.nyu.edu Ph. +1.212.998.0876 Editor in Chief: Journal of Productivity Analysis Editor in Chief: Foundations and Trends in Econometrics Associate Editor: Economics Letters Associate Editor: Journal of Business and Economic Statistics Associate Editor: Journal of Choice Modeling