Gov2001 March 2007

gov2001@lists.gking.harvard.edu

24 participants
33 discussions

[gov2001-l] what's a rug?

by chingos＠fas.harvard.edu

I checked the help file for the rug() function and apparently it adds a rug to the plot. I'm glad to add a rug to my plot, but would like to know what I'll be adding. Also, does anyone know how to make a graph like the one in Gary's slides on the high school/college wage differential (with the confidence intervals around each data point)? Or maybe this is a rug? Matt

17 years, 2 months

[gov2001-l] errors under reparameterization

by tom.laakso＠gmail.com

It was mentioned in class that the standard errors are not invariant under reparameterization. I am not sure how to address this issue in our numeric estimates of the standard error on the MLE. Is there a correction that I missed in class/section? Or should we just derive them analytically? Tom

17 years, 2 months

[gov2001-l] holding at observed values

by blackwel＠fas.harvard.edu

Hello, I was just looking through problem set 5 and I was a little confused by both 1.3 and 1.4. In those problems, we are asked to calculate the difference in expected probability of failure for a given change in temperature, "holding Pressure at its observed values." Does this mean we should calculate a unique expect change in probability at each level of pressure in the data? If so, would we have 3 plots in 1.4? awesome, thanks, matt.

17 years, 2 months

[gov2001-l] 1.6 error message (for plots)

by msen＠fas.harvard.edu

Hi everyone, I'm trying to make my plots for 1.6 work, but I keep getting the following error message when I run the log.likelihood function with my ruler, y = Y.list$Y, and T = 10. I think I'm doing the exact same thing that Holger did in section, so I'm a little confused as to why this isn't working.... Warning messages: 1: longer object length is not a multiple of shorter object length in: par^y 2: longer object length is not a multiple of shorter object length in: (1 - par)^(T - y) thanks for any help! Maya

17 years, 2 months

[gov2001-l] footnote

by jaharris＠fas.harvard.edu

Does a caption suffice for an "informative footnote" in problem 1.5? If not, could someone please clue me in to effectively placing footnotes within the tabular command in latex? Thanks Andy ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Andy Harris Ph.D. Candidate Department of Government Harvard University

17 years, 2 months

[gov2001-l] Question about the data

by luciska＠gmail.com

Hi... So when I download the data and get ylist, the number of "coin flips" out of which the y's were generated is not clear. It looks like it is around 10 but can someone confirm? However, the vector contains very few small numbers so is it out of a larger value? Thanks... Lucia

17 years, 2 months

[gov2001-l] dataset problems?

by htiet＠fas.harvard.edu

Hi All, I'm having problems loading the dataset for p-set 4. When I load the dataset, it's the same exact data from p-set 2-was this on purpose? If not, does anyone know how to turn off the "Previously saved workspace restored" mechanism in R (it does this every time I first open up R)? I tried opening the p-set 4 dataset without loading it, and it's apparently empty without any data. Thanks, Hai

17 years, 2 months

[gov2001-l] question on hessians/standard errors

by msen＠fas.harvard.edu

Hi everyone, I've gotten to the point where my optim function is giving me a reasonable-looking output. But I'm having a hard time adjusting the function so that it gives me the hessian matrixes (so I can get the standard errors for Question 1.5). It's giving me the following error message: > optim(par = 0, fn = log.likelihood, y = Y.list$Y.3, T = 10, hessian = TRUE) Error in optim(par = 0, fn = log.likelihood, y = Y.list$Y.3, T = 10, hessian = TRUE) : non-finite finite-difference value [1] In addition: Warning message: one-diml optimization by Nelder-Mead is unreliable: use optimize in: optim(par = 0, fn = log.likelihood, y = Y.list$Y.3, T = 10, hessian = TRUE) Am I using the wrong approach here? Did I miss something from section? thanks for any help-- Maya

17 years, 2 months

[gov2001-l] Reading Assignment and Other Announcements

by jgrimmer＠fas.harvard.edu

Hey Everyone. To repeat the announcements from class, here is this week's reading assignment: (1) Review Chp 3 from Unifying Political Methodology, particularly the discussion of multivariate probability distributions. (2) Read pages 97-114 in Chapter 5 of UPM (3) Carefully read Gary King, Michael Tomz, and Jason Wittenberg. "Making the Most of Statistical Analyses: Improving Interpretation and Presentation" AJPS. 44, 2 (April 2000) pgs 347-361. (Available through J-STOR) Holger and I also want to encourage folks to attend office hours. Mine are Tuesdays from 3-5 pm and Holger's are Wednesdays from 4-6 pm. Finally, if you can, please attend section. These meetings will provide helpful hints on R code and clarify the points from class to make the homework and replication assignment easier. Cheers, Justin

17 years, 2 months

[gov2001-l] 1.3

by luciska＠gmail.com

Hi... Could someone give me lead on 1.3 in Assignment 4. Having trouble taking the derivative of the log likelihood function in the case of the binomial. Lucia

17 years, 2 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

Gov2001 March 2007