h1

About Me

My name is Christopher David Desjardins. I am a Ph.D. student studying quantitative methods in education in the department of educational psychology at the University of Minnesota. I am also pursing a M.S. in statistics from the University of Minnesota. I am a Minnesota Interdisciplinary Training in Education Research fellow. I am interested in Bayesian statistics, personal probabilities and causal inferences, longitudinal multilevel models, latent variable modeling (specifically structural equation modeling and factor analysis), free software, Debian GNU/Linux, R , and psychology.

Advertisements

17 comments

  1. I have installed Winbugs using your directions. It says that the licence expired in 2004.


  2. Dear Christopher Desjardins,

    I could not locate your email address so I hope you don’t mind the comment.

    I have created an evaluation news site that links together feeds from evaluation blogs across the web.

    As a starting point, I aggregated RSS/Atom feeds from the American Evaluation Association: Evaluation Blogs list http://www.eval.org/Resources/Blogs.asp Your blog was on this list, so it was included.

    Eval Central has just been set up and is not published to search engines. Additionally, there has been no attempt to advertise it’s existence. Before doing so, I am contacting administrators for all the blogs used during the development phase.

    The site has four goals:
    Create a destination news source for readers interested in evaluation.
    Increase traffic to all participating blogs, with little to no attached burden.
    Provide an opportunity for low post frequency bloggers to reach a larger audience.
    Provide an opportunity for evaluators without a blog to easily post to a site with an established audience.

    This is an academic venture on my part. I believe that if we can increase the frequency and diversity of content easily available to evaluators, and make it easier to post evaluation content for an established audience, we can increase the amount of high quality evaluation content available online.

    Here is the site: http://evalcentral.wordpress.com/


    • You’re welcome to add me. But it should be noted that I am not blogging very often and most of the posts that I do are about statistics.


  3. Can you point me to a good example (a paper or R code or both) of MCMCglmm used for modeling effects of hierarchical predictor variables on multinomial response variables?

    thanks in advance,
    pasky (pls. send response to my work e-mail)


    • Hi,
      Have you tried the CourseNotes vignette that comes with MCMCglmm? They are really good and worth examining.


  4. Christopher David Desjardins, i’m a student who study technology management, i want to measure the technology acceptance which my company’s develop but i have a problem with the number of sample, i only have 43 sample so i have been suggested to use bayesian sem, but i don’t understand much about statistic. i’m a newbie in bayesian sem, i try to write a code base on yours but i have a problem with my winbugs syntax…could you help me? which one that i used fix? i will very grateful for your help…

    model{
    for(i in 1:N){
    #model persamaan pengukuran

    for(j in 1:P){
    y[i,j]~dnorm(mu[i,j],psi [j]) I(thd [j,z[i,j]],thd[j,z[i,j]+1])
    ephat[i,j]<-y[i,j] -mu[i,j]

    }

    #faktor Budaya Organisasi
    mu[i,1]<-xi[i,1]
    mu[i,2]<-lam[1]*xi[i,1]
    mu[i,3]<-lam[2]*xi[i,1]

    #faktor Kemampuan Pengguna
    mu[i,4]<-xi[i,2]
    mu[i,5]<-lam[3]*xi[i,2]
    mu[i,6]<-lam[4]*xi[i,2]

    #faktor Mekanisme Dukungan
    mu[i,7]<-xi[i,3]
    mu[i,8]<-lam[5]*xi[i,3]
    mu[i,9]<-lam[6]*xi[i,3]

    #faktor Desain Antarmuka
    mu[i,10]<-xi[i,4]
    mu[i,11]<-lam[7]*xi[i,4]
    mu[i,12]<-lam[8]*xi[i,4]

    #faktor Persepsi Kualitas
    mu[i,13]<-xi[i,5]
    mu[i,14]<-lam[9]*xi[i,5]
    mu[i,15]<-lam[10]*xi[i,5]

    #faktor Persepsi Kemudahan Kegunaan
    mu[i,16]<-eta[i,1]
    mu[i,17]<-lam[11]*eta[i,1]
    mu[i,18]<-lam[12]*eta[i,1]

    #faktor Persepsi Kegunaan
    mu[i,19]<-eta[i,2]
    mu[i,20]<-lam[13]*eta[i,2]
    mu[i,21]<-lam[14]*eta[i,2]
    mu[i,22]<-lam[15]*eta[i,2]

    #faktor Sikap ke arah Penggunaan
    mu[i,23]<-eta[i,3]
    mu[i,24]<-lam[16]*eta[i,3]
    mu[i,25]<-lam[17]*eta[i,3]

    #faktor Persepsi Niat untuk Menggunakan
    mu[i,26]<-eta[i,4]
    mu[i,27]<-lam[18]*eta[i,4]
    mu[i,28]<-lam[19]*eta[i,4]

    #faktor Adopsi E-government
    mu[i,29]<-eta[i,5]
    mu[i,30]<-lam[20]*eta[i,5]

    #model persamaan struktural
    xi[i,1:5] ~dmnorm(u[1:5],phi[1:5,1:5])

    eta[i,1]~dnorm(nu[i,1],pskp)
    nu[i,1]<-gam[1]*xi[i,2]+gam[2]*xi[i,3]+gam[3]*xi[i,4]
    dthat[i,1]<-eta[i,1]-nu[i,1]

    eta[i,2]~dnorm(nu[i,2],pspk)
    nu[i,2]<-gam[4]*xi[i,1]+beta[1]*eta[i,1]
    dthat[i,2]<-eta[i,2]-nu[i,2]

    eta[i,3]~dnorm(nu[i,3],pssp)
    nu[i,3]<-beta[2]*eta[i,2]+beta[3]*eta[i,3]
    dthat[i,3]<-eta[i,3]-nu[i,3]

    eta[i,4]~dnorm(nu[i,4],psnm)
    nu[i,4]<-beta[4]*eta[i,1]+beta[5]*eta[i,2]+gam[5]*xi[i,5]
    dthat[i,4]<-eta[i,4]-nu[i,4]

    eta[i,5]~dnorm(nu[i,5],psae)
    nu[i,5]<-beta[6]*eta[i,4]
    dthat[i,5]<-eta[i,5]-nu[i,5]
    }#akhir dari i

    for (i in 1:5) {u[i]<-0.0}

    #lamda
    var.lam[1]<-8.0*psi[2] var.lam[2]<-8.0*psi[3]

    var.lam[3]<-8.0*psi[5] var.lam[4]<-8.0*psi[6]

    var.lam[5]<-8.0*psi[8] var.lam[6]<-8.0*psi[9]

    var.lam[7]<-8.0*psi[11] var.lam[8]<-8.0*psi[12]

    var.lam[9]<-8.0*psi[14] var.lam[10]<-8.0*psi[15]

    var.lam[11]<-8.0*psi[17] var.lam[12]<-8.0*psi[18] var.lam[13]<-8.0*psi[20]

    var.lam[14]<-8.0*psi[21] var.lam[15]<-8.0*psi[22]

    var.lam[16]<-8.0*psi[24] var.lam[17]<-8.0*psi[25]

    var.lam[18]<-8.0*psi[27] var.lam[19]<-8.0*psi[28]
    var.lam[20]<-8.0*psi[30]

    for (i in 1:20) {lam[i] ~dnorm(1,var.lam[i])}
    for (j in 1:P) {
    psi[j] ~dgamma(10,8)
    sgl[j]<-1/psi[j]
    }

    #gamma
    gam[1]~dnorm(0.4,var.pk)
    gam[2]~dnorm(0.4,var.kp)
    gam[3]~dnorm(0.4,var.kp)
    gam[4]~dnorm(0.4,var.kp)
    gam[5]~dnorm(0.4,var.nm)

    var.pk <-8.0*pspk pspk~dgamma(10,8) sgpk<-1/pspk
    var.kp <-8.0*pskp pskp~dgamma(10,8) sgkp<-1/pskp
    var.sp <-8.0*pssp pssp~dgamma(10,8) sgsp<-1/pssp
    var.nm <-8.0*psnm psnm~dgamma(10,8) sgnm<-1/psnm
    var.ae <-8.0*psae psae~dgamma(10,8) sgae<-1/psae

    #beta
    beta[1] ~dnorm(0.5,var.pk)
    beta[2] ~dnorm(0.5,var.sp)
    beta[3] ~dnorm(0.5,var.sp)
    beta[4] ~dnorm(0.5,var.nm)
    beta[5] ~dnorm(0.5,var.nm)
    beta[6] ~dnorm(0.5,var.ae)

    phi[1:5,1:5] ~dwish(R[1:5,1:5],21)
    phx[1:5,1:5]<-inverse(phi[1:5,1:5])

    }
    #end of model

    DATA
    list(N=43, P=30,
    R=structure(
    .Data=c(10,0,0,0,0,
    0,10,0,0,0,
    0,0,10,0,0,
    0,0,0,10,0,
    0,0,0,0,10
    ),
    .Dim=c(5,5)),
    thd=structure(
    .Data=c(-250,-1.99072046156042,-1.08241139430414,0.983052916910141,250,-250,-200,-1.47752529199845,0.452147411138078,250,-250,-200,-1.08241139430414,0.730448177619092,250,-250,-1.99072046156042,-0.585607161227169,1.08241139430414,250,-250,-1.99072046156042,-0.265404753825216,1.08241139430414,250,-250,-200,-0.656304990872144,0.892559673266593,250,-250,-200,-1.08241139430414,0.517723552818072,250,-250,-1.99072046156042,-1.47752529199845,0.585607161227169,250,-250,-200,-1.99072046156042,0.265404753825216,250,-250,-200,-1.67966118528897,0.656304990872144,250,-250,-1.99072046156042,-1.67966118528897,0.517723552818072,250,-250,-1.67966118528897,-0.892559673266593,0.730448177619092,250,-250,-200,-1.08241139430414,0.80884440410662,250,-250,-200,-1.67966118528897,0.730448177619092,250,-250,-1.67966118528897,-1.19379507272265,0.80884440410662,250,-250,-200,-1.32236537894944,0.98305291691014,250,-250,-200,-0.656304990872144,1.08241139430414,250,-250,-1.99072046156042,-1.32236537894944,0.585607161227169,250,-250,-200,-1.19379507272265,0.656304990872144,250,-250,-1.99072046156042,-1.67966118528897,0.80884440410662,250,-250,-1.99072046156042,-1.19379507272265,0.656304990872144,250,-250,-1.67966118528897,-1.19379507272265,0.452147411138078,250,-250,-200,-1.47752529199845,0.656304990872144,250,-250,-1.99072046156042,-1.47752529199845,0.585607161227169,250,-250,-200,-1.19379507272265,0.730448177619092,250,-250,-1.67966118528897,-0.80884440410662,0.892559673266593,250,-250,-1.99072046156042,-1.67966118528897,0.585607161227169,250,-250,-200,-0.983052916910141,0.80884440410662,250,-250,-200,-1.67966118528897,0.585607161227169,250,-250,-1.99072046156042,-1.08241139430414,0.892559673266593,250),
    .Dim=c(30,5)),
    z=structure(
    .Data=c(3,3,2,3,2,2,4,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,2,2,3,3,3,3,3,2,3,3,3,3,2,3,2,2,2,2,2,2,2,3,3,3,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,2,2,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,3,4,3,3,3,3,3,2,3,4,2,4,4,4,4,3,3,3,3,2,2,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,2,2,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,3,3,3,3,3,3,3,3,3,2,2,2,3,3,3,3,3,4,3,2,3,3,3,3,4,3,2,4,3,3,3,3,3,3,3,3,3,2,3,2,2,2,2,3,3,4,3,3,2,3,3,3,3,2,4,3,3,4,4,3,3,2,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,2,3,4,4,2,2,3,3,4,4,4,4,3,3,3,3,3,4,3,3,3,3,3,4,4,4,3,4,4,4,4,3,3,3,3,3,3,4,3,3,3,3,2,3,3,1,4,4,4,2,3,3,3,3,3,3,1,3,3,3,3,2,3,3,2,2,2,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,2,3,2,4,3,2,2,2,2,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,2,2,2,2,3,3,3,4,4,4,4,3,3,2,4,3,3,3,2,3,4,2,4,4,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,2,3,3,3,3,2,4,4,4,3,3,3,3,3,3,3,3,4,4,3,3,3,3,4,3,1,4,3,4,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,2,3,3,3,3,3,2,3,2,3,2,3,4,3,3,3,3,3,4,4,3,3,3,3,3,3,3,3,3,3,3,4,3,3,3,4,3,3,4,3,3,3,3,3,3,2,2,3,3,4,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,3,4,4,3,4,4,4,4,3,3,3,3,3,3,3,3,3,4,4,3,3,3,3,3,3,3,3,3,2,3,4,3,2,3,1,3,3,3,3,2,2,2,3,3,2,3,2,2,2,4,4,4,2,3,2,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,2,2,3,3,2,3,3,2,2,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,3,3,2,2,1,1,2,2,2,2,2,1,1,2,2,2,1,1,1,2,2,2,2,2,2,2,2,2,3,2,1,1,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,3,4,4,3,3,4,4,3,3,4,4,4,3,3,3,4,4,4,3,3,4,4,4,4,4,4,4,4,4,4,3,3,3,2,2,3,4,4,4,3,3,3,4,4,4,3,3,4,3,3,3,3,3,3,3,3,4,3,4,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,2,3,3,3,3,2,3,3,3,3,3,3,3,2,3,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,3,4,3,3,3,3,3,3,4,3,3,3,4,4,4,3,4,3,3,3,3,4,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,4,3,2,3,4,2,3,4,3,2,1,3,4,1,3,2,1,3,2,1,2,2,1,2,2,1,2,3,4,2,2,3,2,2,2,3,3,3,3,3,3,3,3,3,3,3,3,3,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,3,3,2,3,4,4,4,4,4,4,3,3,4,4,4,4,3,3,4,4,4,4,3,3,4,3,4,3
    ),
    .Dim=c(43,30)))
    )

    INITS
    list(
    lam=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
    psi=c(1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0),
    pspk=1.0, pskp=1.0, pssp=1.0, psnm=1.0, psae=1.0,
    gam=c(0,0,0,0,0),
    beta=c(0,0,0,0,0,0),
    phi=structure(.Data=c(1,0,0,0,0,
    0,1,0,0,0,
    0,0,1,0,0,
    0,0,0,1,0,
    0,0,0,0,1
    ),.Dim=c(5,5)),
    xi=structure(.Data=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
    .Dim=c(43,5)),
    eta=structure(.Data=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
    .Dim=c(43,5))
    )


  5. Hello Christopher,
    I managed to install Winbugs following you instruction,
    but I get the annoying black box trap #101 error.
    after typing
    wine ~/.wine/drive_c/Program Files/WinBUGS14/WinBUGS14.exe
    What I have missed? Thank you
    Anna


    • This could be caused by a different version of Wine. What version of Wine are you using? Also, may I suggest you consider either OpenBUGS or JAGS? Both of which are freely available for Linux?


  6. Dear Christopher Desjardins:

    I come to you because in a material you posted to your blog (
    Bayesian, MCMC, and Multilevel Modeling (a foray Into the Subjective)) at the end you say: “If you need help with syntax Please ask.”

    I explain my situation:
    I’m working on the evolution of the social structure of a particular insect. I want to evaluate whether the colonial size(independent variable, continous) of the species predicts the social structure(variable discrete with 4 states) in such insects. for this I fit a GLMM in the program MCMCglmm of R, since my response variable is discrete with 4 states. the problem lies is not good as well adjusted prior to model.

    the model I’m trying to implement is the following:

    model <- MCMCglmm (EstructuraSOC ~ Colonysize -1,
    random = ~ animal
    rcov = ~ us (trait): units,
    family = "categorical"
    Prior = list (R = list (fix = 1, V = 0.25 * (I + J), n = 3), G = list (G1 = list (V = diag (4), n = 4), G2 = list (V = diag (4), n = 4)))
    data = Data,
    pedigree= Apistree ,
    nitt = 600000,
    burnin = 100000,
    thin = 100)

    with I = J = 3 (since they are 4 categories). to implement this model throws me the following error:

    V is the wrong dimension for Some priorG / priorR elements.

    I have little experience in the use of R or less of MCMCglmm, so if you can help me in the structure of the model, as in the prior adjustment would be very nice.


    • Hi Oscar,
      Have you solved this problem? I have been away for a while. If you haven’t please let me know. Otherwise, R-Sig-Mixed-Models mailing list is a great place to post questions about MCMCglmm as the author hangs around there and answers questions. Best.
      Chris

      https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models


  7. Hi Chris,
    Do you have an email I could talk with you through? I have a couple of stupid questions about your profile analysis R program.


    • desja004 AT umn DOT edu

      I am not extremely knowledgeable about profile analysis but I’d be happy to help or add any thing that you think might be interesting (or if you found a bug).


  8. Hello, I’m a PhD at UW-Madison, and now dealing with a project that needs the idea of Zero-inflation Poisson with lasso penalty. Hope to see your wonderful work soon!


    • Hi,
      Your project sounds interesting. I’ll post soon about both my work with zero-inflated and overdispersed models and penalized regression.


  9. Dear Christopher Desjardins
    i’m MSc candidate of biostatistics and my thesis is about application of generalized bayesian structural equation in educational-health related data(using R and winbugs);
    i have one general question on bayesian analysis:
    what is the strategy of selecting prior distribution?(+say me some good reference about it)
    also i cant find your email adress;i have some special question about my real data)
    thanks


    • Hi,
      I would look at a book by Andrew Gelman or Brad Carlin regarding Bayesian analysis. Regarding a strategy for selecting priors: I would say, in general, you want to minimize their influence unless of course you are doing something like a meta-analysis or collating multiple sources. I think people tend to favor flat or non-informative priors.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: