R packages effects & plm : "error in contrasts" when trying to plot marginal effects
Asked Answered
D

0

2

After reading this answer on error in contrasts and reviewing my data I am still stuck with a problem when trying to combine packages 'plm' and 'effects'. This might be impossible as John Fox does not discuss this possibility in his effects document (link is not allowed apparently - Google: "john fox effects package" if you want to have a look). So if it is indeed impossible please tell me.

I am running a simple regression on a reduced data set

library(plm) ; library(effects) 
shortdata<-plm.data(shortdata,index=c("ID","Year"))
MESS<-plm(paci_to_t ~ paco_to_t + cddom + cddom2,data=shortdata,model="within")
Effect(c('cddom2'),MESS)

and I get the following error (and the typical plm warning message)

Error in `contrasts<-`(`*tmp*`, value = contr.funs[1 + isOF[nn]]) : 
contrasts can be applied only to factors with 2 or more levels

I know it's still a pretty long data structure but I reckon it's worth sharing

 > dput(shortdata)
structure(list(ID = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 
6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L), .Label = c("1", "2", "4", "5", 
"6", "7"), class = "factor"), Year = structure(c(1L, 2L, 3L, 
4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 1L, 
2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 
16L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 
14L, 15L, 16L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 
12L, 13L, 14L, 15L, 16L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 
10L, 11L, 12L, 13L, 14L, 15L, 16L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L), .Label = c("1991", 
"1992", "1993", "1994", "1995", "1996", "1997", "1998", "1999", 
"2000", "2001", "2002", "2003", "2004", "2005", "2006"), class = "factor"), 
    paco_to_t = c(47L, 45L, 50L, 61L, 113L, 76L, 60L, 70L, 75L, 
    70L, 52L, 39L, 29L, 34L, 12L, 12L, 1L, 4L, 6L, 34L, 42L, 
    25L, 38L, 34L, 30L, 27L, 32L, 27L, 7L, 5L, 0L, 0L, 14L, 33L, 
    38L, 36L, 33L, 24L, 22L, 25L, 28L, 65L, 59L, 44L, 49L, 43L, 
    27L, 15L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 2L, 3L, 13L, 22L, 
    11L, 12L, 2L, 1L, 3L, 7L, 11L, 31L, 53L, 24L, 28L, 29L, 22L, 
    38L, 41L, 47L, 40L, 30L, 21L, 8L, 9L, 14L, 59L, 79L, 129L, 
    78L, 91L, 112L, 86L, 104L, 101L, 140L, 83L, 35L, 17L, 13L
    ), paci_to_t = c(535L, 641L, 582L, 634L, 871L, 392L, 421L, 
    285L, 223L, 183L, 52L, 24L, 1L, 5L, 1L, 0L, 2L, 22L, 25L, 
    124L, 216L, 70L, 115L, 78L, 46L, 32L, 14L, 10L, 2L, 1L, 0L, 
    0L, 683L, 626L, 1520L, 385L, 514L, 226L, 166L, 85L, 169L, 
    391L, 160L, 85L, 43L, 23L, 6L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 7L, 13L, 5L, 0L, 0L, 0L, 21L, 120L, 186L, 
    187L, 519L, 113L, 200L, 110L, 41L, 43L, 20L, 28L, 12L, 2L, 
    0L, 0L, 143L, 73L, 368L, 708L, 920L, 533L, 583L, 412L, 256L, 
    212L, 147L, 177L, 84L, 8L, 0L, 0L), cddom = structure(c(4.22657266700715, 
    7.07828323739468, 5.58155937520987, 6.09945741088926, 4.98990473760187, 
    13.1975509132969, 5.60581022431812, 6.4115834088139, 3.19372115108478, 
    3.47380227675641, 4.13404916693487, 2.97544267158703, 2.54471629085273, 
    2.64814971815201, 1.28583544617808, 0.437627780850619, 0.136363636363636, 
    0.164922480620155, 0.981640399790555, 1.61119564479727, 4.27951983102512, 
    4.52440902710094, 2.32837056711187, 6.18186426561852, 4.0123555170895, 
    4.22609174210518, 3.95801989660183, 4.61957981110804, 3.77282177376164, 
    1.66791871414077, 0.546721420916684, 0.265624433322453, 2.41282572727806, 
    2.77748331046807, 7.00223921984389, 3.55837337174436, 5.31590575343992, 
    2.61930006177923, 2.31060371552517, 2.57360416944421, 1.84395565608465, 
    2.92626684365294, 3.61419385672379, 1.80034349383174, 1.50187312925546, 
    1.11122165051967, 0.997854809226777, 0.386006521136941, 0, 
    0, 0, 0, 0, 0, 0, 0, 0.0476190476190476, 0.056547619047619, 
    0.0696251163521699, 0.385345494115747, 0.451639377056987, 
    0.33059260900224, 0.347544282905224, 0.0138314176245211, 
    0.211351052048726, 0.525647451963241, 0.696630753538187, 
    0.666082288178836, 1.09492110512526, 3.8035303566375, 1.00915595031924, 
    0.816810004384515, 0.530090984042535, 0.554438577055268, 
    1.08116003643161, 0.467985507058757, 0.512281275398252, 0.833471981363938, 
    0.506558172321601, 0.498777213205938, 0.548336215316966, 
    0.85237681730237, 0.982747572848003, 6.59758768791534, 8.74094242997363, 
    9.29913186611362, 6.85204239723805, 7.22544455545639, 7.55800542543194, 
    4.85857232261679, 3.3323740421857, 3.49663263434836, 4.73578798222752, 
    3.12245695547519, 1.39134774182295, 0.676122574837844), .Dim = 96L, .Dimnames = list(
        c("1990_ABT", "1991_ABT", "1992_ABT", "1993_ABT", "1994_ABT", 
        "1995_ABT", "1996_ABT", "1997_ABT", "1998_ABT", "1999_ABT", 
        "2000_ABT", "2001_ABT", "2002_ABT", "2003_ABT", "2004_ABT", 
        "2005_ABT", "1990_AKN", "1991_AKN", "1992_AKN", "1993_AKN", 
        "1994_AKN", "1995_AKN", "1996_AKN", "1997_AKN", "1998_AKN", 
        "1999_AKN", "2000_AKN", "2001_AKN", "2002_AKN", "2003_AKN", 
        "2004_AKN", "2005_AKN", "1990_ALL", "1991_ALL", "1992_ALL", 
        "1993_ALL", "1994_ALL", "1995_ALL", "1996_ALL", "1997_ALL", 
        "1998_ALL", "1999_ALL", "2000_ALL", "2001_ALL", "2002_ALL", 
        "2003_ALL", "2004_ALL", "2005_ALL", "1990_ABT", "1990_ABT", 
        "1990_ABT", "1990_ABT", "1990_ABT", "1990_ABT", "1990_ABT", 
        "1990_ABT", "1998_ALT", "1999_ALT", "2000_ALT", "2001_ALT", 
        "2002_ALT", "2003_ALT", "2004_ALT", "2005_ALT", "1990_AMG", 
        "1991_AMG", "1992_AMG", "1993_AMG", "1994_AMG", "1995_AMG", 
        "1996_AMG", "1997_AMG", "1998_AMG", "1999_AMG", "2000_AMG", 
        "2001_AMG", "2002_AMG", "2003_AMG", "2004_AMG", "2005_AMG", 
        "1990_AZN", "1991_AZN", "1992_AZN", "1993_AZN", "1994_AZN", 
        "1995_AZN", "1996_AZN", "1997_AZN", "1998_AZN", "1999_AZN", 
        "2000_AZN", "2001_AZN", "2002_AZN", "2003_AZN", "2004_AZN", 
        "2005_AZN"))), cddom2 = structure(c(17.8639165094919, 
    50.1020935887824, 31.1538050589932, 37.2033807072519, 24.8991492903416, 
    174.175350109063, 31.4251082710696, 41.1084018081777, 10.1998547908863, 
    12.067302257998, 17.0903625146349, 8.85325909190098, 6.47558100093125, 
    7.01269692974855, 1.65337279464797, 0.191518074572238, 0.018595041322314, 
    0.0271994246139054, 0.96361787450096, 2.5959514058137, 18.3142899841373, 
    20.4702770445124, 5.42130949779285, 38.2154457985312, 16.0989967955186, 
    17.8598514126896, 15.665921501896, 21.340517631197, 14.2341841365699, 
    2.78195283698102, 0.298904312089158, 0.0705563395778743, 
    5.8217279902149, 7.71441353992869, 49.03135409192, 12.6620210527394, 
    28.2588539794557, 6.86073281363667, 5.33888953019871, 6.6234384209806, 
    3.40017246160656, 8.56303764026254, 13.06239723398, 3.24123669578228, 
    2.2556228963796, 1.23481355658365, 0.995714220297008, 0.149001034360244, 
    0, 0, 0, 0, 0, 0, 0, 0, 0.00226757369614512, 0.00319763321995465, 
    0.0048476568270532, 0.148491149835309, 0.203978126908423, 
    0.109291473126908, 0.120787028580106, 0.000191308113503912, 
    0.0446692672021035, 0.276305243755448, 0.485294406775182, 
    0.443665614625554, 1.19885222644872, 14.466843173863, 1.01839573206472, 
    0.667178583262631, 0.280996451363183, 0.30740213572707, 1.16890702437679, 
    0.219010434817042, 0.26243210512366, 0.694675543718729, 0.2566011819458, 
    0.248778708413482, 0.300672605028135, 0.726546238674519, 
    0.965792791938642, 43.528163299732, 76.4040745641133, 86.4738534633698, 
    46.9504850135478, 52.2070490239743, 57.1234460108586, 23.6057250140979, 
    11.104716757033, 12.2264397795899, 22.4276878126106, 9.74973743879538, 
    1.93584853867582, 0.457141736205356), .Dim = 96L, .Dimnames = list(
        c("1990_ABT", "1991_ABT", "1992_ABT", "1993_ABT", "1994_ABT", 
        "1995_ABT", "1996_ABT", "1997_ABT", "1998_ABT", "1999_ABT", 
        "2000_ABT", "2001_ABT", "2002_ABT", "2003_ABT", "2004_ABT", 
        "2005_ABT", "1990_AKN", "1991_AKN", "1992_AKN", "1993_AKN", 
        "1994_AKN", "1995_AKN", "1996_AKN", "1997_AKN", "1998_AKN", 
        "1999_AKN", "2000_AKN", "2001_AKN", "2002_AKN", "2003_AKN", 
        "2004_AKN", "2005_AKN", "1990_ALL", "1991_ALL", "1992_ALL", 
        "1993_ALL", "1994_ALL", "1995_ALL", "1996_ALL", "1997_ALL", 
        "1998_ALL", "1999_ALL", "2000_ALL", "2001_ALL", "2002_ALL", 
        "2003_ALL", "2004_ALL", "2005_ALL", "1990_ABT", "1990_ABT", 
        "1990_ABT", "1990_ABT", "1990_ABT", "1990_ABT", "1990_ABT", 
        "1990_ABT", "1998_ALT", "1999_ALT", "2000_ALT", "2001_ALT", 
        "2002_ALT", "2003_ALT", "2004_ALT", "2005_ALT", "1990_AMG", 
        "1991_AMG", "1992_AMG", "1993_AMG", "1994_AMG", "1995_AMG", 
        "1996_AMG", "1997_AMG", "1998_AMG", "1999_AMG", "2000_AMG", 
        "2001_AMG", "2002_AMG", "2003_AMG", "2004_AMG", "2005_AMG", 
        "1990_AZN", "1991_AZN", "1992_AZN", "1993_AZN", "1994_AZN", 
        "1995_AZN", "1996_AZN", "1997_AZN", "1998_AZN", "1999_AZN", 
        "2000_AZN", "2001_AZN", "2002_AZN", "2003_AZN", "2004_AZN", 
        "2005_AZN")), class = "AsIs")), .Names = c("ID", "Year", 
"paco_to_t", "paci_to_t", "cddom", "cddom2"), class = c("plm.dim", 
"data.frame"), row.names = c(1L, 2L, 3L, 4L, 160L, 5L, 162L, 
163L, 164L, 6L, 166L, 7L, 8L, 9L, 170L, 171L, 172L, 173L, 174L, 
175L, 176L, 177L, 178L, 179L, 180L, 181L, 182L, 183L, 184L, 10L, 
186L, 11L, 188L, 12L, 190L, 191L, 192L, 13L, 194L, 14L, 196L, 
197L, 15L, 199L, 200L, 16L, 202L, 17L, 157L, 400L, 401L, 583L, 
524L, 565L, 526L, 527L, 204L, 205L, 206L, 207L, 208L, 209L, 210L, 
211L, 212L, 213L, 214L, 215L, 216L, 217L, 218L, 18L, 19L, 221L, 
222L, 223L, 224L, 20L, 226L, 21L, 22L, 23L, 230L, 231L, 232L, 
233L, 234L, 235L, 236L, 237L, 238L, 239L, 24L, 25L, 26L, 27L))

Now to the best of my ability, I do not see a factor with less than 2 levels so I don't understand where this error comes from. My eventual goal for using the effects package is to be able to plot the marginal effect of cddom (and cddom2 which is the squared version of cddom) on my response variable. Of course in my full data set I have about 15 more controls, none of which are factors with a single level, but I keep getting the same problem.

I hope someone can advise!

EDIT This problem with the effect function changes sometimes. e.g.

Effect(c("cddom","cddom2"),MESS)

gives the following error

Error in mod.matrix %*% mod$coefficients[!is.na(coef(mod))] : non-conformable arguments

Thanks in advance

Simon

Deshabille answered 17/4, 2014 at 11:23 Comment(4)
Looking at your dataset, cddom2 is numeric not a factor.Difficult
I know, none of my data are factors. But why would a function to plot the marginal effects of something be constrained to only plot marginal effects of factors? That does not make any sense to me. Regardless, I guess any value can be coerced to a factor so I do not think that causes the problem.Deshabille
Have you tried making cddom2 a binary factor?Difficult
No I have not tried that. Even if that would work it would not really help me because the marginal effects for a binary factor are not very interesting. I tried running the basic regression in MESS with a variety of explanatory variables and a problem with effect always persists, regardless of the kind of variable I'm using (see edit).Deshabille

© 2022 - 2024 — McMap. All rights reserved.