similar to: suggested addition to model.matrix

Hi, I recently ran into an inconsistency in the way model.matrix.default handles factor encoding for higher level interactions with categorical variables when the full hierarchy of effects is not present. Depending on which lower level interactions are specified, the factor encoding changes for a higher level interaction. Consider the following minimal reproducible example: -------------- >

Dear Spencer, I don't think that the problem of "converting a data frame into a model matrix" is well-defined, because there isn't a unique mapping from one to the other. In your example, you build the model matrix for the additive formula ~ a + b from the data frame matrix containing a and b, using "treatment" contrasts, but there are other possible formulas (e.g.,

Hello Tyler, I want to bring to your attention the following document: "What happens if you omit the main effect in a regression model with an interaction?" (https://stats.idre.ucla.edu/stata/faq/what-happens-if-you-omit-the-main-effect-in-a-regression-model-with-an-interaction). This gives a useful review of the problem. Your example is Case 2: a continuous and a categorical regressor.

Hello Tyler, Thank you for searching for, and finding, the basic description of the behavior of R in this matter. I think your example is in agreement with the book. But let me first note the following. You write: "F_j refers to a factor (variable) in a model and not a categorical factor". However: "a factor is a vector object used to specify a discrete classification"

Hello Tyler, I rephrase my previous mail, as follows: In your example, T_i = X1:X2:X3. Let F_j = X3. (The numerical variables X1 and X2 are not encoded at all.) Then T_{i(j)} = X1:X2, which in the example is dropped from the model. Hence the X3 in T_i must be encoded by dummy variables, as indeed it is. Arie On Thu, Nov 2, 2017 at 4:11 PM, Tyler <tylermw at gmail.com> wrote: > Hi

Hello Tyler, You write that you understand what I am saying. However, I am now at loss about what exactly is the problem with the behavior of R. Here is a script which reproduces your experiments with three variables (excluding the full model): m=expand.grid(X1=c(1,-1),X2=c(1,-1),X3=c("A","B","C")) model.matrix(~(X1+X2+X3)^3-X1:X3,data=m)

Hi, I need to build a function to generate one column for each level of a factor in the model matrix created on an arbitrary formula (instead of using the available contrasts options such as contr.treatment, contr.SAS, etc). My approach to this was first to use the built-in function for contr.treatment but changing the default value of the contrasts argument to FALSE (I named this function

Hello, I want to create a design matrix using R. Can you explain the code which creates the following please? I understand the first part. b=g1(?) does what? dd <- data.frame(a = gl(3,4), b = gl(4,1,12)) # balanced 2-way dd a b 1 1 1 2 1 2 3 1 3 4 1 4 5 2 1 6 2 2 7 2 3 8 2 4 9 3 1 10 3 2 11 3 3 12 3 4 I am using the tree dataset in R. I want to form a reparameterized design

Hi, I found that using different contrasts (e.g. contr.helmert vs. contr.treatment) will generate different fitted probabilities from multinomial logistic regression using multinom(); while the fitted probabilities from binary logistic regression seem to be the same. Why is that? and for multinomial logisitc regression, what contrast should be used? I guess it's helmert? here is an example

Dear experts: Is it possible to create a new function based on stats:::model.matrix.default so that an alternative factor coding is used when the function is called instead of the default factor coding? Basically, I'd like to reproduce the results in 'mat' below, without having to explicitly specify my desired factor coding (identity matrices) in the 'contrasts.arg'. dd

Dear R users, When fitting a lme() object (from the nlme library), is it possible to test interactions *before* main effects? As I understand, R conventionally re-orders all terms such that highest-order interactions come last - but I??d like to know if it??s possible (and sensible) to change this ordering of terms. I??ve tried the terms() command (from aov) but I don??t know if something

Dear R-users, Using the function model.matrix I noticed the following behaviour (example below): Using the formula "~ a + a:b" will give a matrix of the same dimension as using "~ a * b". In the first case there are additional columns for the interaction (to compensate for the missing main-effect). dd <- data.frame(a = gl(3,4), b = gl(4,1,12)) model.matrix(~ a*b, dd,

With the following example using contr.sum for both factors, > dd <- data.frame(a = gl(3,4), b = gl(4,1,12)) # balanced 2-way > model.matrix(~ a * b, dd, contrasts = list(a="contr.sum", b="contr.sum")) (Intercept) a1 a2 b1 b2 b3 a1:b1 a2:b1 a1:b2 a2:b2 a1:b3 a2:b3 1 1 1 0 1 0 0 1 0 0 0 0 0 2 1 1 0 0 1 0

On 6/23/05, RenE J.V. Bertin <rjvbertin at gmail.com> wrote: > Hello, > > I was just having a look at the aov function source code, and see that when the model used does not have an Error term, Helmert contrasts are imposed: > > if (is.null(indError)) { > ... > } > else { > opcons <- options("contrasts") >

Hi Folks, I've encountered something I hadn't been consciously aware of previously, and I'm wondering what the explanation might be. In (on another list) using R to demonstrate the difference between different contrasts in 'lm' I set up an example where Y is sampled from three different normal distributions according to the levels ("A","B","C")

Hi, perhaps this is a stupid question, but i need some help about Helmert contrasts in the Cox model. I have a survival data frame with an unordered factor `group' with levels 0 ... 5. Calculating the Cox model with Helmert contrasts, i expected that the first coefficient would be the same as if i had used treatment contrasts, but this is not true. I this a error in reasoning, or is it

Hi all, I have dataset with 2 independent variable, one (x1) is continuous, the other (x2) is a categorical variable with 2 levels. The dependent variable (y) is continuous. When I run linear regression y~x1*x2, I found that the p value for the continuous independent variable x1 changes when different contrasts was used (helmert vs. treatment), while the p values for the categorical x2 and

Hi everyone, Can someone explain me how to calculate SAS type III sums of squares in R? Not that I would like to use them, I know they are problematic. I would like to know how to calculate them in order to demonstrate that strange things happen when you use them (for a course for example). I know you can use drop1(lm(), test="F") but for an lm(y~A+B+A:B), type III SSQs are only

Is there any logical reason why glm prints out the labels of factor levels after variable names when baseline contrasts (contr.treatment) are used but the codes for the levels when mean contrasts (contr.sum) are used? Jim -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info",

This is my first post to this list so I suppose a quick intro is in order. I've been using SPLUS 2000 and R1.6.2 for just a couple of days, and love S already. I'm reading MASS and also John Fox's book - both have been very useful. My background in stat software was mainly SPSS (which I've never much liked - thanks heavens I've found S!), and Perl is my tool of choice for