I am having trouble manipulating the simulated effects of a circle graph. I've created a pseudo 3D plot using circle size as 3rd variable. The code I attached has simulated data values for all 3 variable, and is plotting a graph showing an interaction with all 3variables. I am trying to create extreme effects by having the circlesize increase, with increased X value, while holding the other variablesomewhat constant (and then do the other). Any suggestions?
_________________________________________________________________ Live.
# Simulation Paramter Settings N <- 5000 # Sample Size MX <- 10 # Mean of X SDX <- 2 # SD of indepdenent part of X MZ <- 10 # Mean of Z SDZ <- 2 # SD of independent part of Z Yerror <- 6 # SD of error in Y a <- 10 # See model above b <- 1 c <- .5 d <- .5 # Simulate X and Z tempvar <- rnorm(N,0,1) # shared X, Z variance X <- rnorm(N,MX, SDX) + tempvar Z <- rnorm(N,MZ, SDZ) + tempvar rm(tempvar) # Simulate Y Y <- a + (b * X) + (c * Z) + (d * X * Z) + rnorm(N,0,Yerror) XYZ <- data.frame(X,Z,Y) summary(XYZ) cov(XYZ) cor(XYZ) # General Plot Parameters proportion <- .05 # Proportion of N to be plotted plotN <- round(proportion * N) # Bubble Plot symbols(x = X[1:plotN], y = Z[1:plotN], circles = Y[1:plotN], xlab = 'X', ylab = 'Z', main = "Y by X and Z", sub = 'Point Size = Y') # Bubble Plot 2 symbols(x = X[1:plotN], y = Z[1:plotN], circles = Y[1:plotN], xlab = 'X', ylab = 'Z', fg='black', bg='grey', main = "Y by X and Z", sub = 'Point Size = Y') # Bubble Plot 3 plot(seq(min(X)-2,max(X)+2,len=10),seq(min(Z)-2,max(Z)+2,len=10),type='n',xlab='X',ylab='Z', main='Y by X and Z',sub='Point Size = Y') rad.ex <- .5 symbols(x = X[1:plotN], y = Z[1:plotN], circles = ((rad.ex*Y[1:plotN])/max(Y[1:plotN])), fg='black', bg='grey', add=T, inches=F) symbols(x = X[1:plotN], y = Z[1:plotN], circles = ((rad.ex*Y[1:plotN])/max(Y[1:plotN])), add=T, inches=F) # rad.ex serves as a scaling parameter for the expansion of the radii of the circles, in place of character expansion (cex). # The add parameter in the symbols() function adds to the existing plot. # The second symbols() function redraws the circles over the fill. # The inches parameter set to FALSE scales the radii to the axis scale. # In the study, it might be important to consider instructions on # reading the bubble plot. For example, everybody is taught that # if the lines cross in an ANOVA line plot, then there is an interaction. # Similarly, it may be important to instruct the participants that if # the comparative sizes of circles changes from one vertical (or horizontal) # line to another, then there is an interaction. # Instruction could either be a variable or a constant in the study. # At some point the code for the bubble plot will become sufficiently complex # that it will be easier to code it into a function that can be called with # various parameters to alter the appearance of the plot.
______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
