Dear Community: I am using SAS 9.4. I am working on a project where I estimate multiple pairwise tetrachoric/polychoric correlations between dichotomous variables. I then take those correlations and make a heatmap. However, the correlations are not very strong, and/or there are more strongly negative correlations than positive ones. What results is a heatmap whose color range scale is centered at -0.25 (i.e., the white or "neutral/null" color is centered at -0.25 instead of, ideally, at 0). This results in a heatmap that provides a "warmer" (pink) color indicating correlation=0. The result is a figure that is harder to interpret.  Here is an example of what this is looking like (I changed my variables' names for this post):   You can see that the the red/pink colors here are a little misleading, since they actually correspond to values close to 0 (no correlation). So my GOAL would be: change the color scale so that the scale is symmetric between [-0.5,0.5] so that the 0 is centered and corresponds to the null/neutral values. Or another approach could be to keep the asymmetric color range here (as shown above: [-1.0,0.5] but change the color values so that white is indicating 0 correlations.   I think I would need to use/alter some RANGEATTR type of variable, but I have little experience with this from my reading, and I was not able to find solutions with the syntax/code examples.   Here is my full code:   *NOTE (!!!): for Sex=Female, the color scale is centered on -0.25, so I will want to figure out how to recenter this; *SEX=Female; Title 'Correlation Matrix'; Title2 'Using PROC FREQ/plcorr Method'; Title3 'Sex = Female'; PROC FREQ Data=work.inpt_data; Where Sex=0; Tables (var1 var2 var3 var4 var5 var6 var7 var8)* (A B C D E F G H I J K L M N O P Q R S T U V W X Y Z AA) / plcorr; Ods output measures=mycorr_sexF_1 (where=(statistic="Tetrachoric Correlation" or statistic="Polychoric Correlation") keep = statistic table value ASE); RUN; *NOTE: correlation stats could not be calculated for some of the alphabet variables; *The ods output statement makes an output dataset called "mycorr_" that has the statistics from the Measures output and I specify that I only want the values for ether the polychoric or tetrachoric correlation and only want to keep certain variables (i.e., statistic, table, and value). But there is an additional 2 steps needed to make this into a matrix (See next)*; Title 'View the Tetrachoric Correlation Output';PROC PRINT Data=work.mycorr_sexF_1; RUN; DATA mycorr_sexF_2; *new dataset name working from the mycorr data made in the previous output statement*; Set mycorr_sexF_1; Group = floor((_n_ - 1)/8); *the number at the end has to be 8, since I have 8 variables for numbered variable overall that will serve to define what the groups are in this specific output*; x = scan(table, 2, " *"); y = scan(table, 3, " *"); Keep group value table x y; RUN; *want to see how the new dataset looks*; Title ''; PROC PRINT Data=work.mycorr_sexF_2; RUN; *nice, this makes a correlation matrix, but it is not in an easily-readable order, so I can transform the data using PROC TRANSPOSE... See next*; Title ''; PROC TRANSPOSE Data=work.mycorr_sexF_2 out=mymatrix_sexF_2(drop = _name_ group) ; Id x; By group; Var value; RUN; *this creates an output dataset that is formatted to look like a matrix, and this is called "mymatrix"... will see what it looks like*; Title ''; PROC PRINT Data=work.mymatrix_sexF_2; RUN; *this is fine, though I would have to color the cells myself in an output file. Likely better to try to get the heatmap to work. *PROC SGPLOT has the heatmap or heatmapparm statement to attempt a heatmap: will try this next: *See: https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/grstatproc/p0y94pnvxskvkmn19u7sx2j043lq.htm; Ods graphics on; Ods html sge=on; Ods trace on; Title '4/27/23: HeatMap: '; Title2 'Sex = Female'; PROC SGPLOT Data=work.mycorr_sexF_2; Heatmapparm x=x y=y colorresponse=value / outline nomissingcolor; *coloresponse statement tells SAS that the third variable is numeric value (correlation stat); Text x=x y=y text=value; *(***this adds the correlation values to the cells, but it looks messy without rescaling it in ods html file); RUN; Ods graphics off; Ods trace off;   What I would like to see happen is something like the following, which was generated from a different project with slightly different range of values: Sorry this looks really messy because I was altering it so much -- but you get the idea: the neutral (no correlation) is centered on 0, and this makes this figure much easier to interpret compared to the one above:   Thank you for any and all help!     ... View more