EDIT:  I now found that if I omit initial plant height ("initht") in the random statement, I do not have the issue with the residuals described below.  Hopefully, this can lend some insight. ---------------   Hi all,   I am fitting a model with PROC MIXED.     I have a continuous response variable "biomass" and 3 predictor variables 1) light (2 level categorical) 2) flood depth (ranked ordinal) and 3) flood duration (ranked ordinal), and I am fitting all 2 and 3 way interactions.   I have 2 random effects, initial plant height and experimental block.   Here is my model statement (flood depth is "sublevel").     PROC MIXED data=RUN1biomass covtest plots=(residualpanel); CLASS Block Light sublevel duration; MODEL AGB=Light Sublevel duration duration*Light SubLevel*Light SubLevel*Duration Sublevel*Light*Duration / influence DDFM=SATTERTH OUTP=work.diagnostics RESIDUAL; RANDOM Block initht; run;   The model runs fine but I don't understand the residuals.  I am attaching the residual panel produced by proc mixed.  Notice that ALL THE RESIDUALS ARE POSITIVE.  How is this even possible??  Is this some glitch or is there any way this actually makes some sense? The studentized residual panels and pearson residual plots look basically the same.  All of the conditional residual plots look good--normal distribution of residuals and random scatter around 0.     Thank you!!    And here is the output from proc mixed  ... View more

I just installed SAS 9.4, and have been using SAS 9.3 with success.  I just opened for the first time in 9.4 and received the following error message when opening SAS.  I also can not run any programs without getting that error, and can not get output as results.  ERROR: Trying to open a key that does not exist: ODS\GUI\RESULTS I will try reinstalling the program if necessary, but thought I would ask here first, since it takes quite a long time to install.  I have a student license through my school, and can not call SAS tech support without paying for it. I am running Windows 7, 64 bit, and using the 64 bit version of SAS 9.4.  Thanks, Meghan ... View more

Thank you to all who read my previous pos, but I unfortunately did not get a response.  After rethinking my problem, I decided that Canonical Correspondence Analysis or Multivariate Multiple Regression might fit the bill.  I am first going to try using Multivariate Multiple Regression, as I can (presumably) do this in GLIMMIX, where I would be able to include random effects (research site is a random effect in my design), and specify non-normal distributions (Poisson, as I am using count data). Briefly, my research question involves modeling how the density of 3 classes of wetland plants (my response variables) varies with plot hydrology (hydro) and leaf area index (LAI).  I want to be able to state results to answer these subquestions: -Does LAI, hydrology, and the interaction between these two factors affect the density of wetland plants? -Do the different classes of plants respond differently to these factors and if so, in what direction? My 2 questions at this point are 1) Can I model this question in GLIMMIX, using multivariate multiple regression (I believe I have to use the BYOBS option) ?  2) Will I be able to draw the conclusions I aim to, as stated above, by using this method? If someone can answer this question, I have more questions about the coding, but I figure I need to start with basics right now! Additionally, if you have other suggestions about how to model the data, please contribute! Meghan MORE INFO ABOUT MY DATA: I have 3 research sites Within each research site, I have multiple plots, each with counts of each class of wetland plant, and measures of plot hydrology and LAI.  ... View more

Thank you SAS community users once again for reading this! This is more of a general stats question regarding the best procedure to answer my question of interest.  I started to address it using poisson regression in proc GLIMMIX, but now am unsure about this decision. Experimental Design: I sampled shrubs (stem counts by species) in quadrats in 3 forests and measured environmental variables in each of those quadrats.  The environmental variable of primary interest the "wetness" of each quadrat. In order to make more ecological sense of the data, I aggregated the shrubs by Wetland Indicator Status (e.g. a measure of the known preference of species for wetland areas). What I want to know is how relationship between stem density and wetness differs between the different indicator groups.  That is, I want to be able to compare the responses of the shrub groups to these conditions.  My initial thought was to make Indicator Status a term in a poisson regression, where a significant interaction between "hydro" (the wetness measure) and "IND" the wetland indicator status of the shrub group  (hydro*ind in the model statement below) would indicate that the relationship between stem count and wetness differs between shrub groups with different indicator statuses.  Basically, I am trying to confirm what you would expect--that the density of shrubs with an affinity for wet soils increase as soil wetness increases, and those with a preference for drier areas decrease with increasing wetness.  There is then a third group of an invasive species that does not have an indicator status, so I am trying to see it its relationship with wetness is more like the wetland or upland species.  However, I am thinking that my approach to the analysis is problematic, because indicator status (IND) is a property of the DV and not an independent variable.  So if anyone can help me come up with a way of answering this question, it would help me immensely.  I do need a method that can handle non-normal data (count data, so it is has many zeroes), and random effects (forest site is a random effect). First attempt here: proc glimmix data=shrub maxopt=100; class site ind;  model count=  hydro   ind    ind*hydro   / dist=poisson link=log solution; random  site; covtest 'Global Test random effects' ZEROG / CL WALD; output out=glimmixout predicted=pred Pearson=PearsonRes; run; THANKS IN ADVANCE!! Sincerely, Meghan Langley aka. Ms. ABD ... View more

Hello all, In advance, please let me thank whoever reads this entire post.  I have been working very hard to understand and build this model, and stats are not my strength.  Any assistance you can provide in answering these few remaining questions I have will be greatly appreciated. I am using a mixed model poisson regression to model my data (because I have random effects), and have been working with another statistician to make sure that is the correct model.  However, I still have a few BASIC but not completely resolved questions that I am hoping someone can help me with. Mostly, I am not entirely certain what ALL of the model assumptions are.  I KNOW that I need to check for overdispersion, and have been using the negative binomial model when that is the case.  However, I am not sure what other assumptions I need to validate.  Namely: 1) Do the model residuals need to be normally distributed? 2) I have read that for mixed models in general, the random effects are assumed to be randomly distributed.  Is this true for poisson as well and how is this assessed in SAS (I am using Proc GLIMMIX)?  I am not necessarily asking for complete code, just a general approach. 3) Assumptions of equal variance:  Is this an assumption of this model ? (please read more detailed questions below)      a. Do I need to look for random scatter when I plot the overall model residuals against the linear predictor values.      b.  Do I need to look for random scatter when I plot the residuals against the individual predictor values (for each independent variable.      c.  If A and/or B are true, what can/should I do when I see a cone shaped pattern indicating homoscedasticity.   THANK YOU!! Meghan ... View more

Hello Sas Users (my second post today and ever!) I am overwhelmed by the number of available statistical procedures in sas, and am hoping for someone to nudge me in the right direction.  Stats are not my strength, so any attempt you could make to simplify your explanation would be appreciated. I am an ecologist.  My research question:  how do various environmental variables (such as light, soil moisture, etc) affect the density of particular tree seedlings (or saplings) So, basically, for each analysis, I have one response variable (seedling density) and multiple predictor variables. I ran this analysis in proc reg, specifying a backwards regression.  However, as the response variables are based on counts they are very non-normal (heavily skewed to right, because of many zeroes).  Additionally, many of my predictor variables are heavily skewed to the right or left.  I have tried various transformations of both the predictor and response variables to satisfy the assumptions of homoscedasticity and normality of residuals--and have given up on that approach. I am reading about a lot of different procedures now, but am just not sure which would be the best for me to start learning about and working with. My sample size is n=60. I also should add that scatter plots of individual predictors vs the dependent variables, do not suggest a particularly strong relationship with any one variable.  Any advice would be appreciated.  Also, please do not hesitate to ask if you would like me to supply you with more information. Thanks, Meghan ... View more