Specific results from mixed models fit with lme

[matccarp]

I feel like this is a simple and stupid question but I cannot seem to find an answer that works.

How do I calculate specific statistical results in a mixed model analysis with 'lme()'?

Data:

year	Plot	SeedingDate	Rep	Treat	CalDays
2017	11	Sep	1	Controlled	161
2017	12	Sep	1	Deep Fraze	35
2017	13	Sep	1	Verticut	161
2017	14	Sep	1	Less Fraze	84
2017	15	Sep	1	Scalped	63
2017	21	Sep	2	Less Fraze	63
2017	22	Sep	2	Deep Fraze	49
2017	23	Sep	2	Verticut	84
2017	24	Sep	2	Scalped	84
2017	25	Sep	2	Controlled	84
2017	31	Sep	3	Scalped	35

Code:
setwd("/Users/mc1499/Documents/Thesis Measurements/2017 & 2018 Fall/GDD Rye")
dat<-read.csv("GDD Rye CSV.csv")
head(dat)
block<-as.factor(dat$Rep)
trt<-as.factor(dat$Treat)
seeddate<-as.factor(dat$SeedingDate)
yr<-as.factor(dat$year)
str(dat)

DIA0WAS<-lme(CalDays ~ Treat*SeedingDate,random = ~1|year/Rep,data=dat)
anova(DIA0WAS)
predictmeans(DIA0WAS,"Treat:SeedingDate",pairwise=TRUE)

I am trying to graph my data and the programs needs either % critical value or standard deviation, what code can I use to get that from my data?

[andresrcs]

Could you please turn this into a minimal REPRoducible EXample (reprex) for your issue?

This is what I get if I try to run your code as a reprex

dat <- data.frame(stringsAsFactors=FALSE,
                  year = c(2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017,
                           2017),
                  Plot = c(11, 12, 13, 14, 15, 21, 22, 23, 24, 25, 31),
                  SeedingDate = c("Sep", "Sep", "Sep", "Sep", "Sep", "Sep", "Sep", "Sep",
                                  "Sep", "Sep", "Sep"),
                  Rep = c(1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 3),
                  Treat = c("Controlled", "Deep Fraze", "Verticut", "Less Fraze",
                            "Scalped", "Less Fraze", "Deep Fraze", "Verticut",
                            "Scalped", "Controlled", "Scalped"),
                  CalDays = c(161, 35, 161, 84, 63, 63, 49, 84, 84, 84, 35)
)

library(nlme)

block<-as.factor(dat$Rep)
trt<-as.factor(dat$Treat)
seeddate<-as.factor(dat$SeedingDate)
yr<-as.factor(dat$year)
str(dat)
#> 'data.frame':    11 obs. of  6 variables:
#>  $ year       : num  2017 2017 2017 2017 2017 ...
#>  $ Plot       : num  11 12 13 14 15 21 22 23 24 25 ...
#>  $ SeedingDate: chr  "Sep" "Sep" "Sep" "Sep" ...
#>  $ Rep        : num  1 1 1 1 1 2 2 2 2 2 ...
#>  $ Treat      : chr  "Controlled" "Deep Fraze" "Verticut" "Less Fraze" ...
#>  $ CalDays    : num  161 35 161 84 63 63 49 84 84 84 ...

DIA0WAS<-lme(CalDays ~ Treat*SeedingDate, random = ~1|year/Rep, data=dat)
#> Error in `contrasts<-`(`*tmp*`, value = contr.funs[1 + isOF[nn]]): contrasts can be applied only to factors with 2 or more levels
anova(DIA0WAS)
#> Error in anova(DIA0WAS): objeto 'DIA0WAS' no encontrado
predictmeans(DIA0WAS,"Treat:SeedingDate",pairwise=TRUE)
#> Error in predictmeans(DIA0WAS, "Treat:SeedingDate", pairwise = TRUE): no se pudo encontrar la función "predictmeans"

^{Created on 2019-06-21 by the reprex package (v0.3.0)}

[matccarp]

Okay, I am not sure if this it correct because it keeps giving me an error for my "lme" and anova but that is not the case. Those are running fine I am just trying to find a code to get %CV or standard deviation.

setwd("/Users/mc1499/Documents/Thesis Measurements/2017 & 2018 Fall/GDD Rye")
dat<-read.csv("GDD Rye CSV.csv")
block<-as.factor(dat$Rep)
trt<-as.factor(dat$Treat)
seeddate<-as.factor(dat$SeedingDate)
yr<-as.factor(dat$year)

DIA0WAS<-lme(CalDays ~ Treat*SeedingDate,random = ~1|year/Rep,data=dat)
#> Error in lme(CalDays ~ Treat * SeedingDate, random = ~1 | year/Rep, data = dat): could not find function "lme"
anova(DIA0WAS)
#> Error in anova(DIA0WAS): object 'DIA0WAS' not found

[andresrcs]

@matccarp

I am not sure if this it correct because it keeps giving me an error for my "lme" and anova but that is not the case.

You are getting those erros because you are not including library calls and sample data in your example, if you want to make a proper reprex, please read the content of the link I gave you before and try to follow along.

[matccarp]

I am sorry. I did, I just didn't catch that.

library(nlme)
setwd("/Users/mc1499/Documents/Thesis Measurements/2017 & 2018 Fall/GDD Rye")
dat<-read.csv("GDD Rye CSV.csv")
block<-as.factor(dat$Rep)
trt<-as.factor(dat$Treat)
seeddate<-as.factor(dat$SeedingDate)
yr<-as.factor(dat$year)

DIA0WAS<-lme(CalDays ~ Treat*SeedingDate,random = ~1|year/Rep,data=dat)
anova(DIA0WAS)
#> numDF denDF F-value p-value
#> (Intercept) 1 98 203.00155 <.0001
#> Treat 4 98 36.74186 <.0001
#> SeedingDate 2 98 24.95223 <.0001
#> Treat:SeedingDate 8 98 5.09245 <.0001

[matccarp]

I feel like this is a simple and stupid question but I cannot seem to find an answer that works.

How do I calculate specific statistical results in a mixed model analysis with 'lme()'?

Data:

year	Plot	SeedingDate	Rep	Treat	CalDays
2017	11	Sep	1	Controlled	161
2017	12	Sep	1	Deep Fraze	35
2017	13	Sep	1	Verticut	161
2017	14	Sep	1	Less Fraze	84
2017	15	Sep	1	Scalped	63
2017	21	Sep	2	Less Fraze	63
2017	22	Sep	2	Deep Fraze	49
2017	23	Sep	2	Verticut	84
2017	24	Sep	2	Scalped	84
2017	25	Sep	2	Controlled	84
2017	31	Sep	3	Scalped	35

Code:
library(nlme)
setwd("/Users/mc1499/Documents/Thesis Measurements/2017 & 2018 Fall/GDD Rye")
dat<-read.csv("GDD Rye CSV.csv")
block<-as.factor(dat$Rep)
trt<-as.factor(dat$Treat)
seeddate<-as.factor(dat$SeedingDate)
yr<-as.factor(dat$year)

DIA0WAS<-lme(CalDays ~ Treat*SeedingDate,random = ~1|year/Rep,data=dat)
anova(DIA0WAS)

I am trying to graph my data and the program needs either % critical value or standard deviation, but anova does not give me those.
What code can I use to get that from my data?

[matccarp]

What am I not doing right?

Obviously, being that you deleted/merged my post with the same post from 3 days ago that no one has responded to, I am again messing up something.

[andresrcs]

You are not supposed to duplicate open topics even if the previous one doesn't have an answer, if you want to attract attention towards your topic try to make a proper reproducible example to make it easier for responders to help you, take a look to this FAQ about how to make good R related questions.

FAQ: Tips for writing R-related questions

First and foremost, please ask! A core goal of community is to be a friendly place to chat about topics related to data science, R, and RStudio. We know that posting to technical forums can be intimidating. But know that many here would love to see you overcome your inhibition and engage with us. Here are a few tips some folks here think might be helpful. Before you post Check Out R Documentation - R has built in documentation on packages and functions. For example typing ?lm into your R console will open the documentation on the lm function. Search - Be sure to search for the basic keywords of your question with your favorite search engine, Stack Overflow, and R-Help. Stack Overflo…

[matccarp]

Version:1.0 StartHTML:0000000107 EndHTML:0000003814 StartFragment:0000000127 EndFragment:0000003796

head(dat, 5)[, c('Treat', 'CalDays', 'SeedingDate')]
#> Error in head(dat, 5): object 'dat' not found
library(nlme)
  df <- data.frame(Treat = c('Controlled', 'Deep Fraze', 'Verticut', 'Less Fraze', 'Scalped'), SeedingDate = c('Sep', 'Sep', 'Sep', 'Sep', 'Sep'), CalDays = c(161, 35, 161, 84, 63))
  DIA0WAS<-lme(data = df, CalDays ~ Treat*SeedingDate)
#> Error in getGroups.data.frame(dataMix, groups): invalid formula for groups
  anova(DIA0WAS)
#> Error in anova(DIA0WAS): object 'DIA0WAS' not found

Okay, this has all kinds of errors and I understand that. I have worked on it all morning and can not get it to not give me an error when putting it in this format. The errors are not the problem I can run the code fine with the code I am using. All I need is an additional code for a way to get % critical value or standard deviation.

I feel like it is something as simply as "summary(dat)" that is why I feel that the errors are not a big deal. Does anyone have a simple code to get either of these value for your data?

Thank you.