How to obtain individual slope coefficients for a growth curve model

[ed_coder]

I am trying to obtain individual slope coefficients in a growth curve model I am running on RStudio.

I'm using the lme4 package.

base <- lmer(X1 ~ Week + (Week | id),
data=Data)

beta <- coef(base)$id
summary(beta)

When I run this code, instead of extracting individual slopes the 'beta' variable shows:
Length = 0, Class = Null, Mode = Null.

What am I doing incorrectly to not get the individual slope coefficients?

[andresrcs]

Is this what you are trying to do?
(just in case, I'm using iris dataset as example data)

library(lme4)
base <- lmer(Sepal.Length ~ Sepal.Width + (Sepal.Width | Species),
             data=iris)
#> singular fit
summary(base)
#> Linear mixed model fit by REML ['lmerMod']
#> Formula: Sepal.Length ~ Sepal.Width + (Sepal.Width | Species)
#>    Data: iris
#> 
#> REML criterion at convergence: 193.7
#> 
#> Scaled residuals: 
#>     Min      1Q  Median      3Q     Max 
#> -2.8728 -0.5932 -0.0753  0.4327  3.3306 
#> 
#> Random effects:
#>  Groups   Name        Variance Std.Dev. Corr
#>  Species  (Intercept) 0.39789  0.6308       
#>           Sepal.Width 0.01394  0.1181   1.00
#>  Residual             0.19074  0.4367       
#> Number of obs: 150, groups:  Species, 3
#> 
#> Fixed effects:
#>             Estimate Std. Error t value
#> (Intercept)   3.3805     0.4891   6.912
#> Sepal.Width   0.8129     0.1264   6.431
#> 
#> Correlation of Fixed Effects:
#>             (Intr)
#> Sepal.Width -0.157
#> convergence code: 0
#> singular fit
base@beta
#> [1] 3.3804685 0.8129194

^{Created on 2019-01-28 by the reprex package (v0.2.1)}

[ed_coder]

Well yes and no. The first part of my code works well. I get the fixed and random effect coefficients for the group but I'm unable to get slope values for each individual participant in the sample.

[andresrcs]

It's hard to help you any further if you don't provide a reproducible example, we have no way to know the structure of your real data.

FAQ: What's a reproducible example (`reprex`) and how do I create one?

Why reprex? Getting unstuck is hard. Your first step here is usually to create a reprex, or reproducible example. The goal of a reprex is to package your code, and information about your problem so that others can run it and feel your pain. Then, hopefully, folks can more easily provide a solution. What's in a Reproducible Example? Parts of a reproducible example: background information - Describe what you are trying to do. What have you already done? complete set up - include any library() calls and data to reproduce your issue. data for a reprex: Here's a discussion on setting up data for a reprex make it run - include the minimal code required to reproduce your error on the data…

[ed_coder]

library(foreign)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library (lme4)
#> Loading required package: Matrix
CBMY <- read.spss("/Users/jd43652/Desktop/Test.sav", 
                   use.value.labels = TRUE, to.data.frame = TRUE, sep = "\t", dec = ".")
colSums (is.na(CBMY))
#>       sid      Time TestScore 
#>         0         0        66
CBM <- na.omit(CBMY)
base <- lmer(CBM$TestScore ~ 1 + CBM$Time + (1 + CBM$Time | CBM$sid),
               data=CBM, REML=FALSE)
#> singular fit
summary(base)
#> Linear mixed model fit by maximum likelihood  ['lmerMod']
#> Formula: CBM$TestScore ~ 1 + CBM$Time + (1 + CBM$Time | CBM$sid)
#>    Data: CBM
#> 
#>      AIC      BIC   logLik deviance df.resid 
#>   1317.9   1343.6   -652.9   1305.9      537 
#> 
#> Scaled residuals: 
#>      Min       1Q   Median       3Q      Max 
#> -3.07945 -0.61070 -0.01133  0.55039  2.83987 
#> 
#> Random effects:
#>  Groups   Name        Variance  Std.Dev. Corr
#>  CBM$sid  (Intercept) 0.3029561 0.55041      
#>           CBM$Time    0.0004888 0.02211  1.00
#>  Residual             0.4911642 0.70083      
#> Number of obs: 543, groups:  CBM$sid, 80
#> 
#> Fixed effects:
#>             Estimate Std. Error t value
#> (Intercept) -0.59942    0.09085  -6.598
#> CBM$Time     0.15074    0.01542   9.776
#> 
#> Correlation of Fixed Effects:
#>          (Intr)
#> CBM$Time -0.538
#> convergence code: 0
#> singular fit
beta <- coef(base)$sid
summary(beta)
#> Length  Class   Mode 
#>      0   NULL   NULL

[andresrcs]

@ed_coder

"/Users/jd43652/Desktop/Test.sav"

Still not reproducible since we don't have access to this file

[ed_coder]

Here you go:
https://utexas.box.com/s/942slnbi9qfu2q38r6my0ckdoyrwq4y2

[ed_coder]

Is the link helpful in accessing the data?

[andresrcs]

Yes, here is a minimal example solution with just the first 20 rows of your data.

library(lme4)
#> Loading required package: Matrix

CBM <- data.frame(
    sid = c(210103, 210103, 210103, 210103, 210103, 210103, 210103,
            210105, 210105, 210105, 210105, 210105, 210105, 210105,
            210202, 210202, 210202, 210202, 210202, 210202),
    Time = c(1, 2, 3, 4, 5, 6, 7, 1, 2, 3, 4, 5, 6, 7, 1, 2, 3, 4, 5, 6),
    TestScore = c(-1.45496720585195, -0.260188842483168, -1.45496720585195,
                  -1.75366179669414, -1.45496720585195, -0.857578024167559,
                  -1.45496720585195, -1.75366179669414, 0.337200339201222,
                  -0.857578024167559, -1.15627261500975, -0.558883433325363,
                  0.0385057483590271, -0.857578024167559, -0.857578024167559, 1.23328411172781,
                  -1.15627261500975, -1.45496720585195, -1.15627261500975,
                  0.635894930043418)
)
base <- lmer(CBM$TestScore ~ 1 + CBM$Time + (1 + CBM$Time | CBM$sid),
             data=head(CBM,20), REML=FALSE)
#> singular fit
coef(base)$`CBM$sid`$`(Intercept)`
#> [1] -0.8681693 -0.8129559 -0.8007325
coef(base)$`CBM$sid`$`CBM$Time`
#> [1] -0.004295916  0.007774753  0.010447016

^{Created on 2019-02-02 by the reprex package (v0.2.1)}

[ed_coder]

@andresrcs

coef(base)$CBM$sid$(Intercept)

Thank you so much for helping me fix this problem. I was able to obtain individual participants intercepts and slopes. For future reference, can you tell me why my initial individual slope output was "NULL"? Was it because I did not add `(Intercept)` and CBM$Time to my code?

[andresrcs]

Yes, because those are the column names of the output

coef(base)
#> $`CBM$sid`
#>        (Intercept)     CBM$Time
#> 210103  -0.8681693 -0.004295916
#> 210105  -0.8129559  0.007774753
#> 210202  -0.8007325  0.010447016
#> 
#> attr(,"class")
#> [1] "coef.mer"