R: Problem with loop showing variable names

Question

I want to create a loop using variable names instead of numbers but I'm struggling with it.

I have over 1000 variables in my data but the structure looks like this:

#Reproducible data
id <- rep(c("1","2","3","4","5","6"),3)
sequence <- rep(c("1","2","1","2","1","1"),3)
treatment <- c(rep(c("A"), 6), rep(c("B"), 6),rep(c("C"), 6))
var1 <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
var1_base <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
var2 <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
var2_base <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
var3 <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
var3_base <- c(rnorm(3, 1, 0.4), rnorm(3, 3, 0.5), rnorm(3, 6, 0.8), rnorm(3, 1.1, 0.4), rnorm(3, 0.8, 0.2), rnorm(3, 1, 0.6))
DF <- data.frame(id,sequence,treatment, var1, var2, var3, var1_base, var2_base, var3_base) %>%
  mutate(id = factor(id),
         sequence = factor(sequence),
         treatment = factor(treatment, levels = c("A","B","C")))

> head(DF)
  id sequence treatment      var1      var2      var3 var1_base var2_base var3_base
1  1        1         A 0.5488589 1.3045888 0.2367363 1.2646227 1.2241417 0.1968524
2  2        2         A 1.0201801 1.3480361 0.9944096 0.3625067 0.8987885 1.5868442
3  3        1         A 0.7269204 0.7091029 1.2025266 0.1238612 1.8828400 0.8687552
4  4        2         A 3.3240269 3.3133104 3.2251780 2.4116230 2.6284785 2.6027341
5  5        1         A 3.3051822 2.4542786 2.1687379 3.5250026 3.2231797 2.9990167
6  6        1         A 2.7436715 2.7419527 3.8349072 2.9971485 3.0528477 2.6970430

I want to create a linear mixed model with var as the outcome; treatment, var_base (baseline), and sequence as the fixed effect; id as a random effect.

To code it one by one, it would look like this:

lm1 <- lmer(var1 ~ var1_base + treatment + sequence + (1|id), data = DF)

But since I have over 1000 vars, it wouldn't make sense to do it individually. I tried writing for loop but did not turn out to be what I expected.

#Approaches 1--it worked but I want the result to show "var" instead of "[[1]]"

lm_output <- list()

for(i in 4:6){
  lm1 <-lmer(DF[[i+3]] ~ DF[[i]] + Treatment+  sequence + (1|id), data = DF)
  summary(lm1)
  lm_output[[i]] <- summary(lm1)
}
>print(lm_output[1:6])

[[1]]
NULL

[[2]]
NULL

[[3]]
NULL

[[4]]
Fixed effects:
            Estimate Std. Error      df t value Pr(>|t|)   
(Intercept)   0.8995     0.6129 13.0000   1.468  0.16598   
DF[[i]]       0.6772     0.1860 13.0000   3.641  0.00299 **
TreatmentB    0.1621     0.6885 13.0000   0.235  0.81751   
TreatmentC   -0.3112     0.7049 13.0000  -0.441  0.66611   
sequence2    -0.1001     0.5715 13.0000  -0.175  0.86367   


[[5]]
Fixed effects:
             Estimate Std. Error        df t value Pr(>|t|)    
(Intercept)  0.137752   0.365302 11.104560   0.377    0.713    
DF[[i]]      0.729762   0.071874  9.810327  10.153 1.61e-06 ***
TreatmentB   0.531048   0.332585  9.144490   1.597    0.144    
TreatmentC   0.060414   0.343280  9.185060   0.176    0.864    
sequence2   -0.001702   0.440920  4.000881  -0.004    0.997    


[[6]]
Fixed effects:
                 Estimate Std. Error        df t value Pr(>|t|)    
    (Intercept)  0.765739   0.446747 13.000000   1.714    0.110    
    DF[[i]]      0.783985   0.132198 13.000000   5.930 4.98e-05 ***
    TreatmentB   0.006516   0.554550 13.000000   0.012    0.991    
    TreatmentC  -0.312968   0.515562 13.000000  -0.607    0.554    
    sequence2   -0.762799   0.436095 13.000000  -1.749    0.104

Is there a way to transform [[4]] --> var1, [[5]] --> var2..., so it's more intuitive and easier to manage the data?

#Approaches 2--Tried storing vars name as a vector first and ran. Did not work

responseList <- names(DF)[c(4:6)]

lm_output2 <- list()

for(i in n){
  lm2<-lmer(get(n+3) ~ get(n) + Treatment+  sequence + (1|id), data = DF)
  summary(lm2)
  lm_output2[[i]] <- summary(lm2)
}

> Error in n + 3 : non-numeric argument to binary operator

I understand this error because in this case, n is not numeric so it would fail to do get (n+3). But I don't know how can I specify var and var_base in the same loop.

Any suggestion is appreciated, thank you!

A. S. K. · Accepted Answer · 2020-06-08 21:37:27Z

1

You can build the formula for lmer as a string. So we could loop over vars (1, 2, 3, etc.) and concatenate formula from the desired variable names, like this:

library(lme4)
lm_output <- list()
for(i in 1:3) {
  outcome_var = paste("var", i, sep = "")
  base_var = paste(outcome_var, "base", sep = "_")
  form = as.formula(paste(outcome_var,
                          " ~ ",
                          base_var,
                          " + treatment + sequence + (1 | id)",
                          sep = ""))
  lm1 = lmer(form, data = DF)
  summary(lm1)
  lm_output[[i]] <- summary(lm1)
}

answered Jun 8, 2020 at 21:37

A. S. K.

2,85616 silver badges27 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Daisy Chang Over a year ago

Thank you! This is very helpful.

Collectives™ on Stack Overflow

R: Problem with loop showing variable names

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related