Repeated measure anova error

[Jenna56]

Hi,

I am trying to do rep measures ANOVA using the rSTATIX package but it is showing the following

Error: Problem with mutate() input data. x 0 (non-NA) cases i Input data is map(.data$data, .f, ...).

I am not sure whats the problem is, tried everything but unsuccessful. Worked fine for normal two-way ANOVA but not for repeated. Please help.

[benediktclaus]

Hi @Jenna56

could you provide us with a reprex (reproducible example)? As you can see below, a repeated measures ANOVA can be done as expected with rstatix.

library(tidyverse)
library(rstatix)

ChickWeight %>% 
  anova_test(dv = weight, wid = Chick, within = Time, between = Diet)
#> ANOVA Table (type III tests)
#> 
#> $ANOVA
#>      Effect DFn DFd       F         p p<.05   ges
#> 1      Diet   3  41   5.075  4.00e-03     * 0.161
#> 2      Time  11 451 280.945 6.41e-194     * 0.769
#> 3 Diet:Time  33 451   3.766  9.34e-11     * 0.118
#> 
#> $`Mauchly's Test for Sphericity`
#>      Effect        W         p p<.05
#> 1      Time 2.68e-17 1.03e-251     *
#> 2 Diet:Time 2.68e-17 1.03e-251     *
#> 
#> $`Sphericity Corrections`
#>      Effect   GGe      DF[GG]    p[GG] p[GG]<.05   HFe      DF[HF]    p[HF]
#> 1      Time 0.114 1.26, 51.48 2.01e-24         * 0.116 1.28, 52.34 8.63e-25
#> 2 Diet:Time 0.114 3.77, 51.48 1.00e-02         * 0.116 3.83, 52.34 1.00e-02
#>   p[HF]<.05
#> 1         *
#> 2         *

^{Created on 2021-10-06 by the reprex package (v2.0.1)}

[Grindelwald146]

I am experiencing the same issue but only when trying to pass a formula instead of the way you specified the ANOVA. Attached you can find the data I am using for my analysis, which looks like:

# A tibble: 344 × 4
   subj_id exp_cond     dimension          score
     <dbl> <chr>        <chr>              <dbl>
 1       7 control      edu_mot1_challenge  3.75
 2       7 control      edu_mot1_interest   3.4 
 3       7 control      edu_mot1_success    2.75
 4       7 control      edu_mot1_failure    1.2 
 5       7 control      edu_mot2_challenge  5   
 6       7 control      edu_mot2_interest   5.6 
 7       7 control      edu_mot2_success    6.5 
 8       7 control      edu_mot2_failure    2.6 
 9       1 experimental edu_mot1_challenge  5.5 
10       1 experimental edu_mot1_interest   4.6 
# … with 334 more rows

When I execute my repeated measures ANOVA the same way you did, it works.

data %>% 
    anova_test(dv = score,
               wid = subj_id,
               within = dimension,
               between = exp_cond,
               type = 2)

My output then looks like:

ANOVA Table (type II tests)

$ANOVA
              Effect DFn DFd         F        p p<.05      ges
1           exp_cond   1  41  0.000165 9.90e-01       1.24e-06
2          dimension   7 287 74.487000 7.46e-61     * 5.57e-01
3 exp_cond:dimension   7 287  0.478000 8.50e-01       8.00e-03

$`Mauchly's Test for Sphericity`
              Effect    W        p p<.05
1          dimension 0.07 1.29e-10     *
2 exp_cond:dimension 0.07 1.29e-10     *

$`Sphericity Corrections`
              Effect   GGe      DF[GG]    p[GG] p[GG]<.05  HFe       DF[HF]    p[HF] p[HF]<.05
1          dimension 0.515 3.6, 147.69 1.48e-32         * 0.57 3.99, 163.67 8.30e-36         *
2 exp_cond:dimension 0.515 3.6, 147.69 7.32e-01           0.57 3.99, 163.67 7.51e-01 

However, when I try to pass a formula like so, which resembles the formula notation given on the help page, then I only see an error.

data %>% 
    anova_test(score ~ exp_cond*dimension + Error(subj_id/(exp_cond*dimension)))

Within-Ss ANOVA (repeated measures ANOVA): y ~ w1*w2 + Error(id/(w1*w2))

Error message:

Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, ...) : 
  0 (non-NA) cases

Is this really a bug or did I just specify the formula the wrong way?

Here's my sys.info output:

Sys.info()
                                                                                                sysname 
                                                                                               "Darwin" 
                                                                                                release 
                                                                                               "20.6.0" 
                                                                                                version 
"Darwin Kernel Version 20.6.0: Mon Aug 30 06:12:20 PDT 2021; root:xnu-7195.141.6~3/RELEASE_ARM64_T8101" 
                                                                                               nodename 
                                                                            "XXXX-macbook-m1.fritz.box" 
                                                                                                machine 
                                                                                                "arm64" 
                                                                                                  login 
                                                                                                 "root" 
                                                                                                   user 
                                                                                                 "XXXX" 
                                                                                         effective_user 
                                                                                                 "XXXX" 

data.csv

[Grindelwald146]

OK I just realised that it was my bad because I needed to specify the formula for a mixed ANOVA. Sometimes typing things out really helps to change one's POV :)