Analyse time clusters

[MonicaBarbir]

Hello Eyetracking R,

I am having some problems analysing time clusters. I get the following error in the script that had worked a couple days ago: Error in .subset2(x, i, exact = exact) : recursive indexing failed at level 2

I have recently updated Eyetracking R, and other necessary packages such as Dplyr. Since having the problem, I've gone about updating every package just in case, but to no avail.

I had had a similar problem before, but it had worked out when I downgraded Dplyr (which I believe cannot do now, because the new version of Eyetracking R requires the new version of Dplyr).

Any ideas what it may be caused by?

Thank you, Monica

[MonicaBarbir]

Re-hello EyetrackingR,

Update that might help with figuring out the source of the problem: I just ran analyse time clusters on the data set provided with eyetrackingR and I got the following error - Error in *tmp*[[jj]] : subscript out of bounds

Thanks again, Monica

[jwdink]

Hey @MonicaBarbir ,

Can you provide the code for a reproducible example?

Thanks!

[respatte]

Hi everyone,

I'm having the same issue so I'm just going to add it here rather than open a new one. There are a couple issues with analyze_time_clusters:

• treatment-coded variables can't be used with within_subj = F in analyze_time_clusters as the treatment_level argument isn't passed from make_time_cluster_data into analyze_time_clusters:

  ## Determine clusters
  LT.time_cluster_tail <- make_time_cluster_data(LT.time_course_tail,
                                                 predictor_column = "Condition",
                                                 treatment_level = "No Label",
                                                 aoi = "Tail",
                                                 test = "lmer",
                                                 threshold = 1.5,
                                                 formula = ArcSin ~ Condition +
                                                   (1 | Participant) +
                                                   (1 | Stimulus))

  ## Run analysis
  LT.time_cluster_tail.analysis <- analyze_time_clusters(LT.time_cluster_tail,
                                                         formula = ArcSin ~ Condition +
                                                           (1 | Participant) +
                                                           (1 | Stimulus),
                                                         within_subj = F,
                                                         parallel = T,
                                                         samples = 200)
# Error in get_resampled_sum_stat_btwn(data, rows_of_participants, attrs,  : 
#   task 1 failed - "
# The term 'Condition' was not found in your model.
# This can happen if your predictor is treatment-coded, in which case you should
#   specify the `treatment_level` argument.
# For example, if 'Target' is treatment coded, then it becomes 'TargetInanimate'
#   in the model, and you should set `treatment_level = 'Inanimate'`.
# The terms in the model were: '(Intercept)', 'ConditionNo Label'"

• directly supplying the treatment level in the predictor_column name (in make_time_cluster_data) raises an error when running analyze_time_clusters (different error depending on the value given to within_subj):

  ## Determine clusters
  LT.time_cluster_tail <- make_time_cluster_data(LT.time_course_tail,
                                                 predictor_column = "ConditionNo Label",
                                                 aoi = "Tail",
                                                 test = "lmer",
                                                 threshold = 1.5,
                                                 formula = ArcSin ~ Condition +
                                                   (1 | Participant) +
                                                   (1 | Stimulus))

  ## Run analysis with within_subj = T
  LT.time_cluster_tail.analysis <- analyze_time_clusters(LT.time_cluster_tail,
                                                         formula = ArcSin ~ Condition +
                                                           (1 | Participant) +
                                                           (1 | Stimulus),
                                                         within_subj = T,
                                                         parallel = T,
                                                         samples = 200)
# Error in analyze_time_clusters.time_cluster_data(LT.time_cluster_tail,  : 
#   The column 'ConditionNo Label' not found in your data.

## Run analysis with within_subj = F
  LT.time_cluster_tail.analysis <- analyze_time_clusters(LT.time_cluster_tail,
                                                         formula = ArcSin ~ Condition +
                                                           (1 | Participant) +
                                                           (1 | Stimulus),
                                                         within_subj = F,
                                                         parallel = T)
# Error in get_resampled_sum_stat_btwn(data, rows_of_participants, attrs,  : 
#   task 1 failed - "subscript out of bounds"

I'm not totally sure I understand the use of within_subj however, so that might be the issue. In my experiment, each participant is assigned to one of two conditions, and so Condition should be a between-subject factor, hence why I want to use within_subj = F; am I wrong?

If you want my data to run the code on it, you can find it on my experiment's GitHub repository. You need to open the stats/InfantsAnalysis.R file, run the blocks LIBRARY IMPORTS and GATHER DATA, then in the block LOOKING TIME ANALYSIS: TIME COURSE you can run the DATA PREPARATION lines and finally play with the code starting line 160.

[arielleborovsky]

I also encountered this issue and found that the problem had to do with specifying the make_time_sequence_data function. In my case, both "condition" and "participant" were wrongly included in the "summarize_by" argument. Moving "condition" to the "predictor_columns" argument cleared up the issue for me. Hope this helps you too!

Example in code:

WRONG:
timecourse <- make_time_sequence_data(df, time_bin_size = 50, aois = c("AOI1", "AOI2", "AOI3", "AOI4"), predictor_columns = c("Sex"), summarize_by = c("Participant", "Condition"))

CORRECT: timecourse <- make_time_sequence_data(df, time_bin_size = 50, aois = c("AOI1", "AOI2", "AOI3", "AOI4"), predictor_columns = c("Sex", "Condition"), summarize_by = c("Participant"))