Error when using the summary() function.

[enci91]

Hi everyone,

I am encountering an error when using the summary() function on a pls-sem model.
The error message is:

> summary_pls <- summary(pls_model)
Error in `[<-`(`*tmp*`, int_components, dv, value = NA) : 
  subscript out of bounds

The model is calculated as normal, and can even be bootstrapped without issues, however no summary can be generated from it.

My code is as follows:

# Data prep
setwd("[my directory]")
urdf <- read_excel("[my dataset]")

#PLS-SEM
measurements <- constructs(
  composite("Gender", single_item("T1_kjønn")),
  composite("Age", single_item("T1_alder")),
  reflective("Perf_Expectancy_T1", multi_items("T1_pe", 1:4)),
  reflective("Fac_Conditions_T1", multi_items("T1_fc", 1:4)),
  reflective("Beh_Intentions_T1", multi_items("T1_bi", 1:3)),
  reflective("Eff_Expectancy_T1", multi_items("T1_ee", 1:4)),
  reflective("Soc_Influence_T1", multi_items("T1_si", 1:3)),
  reflective("Hed_Motivation_T1", multi_items("T1_hm", 1:3)),
  reflective("Habit_T1", multi_items("T1_ht", 1:4)),
  reflective("Experience", single_item("T1_ex1")),
  reflective("Use_Frequency_T1", single_item("T1_use1")),
  reflective("AIAS-4_T1", multi_items("T1_aias", 1:4)),
  reflective("Perf_Expectancy_T2", multi_items("T2_pe", 1:4)),
  reflective("Fac_Conditions_T2", multi_items("T2_fc", 1:4)),
  reflective("Beh_Intentions_T2", multi_items("T2_bi", 1:3)),
  reflective("Eff_Expectancy_T2", multi_items("T2_ee", 1:4)),
  reflective("Soc_Influence_T2", multi_items("T2_si", 1:3)),
  reflective("Hed_Motivation_T2", multi_items("T2_hm", 1:3)),
  reflective("Habit_T2", multi_items("T2_ht", 1:4)),
  reflective("Experience_T2", single_item("T2_ex1")),
  reflective("Use_Frequency_T2", single_item("T2_use1")),
  reflective("AIAS-4_T2", multi_items("T2_aias", 1:4)),
  interaction_term(iv = "Beh_Intentions_T1", moderator = "Gender"),
  interaction_term(iv = "Beh_Intentions_T1", moderator = "Age"),
  interaction_term(iv = "Beh_Intentions_T1", moderator = "Experience")
)

structure <- relationships(
  paths(from = c("Beh_Intentions_T1", "Beh_Intentions_T1*Gender", "Beh_Intentions_T1*Age", "Beh_Intentions_T1*Experience"), to = "Beh_Intentions_T2"),
  paths(from = "Use_Frequency_T1", to = "Use_Frequency_T2"),
  paths(from = "Beh_Intentions_T1", to = "Use_Frequency_T1"),
  paths(from = "Beh_Intentions_T2", to = "Use_Frequency_T2"),
  paths(from = c("Perf_Expectancy_T1", "Fac_Conditions_T1", "Eff_Expectancy_T1", "Soc_Influence_T1", "Hed_Motivation_T1", "Habit_T1", "AIAS-4_T1"), to = "Beh_Intentions_T1"),
  paths(from = c("Perf_Expectancy_T2", "Fac_Conditions_T2", "Eff_Expectancy_T2", "Soc_Influence_T2", "Hed_Motivation_T2", "Habit_T2", "AIAS-4_T2"), to = "Beh_Intentions_T2")
)  

pls_model <- estimate_pls(data = urdf, 
                          measurement_model = measurements, 
                          structural_model = structure
)

plot_sem <- plot(pls_model)
summary_pls <- summary(pls_model)

I'm running the code on an M1 Macbook Pro using R Studio, with the following version specification:

> version
               _                           
platform       aarch64-apple-darwin20      
arch           aarch64                     
os             darwin20                    
system         aarch64, darwin20           
status                                     
major          4                           
minor          4.1                         
year           2024                        
month          06                          
day            14                          
svn rev        86737                       
language       R                           
version.string R version 4.4.1 (2024-06-14)
nickname       Race for Your Life

Someone else received the same error message (#330), but in that thread the user was attempting to fit a regression path model, which is not this case.

Let me know what you guys think!

[NicholasDanks]

Dear @enci91
Thanks for taking the time and effort to so neatly present your issue. It is ver much appreciated.
In this case, I believe the problem arises because you have not fully specified all direct paths. That is, for example you have at least one interaction term that is connected to an endogenous variable but that the direct effects have been omitted (IV*MV)

I usually try a little plot(measurement_model) or plot(structural_model) and then I don an eyeball scan to ensure that all the direct effects are there along with the interaction term. Take a look above I ran plot on your two models (without data).
I see you have an interaction Beh_Intentions_T1*Gender pointing at Beh_Intentions_T2 but neither of the direct effects Beh_Intentions_T1 or Gender which must also be specified explicitly.

I really like that plot function applied to a measurement model or structural model object because it tells you a lot about your spelling and variables and helps to visualize the model in a way that is more digestible.

Thanks!