0

我想知道我们是否可以在 PLS-SEM 中使用二进制变量（编码 0 和 1）作为主持人？我读到可以将它们作为构造包含在 PLS-SEM 的测量模型中，如我的代码中所示。我使用了产品指标的方法，它奏效了。尽管如此，它不适用于两步法。这种方法可行吗？

此外，我发现这个来源认为可以在 R 中的 PLS SEM 中将二进制变量视为潜在变量（构造）：https ://www.gastonsanchez.com/PLS_Path_Modeling_with_R.pdf

有没有人对此有进一步的参考？


library(seminr)
# Create measurement model lockdown with moderation 
mm_lock_mod <- constructs(
  composite("l1", multi_items("x", c(10,12)), weights = mode_B),
  composite("l2", multi_items("x", 21:24), weights = mode_A),
  composite("l3", multi_items("x", c(29,30,32)), weights = mode_A),     
  composite("l4", multi_items("x", 43:45), weights = mode_A),  
  composite("l5", multi_items("x", 49:52), weights = mode_A),   
  composite("l6", multi_items("x", 58:59), weights = mode_A), 
  composite("l7", multi_items("x", 63:65), weights = mode_A), 
  composite("l8", multi_items("x", 69:70), weights = mode_A), 
  composite("l10", multi_items("x", 82:83), weights = mode_A),
  composite("l11", multi_items("x", 87:89), weights = mode_A),   
  composite("l12", single_item("x96")),
  interaction_term( iv = "l6", moderator = "l12", method = product_indicator),
  interaction_term( iv = "l7", moderator = "l12", method = product_indicator), 
  interaction_term( iv = "l4", moderator = "l12", method = product_indicator), 
  interaction_term( iv = "l10", moderator = "l12", method = product_indicator),
  interaction_term( iv = "l11", moderator = "l12", method = product_indicator) 
  ) 
# Create structural model lockdown with moderation 
sm_lock_mod <- relationships(
  paths( from = c("l1"), to = c("l2")), 
  paths( from = c("l2"), to = c("l3")), 
  paths( from = c("l2","l3"), to = c("l4")),
  paths( from = c("l3","l4","l6","l7","l12","l4*l12","l6*l12","l7*l12"), to = c("l5")), 
  paths( from = c("l3","l4","l5","l12","l10","l11","l10*l12","l11*l12"), to = c("l8")) 
)

# Estimate the new model with moderator
plsmodel_lock_mod <- estimate_pls(
  data = ds2,
  measurement_model = mm_lock_mod,
  structural_model = sm_lock_mod,
  missing = mean_replacement,
  missing_value = -99)
# Extract the summary
summary_plsmodel_lock_mod <- summary(plsmodel_lock_mod)
summary_plsmodel_lock_mod```
4

0 回答 0