Misidentification 6: X1 Regressed on EX31-EX33, Y1 on X1 and LEN1-LEN4

For the sixth misidentification, we regressed X₁ on EX₃₁-EX₃₃, Y₁ on X₁ and LEN₁-LEN₄ using Lavaan (illustrated in the Figure below). This was replicated 10,000 times – using R-loops – while randomly varying the effects for each causal pathway between the constructs in the network excluding the direct effect of X₁ on Y₁ (true direct causal effect = 1.00; N = 10,000; R-Code).

As illustrated in the figure below, the misidentification of the structural association where X1 on EX₃₁-EX₃₃, Y₁ on X1 and LEN₁-LEN₄, produced an average slope coefficient of b = 3.387 for the relationship between X₁ and Y₁. This suggests that a 1 point increase in X₁ directly causes a 3.387 increase in Y₁, on average. Considering the past models, this is extremely close to the true direct effect of X₁ on Y₁ (As a reminder, the true slope coefficient is 1.00).