From 29ae0ae3ef3f897c2c6e52067e79e233f5f1de1b Mon Sep 17 00:00:00 2001
From: Phillip Alday <phillip.alday@beacon.bio>
Date: Tue, 9 Jul 2024 17:12:07 +0200
Subject: [PATCH] add a comment

---
 test/mixedmodels.jl | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/test/mixedmodels.jl b/test/mixedmodels.jl
index d3f0654..61fb82d 100644
--- a/test/mixedmodels.jl
+++ b/test/mixedmodels.jl
@@ -26,6 +26,11 @@ eff = effects(design, model)
 @test eff.y + eff.err ≈ eff.upper
 
 β = fixef(model)
+# note that in the design above, we had a = 2
+# β[1] = intercept
+# β[2] = a, multiply by 2 because that's what we put in the design
+# β[3] = whatever of x|x2 wasn't pivoted out, "multiply by 1" (ie do nothing)
+#        because we had keep_coef = 1 in the design
 @test only(eff.y) ≈ β[1] + 2 * β[2] + β[3]
 
 # there is one bit of weirdness -- the removed coefficients behave like any other