Fixed confidence interval test again

Author: dscolby

src/inference.jl

@@ -335,7 +335,7 @@ julia> p_value_and_std_err(null_dist, CausalELM.mean(null_dist))
 """
 function p_value_and_std_err(null_dist, test_stat)
     n = length(null_dist)
-    extremes = length(null_dist[abs(test_stat) .<= null_dist])
+    extremes = length(null_dist[abs(test_stat) .<= abs.(null_dist)])
     pvalue = extremes / n
     stderr = (sum([(test_stat .- x)^2 for x in null_dist]) / (n - 1)) / sqrt(n)
 

test/test_inference.jl

@@ -7,23 +7,23 @@ Float64.([rand() < 0.4 for i in 1:100])
 
 g_computer = GComputation(x, t, y)
 estimate_causal_effect!(g_computer)
-g_inference = CausalELM.generate_null_distribution(g_computer, 100)
+g_inference = CausalELM.generate_null_distribution(g_computer, 1000)
 p1, stderr1 = CausalELM.p_value_and_std_err(g_inference, CausalELM.mean(g_inference))
 lb1, ub1 = CausalELM.confidence_interval(g_inference)
 p11, stderr11, lb11, ub11 = CausalELM.quantities_of_interest(g_computer, 100)
 summary1 = summarize(g_computer, n=100, inference=true)
 
 dm = DoubleMachineLearning(x, t, y)
 estimate_causal_effect!(dm)
-dm_inference = CausalELM.generate_null_distribution(dm, 100)
+dm_inference = CausalELM.generate_null_distribution(dm, 1000)
 p2, stderr2 = CausalELM.p_value_and_std_err(dm_inference, CausalELM.mean(dm_inference))
 lb2, ub2 = CausalELM.confidence_interval(dm_inference)
 summary2 = summarize(dm, n=100)
 
 # With a continuous treatment variable
 dm_continuous = DoubleMachineLearning(x, t, rand(1:4, 100))
 estimate_causal_effect!(dm_continuous)
-dm_continuous_inference = CausalELM.generate_null_distribution(dm_continuous, 100)
+dm_continuous_inference = CausalELM.generate_null_distribution(dm_continuous, 1000)
 p3, stderr3 = CausalELM.p_value_and_std_err(
     dm_continuous_inference, CausalELM.mean(dm_continuous_inference)
 )
@@ -37,7 +37,7 @@ summary4 = summarize(its, n=100)
 summary4_inference = summarize(its, n=100, inference=true)
 
 # Null distributions for the mean and cummulative changes
-its_inference1 = CausalELM.generate_null_distribution(its, 100, true)
+its_inference1 = CausalELM.generate_null_distribution(its, 1000, true)
 its_inference2 = CausalELM.generate_null_distribution(its, 10, false)
 lb4, ub4 = CausalELM.confidence_interval(its_inference1)
 p4, stderr4 = CausalELM.p_value_and_std_err(its_inference1, CausalELM.mean(its_inference1))
@@ -49,7 +49,7 @@ summary5 = summarize(slearner, n=100)
 
 tlearner = TLearner(x, t, y)
 estimate_causal_effect!(tlearner)
-tlearner_inference = CausalELM.generate_null_distribution(tlearner, 100)
+tlearner_inference = CausalELM.generate_null_distribution(tlearner, 1000)
 lb6, ub6 = CausalELM.confidence_interval(tlearner_inference)
 p6, stderr6 = CausalELM.p_value_and_std_err(
     tlearner_inference, CausalELM.mean(tlearner_inference)
@@ -59,7 +59,7 @@ summary6 = summarize(tlearner, n=100)
 
 xlearner = XLearner(x, t, y)
 estimate_causal_effect!(xlearner)
-xlearner_inference = CausalELM.generate_null_distribution(xlearner, 100)
+xlearner_inference = CausalELM.generate_null_distribution(xlearner, 1000)
 lb7, ub7 = CausalELM.confidence_interval(xlearner_inference)
 p7, stderr7 = CausalELM.p_value_and_std_err(
     xlearner_inference, CausalELM.mean(xlearner_inference)
@@ -73,29 +73,29 @@ summary9 = summarize(rlearner, n=100)
 
 dr_learner = DoublyRobustLearner(x, t, y)
 estimate_causal_effect!(dr_learner)
-dr_learner_inference = CausalELM.generate_null_distribution(dr_learner, 100)
+dr_learner_inference = CausalELM.generate_null_distribution(dr_learner, 1000)
 lb8, ub8 = CausalELM.confidence_interval(dr_learner_inference)
 p8, stderr8 = CausalELM.p_value_and_std_err(
     dr_learner_inference, CausalELM.mean(dr_learner_inference)
 )
 summary10 = summarize(dr_learner, n=100)
 
 @testset "Generating Null Distributions" begin
-    @test size(g_inference, 1) === 100
+    @test size(g_inference, 1) === 1000
     @test g_inference isa Array{Float64}
-    @test size(dm_inference, 1) === 100
+    @test size(dm_inference, 1) === 1000
     @test dm_inference isa Array{Float64}
-    @test size(dm_continuous_inference, 1) === 100
+    @test size(dm_continuous_inference, 1) === 1000
     @test dm_continuous_inference isa Array{Float64}
-    @test size(its_inference1, 1) === 100
+    @test size(its_inference1, 1) === 1000
     @test its_inference1 isa Array{Float64}
     @test size(its_inference2, 1) === 10
     @test its_inference2 isa Array{Float64}
-    @test size(tlearner_inference, 1) === 100
+    @test size(tlearner_inference, 1) === 1000
     @test tlearner_inference isa Array{Float64}
-    @test size(xlearner_inference, 1) === 100
+    @test size(xlearner_inference, 1) === 1000
     @test xlearner_inference isa Array{Float64}
-    @test size(dr_learner_inference, 1) === 100
+    @test size(dr_learner_inference, 1) === 1000
     @test dr_learner_inference isa Array{Float64}
 end
 
@@ -117,23 +117,23 @@ end
 end
 
 @testset "Confidence Intervals" begin
-    @test lb1 !== NaN && ub1 !== NaN
-    @test lb2 !== NaN && ub2 !== NaN
-    @test lb3 !== NaN && ub3 !== NaN
-    @test lb4 !== NaN && ub4 !== NaN
-    @test lb6 !== NaN && ub6 !== NaN
-    @test lb7 !== NaN && ub7 !== NaN
-    @test lb8 !== NaN && ub8 !== NaN
+    @test lb1 < g_computer.causal_effect < ub1
+    @test lb2 < dm.causal_effect < ub2
+    @test lb3 < dm_continuous.causal_effect < ub3
+    @test lb4 < CausalELM.mean(its.causal_effect) < ub4
+    @test lb6 < CausalELM.mean(tlearner.causal_effect) < ub6
+    @test lb7 < CausalELM.mean(xlearner.causal_effect) < ub7
+    @test lb8 < CausalELM.mean(dr_learner.causal_effect) < ub8
 end
 
 @testset "All Quantities of Interest" begin
-    @test lb11 !== NaN && ub11 !== NaN
+    @test lb11 < g_computer.causal_effect < ub11
     @test 1 >= p11 >= 0
     @test stderr11 > 0
-    @test lb44 !== NaN && ub44 !== NaN
+    @test lb44 < CausalELM.mean(its.causal_effect) < ub44
     @test 1 >= p44 >= 0
     @test stderr44 > 0
-    @test lb66 !== NaN && ub66 !== NaN
+    @test lb66 < CausalELM.mean(tlearner.causal_effect) < ub66
     @test 1 >= p66 >= 0
     @test stderr66 > 0
 end