Skip to main content

Table 3 Descriptive model selection frequencies (DMSF) for investigated simulation scenarios

From: Statistical model building: Background “knowledge” based on inappropriate preselection causes misspecification

Scenario

S(*)_a: Univariable selection,

S(*)_b: Univariable selection,

S(*)_c: Backward elimination,

 

αin=0.05

αin=0.2

AIC

S1_(*)

R1: 0.271

R1: 0.648

R1: 1

 

R2: 0.028

R2: 0.212

R2: 0.997

 

R3: 0.001

R3: 0.024

R3: 0.894

S2_(*)

R1: 0.448

R1: 0.805

R1: 1

 

R2: 0.086

R2: 0.382

R2: 1

 

R3: 0.006

R3: 0.069

R3: 0.999

S3_(*)

R1: 0.912

R1: 0.992

R1: 1

 

R2: 0.581

R2: 0.889

R2: 1

 

R3: 0.170

R3: 0.505

R3: 1

S4_(*)

R1: 0.340

R1: 0.711

R1: 1

 

R2: 0.044

R2: 0.255

R2: 0.999

 

R3: 0.001

R3: 0.037

R3: 0.929

S5_(*)

R1: 0.401

R1: 0.764

R1: 1

 

R2: 0.064

R2: 0.323

R2: 1

 

R3: 0.003

R3: 0.054

R3: 0.966

S6_(*)

R1: 0.678

R1: 0.917

R1: 1

 

R2: 0.161

R2: 0.489

R2: 1

 

R3: 0.009

R3: 0.097

R3: 0.959

S7_(*)

R1: 0.644

R1: 0.899

R1: 1

 

R2: 0.107

R2: 0.419

R2: 0.998

 

R3: 0.005

R3: 0.072

R3: 0.929

S8_(*)

R1: 0.705

R1: 0.933

R1: 1

 

R2: 0.200

R2: 0.557

R2: 1

 

R3: 0.019

R3: 0.140

R3: 1

S9_(*)

R1: 0.834

R1: 0.970

R1: 1

 

R2: 0.362

R2: 0.729

R2: 1

 

R3: 0.029

R3: 0.184

R3: 0.965

S10_(*)

R1: 0.840

R1: 0.977

R1: 1

 

R2: 0.403

R2: 0.774

R2: 1

 

R3: 0.052

R3: 0.267

R3: 1

  1. R1 Rule 1 (one out of three preceding studies identifies predictor); R2 Rule 2 (Two out of three preceding studies identify predictor); R3 Rule 3 (all three preceding studies identify predictor)