Three stata do programs are needed:

Pr1. running

Pr2. set of single funnels

Pr3. joint programs

Here Pr1 make the full set of simulations using Pr3.

While Pr2 make one set of funnels.

Program Pr3 does not run separately.

To run the programs download then into a folder. Make sure that the
cd line "E:\Google drev\MARTIN\ARBEJDE\AAC" is changed to refer to the said filder.

You should set the following in Pr1 and Pr2:

the settings made are the ones replicating the paper

****Parameters:

local ci = 0.05 // set confidence level

local r = 1000 // number of funnels

local N = 500 // number of regressions in each funnel

local obs = 20 // number of observations of the regressions. Start of m-loop, see paper

local mint = 2 // minimum acceptable t value (needed to determine point a)

****Variables:

local x_mean = 0 // mean of x

local e_mean = 0 // mean of error term

local x_sd = 2 // standard deviation of x

local e_sd = 10 // standard deviation of error term

local beta = 1 // set true parameter for beta

****Set j, one or more values
foreach j in 1 2 5 10 15 23 34 50 {

For all j's your pc will run for about a week, so take one at a time

When running, it needs a prompt once or twice in the beginning

But once the counting starts it just runs and set one '.' for each funnel

Output:

The program generates two folders in the program folder

dataset with stata.dta output for each SR

output with excell files

Both sets of output has the same format as Tables 2 to 6 in the paper