MATLAB code supporting the paper "Comparison of a phased experimental approach and a single randomized clinical trial for developing multicomponent behavioral interventions"

Here we provide the code corresponding to the methods described in the appendix and the results reported in the paper. To replicate the results, download all the programs into a single folder (e.g. workspace) and run the following driver code:

Driver for the entire simulation

The driver calls the following routine to conduct one simulation:

Operationalization of both classical and phased experimental approach

Operationalization of classical approach - called by the previous routine

Routine for conducting follow-up experiments

Following is a list of routines that sets the true parameter values and a priori beliefs. Save these .m files to the same workspace used in running the driver:

True parameter values

Mean of eta parameters

Mean of beta parameters

Mean of alpha parameters

Standard deviation of eta parameters

Standard deviation of beta parameters

Standard deviation of alpha parameters

Implementation of Algorithm 1 in the appendix and its variants:

find_best_comb

find_best_comb_aliasing

find_best_comb_aliasing_1

find_best_comb_aliasing_2

find_best_comb_aliasing_3

Other supporting code:

compute_Ad_values

compute_M_values

compute_response

designing_routine1

designing_routine2

designing_routine3

designing_routine4

designing_routine5

find_alias

find_alias_1_2

find_parents

find_max_response

find_sigma_values

find_the_rest

generate_doe_response

generate_response

get_err_type

get_sci_sigma

get_sci_values

make_design_21

make_design_22

make_design_32

make_design_33

make_design_34

monte_carlo_est

natures_choice_new

simulate_from_prior

screening_design_construction

srswor

std_eff_size

std_prior_eff

write_output_properly