(October 19, 2022 at 9:09 am)FlatAssembler Wrote: No, I think they determine the cause and effect using the p-values. I think the p-value can be estimated analitically here, but I am not willing to spend time relearning statistics just for this, so I am going to estimate it numerically using this JavaScript program (just like I numerically estimated the p-value in the latest linguistics paper I have published):
That program prints:Code:const how_many_people_in_the_general_population_have_low_blood_pressure=1/3; // I think it's even less than that, but let's go with that.
const how_many_people_with_heart_attacks_have_low_blood_pressure=0.41;
const how_many_people_were_in_the_study=100; // And I am quite sure there were more, but let's go with that.
const how_many_times_will_we_run_the_simulation=10_000;
let how_many_times_did_we_get_the_expected_result=0;
for (let i=0; i<how_many_times_will_we_run_the_simulation; i++) {
let counter=0;
for (let i=0; i<how_many_people_were_in_the_study; i++)
counter+=Math.random()<how_many_people_in_the_general_population_have_low_blood_pressure;
if (counter/how_many_people_were_in_the_study>how_many_people_with_heart_attacks_have_low_blood_pressure)
how_many_times_did_we_get_the_expected_result++;
}
console.log(`The p-value is ${how_many_times_did_we_get_the_expected_result/how_many_times_will_we_run_the_simulation*100}%.`);
Since the p-value is less than 5%, we conclude the relationship is causal.Code:The p-value is 4.32%.
What a lovely example of garbage in - garbage out. Here's what the people who did actual research said:
Mills KT, Stefanescu A, He J. The global epidemiology of hypertension. Nat Rev Nephrol. 2020 Apr;16(4):223-237.
The abstract begins, "Hypertension is the leading cause of cardiovascular disease and premature death worldwide."
And goes on to say, "Variations in the levels of risk factors for hypertension, such as high sodium intake..."
Feel free to read the entire article. Peer-reviewed or it didn't happen.