I am a beginner trying to learn statistics. I have a questions related to randomising simulations to accept or reject null hypotheses.
A random experiment was conducted with 48 bank managers to determine if there is a gender discrimination at the work place.
There were 48 IDENTICAL RESUMES, 24 of them belonged to male applicants, 24 belonged to female applicants.
24 managers received 24 RESUMES belonging to MALES, the distribution of the resumes to the managers was assigned at RANDOM.
24 managers received 24 RESUMES belonging to FEMALES, the distribution of the resumes to the managers was assigned at RANDOM.
Out of 24 males, 21 were promoted.
Out of 24 females, only 15 were promoted.
Meaning that the difference between the male promotion rate and female promotion rate was roughly 29%
NULL HYPOTHESIS: There was no discrimination. In other words, the 29% was due to chance.
ALTERNATIVE HYPOTHESIS: There was discrimination.
The same experiment was simulated 100 times and a dot plot was drawn.
X AXIS - Promotion percentage difference between male and female observed from -100% to +100%. Centre being 0%, left extreme being -100% and right extreme being 100%.
Y AXIS - DOT COUNT from each simulation.
I hope you can visualise it. The video explained that, since from the 100 randomised simulations, since there were approximately only 2 or 3 data (represented by the DOTS) that had the difference of 29% just like the data we observed and LOTS of data (represented by the DOTS) which had -3% to 3% (negative 3% meaning, that females were favoured) we can reject the null hypothesis.
My questions is, if in these 100 randomised simulations, the dot plot showed that there were in fact MAJORITY (largest no. of DOTS) at the 29% mark, then we must ACCEPT the null hypotheses? That the discrimination occurred due the chance since even after randomising (making it appear the everything is due to chance) we still obtained the same data that the FIRST record of 29% difference showed?