cause and effect a statistical riddle

cause and effect

a statistical riddle

a new medicine was invented for preventing heart attack.

it was cllimed, it was good for people. statistics were quoted.

120 persons were examined, sixty who used the medicine. other 60 were control group did not use the medicine.

results over a longish period of time

control group     total                                     60

heart attack                       13

                               no heart attack                 47

treatment group               total                      60

                               heart attack                       11

                               no hear attack                   49

conclusion --- percentage in cntrol group is 21.75; in treatment group 18.67; treatmenr can prevent heart atack

the some one went deeper into statistics. in both the groups (treamnet and control) there were 20 women and 40 men.

the results

	control group		treatment group
	heart attack	no heart attack	heart attack	no heart attack
female	1	19	3	37
male	12	28	8	12
total	13	47	11	49

conclusion – five percent of females in control group had heart attacks; number in treatment ggroup is 7-5 percent.

so it is good for women.

               -- - 30 percent of men in control group had hear attack;  40 percent of men in treatment group had heart attack

so the drug is nad for men

net result – treatment is good for women, bad for men (but overall, it is good)

this is called simpson’s paradox.

(from “the book of why” by judea pearl and dana mackenzie – penguin books – 2018)