Of the result happening by coincidence chance is less

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: E.g., lots of smoking is correlated with lung cancer Correlations can range anywhere from: -1 to +1 – A perfect, positive correlation (+1) is, for example, the perfect, observation that “when the sun comes up, it gets lighter outside”. outside”. – A perfect, negative correlation (-1) is, “as the sun goes perfect, down, darkness increases” down, – An example of a zero correlation between events is me An sneezing and CY Leung getting kicked out of office. sneezing Statistics VII - Findings If there are so many findings that are less If than 5% likely to have been caused by chance, why do they contradict each other all the time? all – E.g., one day there is a new cure for cancer; the E.g., next day it “doesn’t work” next Problem: one study doesn’t make it true – Hence, meta-analysis; combines data from Hence, many studies and analyzes it all together many Strange Findings Strange The more firemen fighting a fire, the more damage there is The going to be. – Therefore firemen cause damage. Therefore – True??? As ice cream sales increase, the rate of drowning deaths As increases sharply. – Therefore, ice cream causes drowning. – True?? True?? Sleeping with one's shoes on is strongly correlated with Sleeping waking up with a headache. – Therefore, sleeping with one's shoes on causes headache. Therefore, – True?? One of the biggest statistical rules One Correlation DOES NOT PROVE causation – Just because two events happen together does not Just mean that one causes the other!! mean – There are often different reasons for the correlation of There two events two – Often these reasons explain the relationship better than Often the causation of one event by the other the – So if someone says “I took X medicine and my cat got So sick, so this medicine makes my cat get sick”, you should question the causation! question Questions?? Questions??...
View Full Document

This note was uploaded on 10/03/2013 for the course SOCI 1000 taught by Professor Tim during the Fall '13 term at CUHK.

Ask a homework question - tutors are online