This is a classic misuse of statistical inference. When a statistical test is done at the standard (95%) significance level, it means that there is a 95% chance that the relationship noted is not due to random chance. A fun thing that one can do with astrological signs is run 2 significance tests relating each sign to a variable for both positive and negative correlation. Even assuming that there is no real relationship, in general one out of every 20 tests will show 'statistically significant' results. You are therefore likely to find at least one significant correlation between a specific sign and anything you test it against.

For this reason, it is important to state a specific hypothesis (i.e. Leos have more frequent accidents) before going fishing for any specific relationship because you are going to make an error once out of every 20 tests on average. One can do the same with breakfast cereal if one wishes.

skullaria wrote:
"According to a study by InsuranceHotline.com, a Web site that quotes drivers on insurance rates, astrological signs are a significant factor in predicting car accidents.

The study, which looked at 100,000 North American drivers' records from the past six years, puts Libras (born September 23-October 22) followed by Aquarians (January 20-February 18) as the worst offenders for tickets and accidents

Leos (July 23-August 22) and then Geminis (May 21-June 20) were found to be the best overall.

Even age, another variable for determining insurance rates, is less of a consideration to Romanov. The cutoff line for being considered a higher risk driver is 24 years of age; 25-year-olds are considered not-high risk.

"I'd rather get into a car with a 24-year-old Leo than a 25-year-old Aries," Romanov said....

Leos, described along with the study results on InsuranceHotline.com/a10.html ..."

zeugma wrote:
This is a classic misuse of statistical inference. When a statistical test is done at the standard (95%) significance level, it means that there is a 95% chance that the relationship noted is not due to random chance. A fun thing that one can do with astrological signs is run 2 significance tests relating each sign to a variable for both positive and negative correlation. Even assuming that there is no real relationship, in general one out of every 20 tests will show 'statistically significant' results. You are therefore likely to find at least one significant correlation between a specific sign and anything you test it against.

For this reason, it is important to state a specific hypothesis (i.e. Leos have more frequent accidents) before going fishing for any specific relationship because you are going to make an error once out of every 20 tests on average. One can do the same with breakfast cereal if one wishes.

skullaria wrote:
"According to a study by InsuranceHotline.com, a Web site that quotes drivers on insurance rates, astrological signs are a significant factor in predicting car accidents.

The study, which looked at 100,000 North American drivers' records from the past six years, puts Libras (born September 23-October 22) followed by Aquarians (January 20-February 18) as the worst offenders for tickets and accidents

Leos (July 23-August 22) and then Geminis (May 21-June 20) were found to be the best overall.

Even age, another variable for determining insurance rates, is less of a consideration to Romanov. The cutoff line for being considered a higher risk driver is 24 years of age; 25-year-olds are considered not-high risk.

"I'd rather get into a car with a 24-year-old Leo than a 25-year-old Aries," Romanov said....

Leos, described along with the study results on InsuranceHotline.com/a10.html ..."

It is interesting, but perhaps even more interesting TO ME is that most people that BELIEVE in astrology would also find fault with the numbers - insisting that the study should have been more specific to astrology charts and not 'just a sun sign.' In any event,' I'm a Gemini - give me a damn discount!' ... sounds very good to me!