Try the experiment yourself!

Heck, maybe you disagree with us. You may be able to prove us wrong! Here's how to make an earthquake prediction experiment of your own:

First, you need TONS of information and a source to get it from. A college library (or just a really, really huge one) should have old newspapers on microfilm. Pick a newspaper that carries lost pet ads and then find out the latitude and longitude of the town where the newspaper is so you can pick earthquakes nearby.

Get the dates of earthquakes with magnitude greater than 4.5 that are near the town of the newspaper. If the town is in Northern California you can do this by using this webpage. Magnitude 4.5 and above is good because these earthquakes will be widely felt but you can adjust this and the distance from the town to get a nice sized data set. We aimed for 10 earthquakes.

For towns outside of Northern California try looking here for information on your area.

Then pick your random dates making sure to steer clear of your earthquake days or anywhere near them. Why? See our intrroduction page for an explanation.

Now find the newspapers for every one of thoses dates (Random or Earthquake) and jot down the number of pets lost on those dates. Be sure to write down their date and whether or not it's a random day beside it. Then, lay out a chart to write your information down. Make a column for the date, whether it's a Random or Earthquake day, lost pets, and the total of lost pets.

Now that you have your data, examine it to see if the number of lost pet ads can be used to predict the earthquakes.