Frank Landis' Blog

Astrological Experimentation
January 30, 2011, 12:45 pm
Filed under: Astrology, Experiments, Real Science | Tags: , ,

I posted a response to a Discover online article about how astrology is like racism. Then I realized that I could post it here too.

Years ago, I came up with a way to objectively test astrology and personal horoscopes. It’s simple, and any experimenter can do it if he or she can find a bunch of willing participants and convince them to spend a few hours rating a bunch of horoscopes. I’ve described my results below, and I encourage other people to try it, as a psych experiment or just for fun.

Experimental Design:

Hypothesis: If astrology is useful, then a person’s horoscope should apply to them more than someone else’s horoscope does.
Null hypothesis: subjects will either rate all horoscopes approximately the same, and/or most subjects will find other people’s horoscopes more relevant to their lives than they do their own.

1. Find a website that gives out free, nine planet, twelve house horoscopes.
2. Recruit a bunch of experimental subjects. I’d suggest 10, and fewer than five is problematic. Get their birthplace, birth date and birth time information.
3. Compile everyone’s horoscope from the same website. The experimenter should strip out any identifying information (for example, anything that says Libra, Virgo, etc), and the subjects should not see their horoscopes prior to the experiment. Typically horoscopes are printed as a list of paragraph statements, one for each planet and house.
4. If you want, you can even add in randomly generated horoscopes.
5. If you want to make it simpler, you can do one more step. People who were born in the same year tend to have some of the same planets and houses (particularly for the outer planets, which move very slowly). To make it easier for the subjects, you can compile all the paragraphs into one long paper, and have everyone rate every paragraph once. You will have to create a key for which paragraph goes with which horoscope to compile the stats, but this saves on work for the subjects.
6. Have everyone rate EVERY horoscope, every paragraph, on whether that paragraph applies to them or not (I suggest: 1 pt if the paragraph is relevant to the subject’s life, 0 if it’s neutral, -1 if the paragraph does not apply to the subject’s life).

7. Compile every person’s scoring of all horoscope paragraphs. Add up the scores per horoscope.
8. If astrology is true, the prediction is that each person should have scored their own horoscope higher than they scored those of the other participants. The stats for this are a bit more complicated than ranking individual scores, because just by chance, you would expect some people to pick their own horoscopes as the most applicable. Still, it’s not hard, and if the stats look too ugly, simply post how people rated their own and other horoscopes.
9. Collect post-test impressions from the subjects, distribute the results, and have fun talking about it.

When I ran this with 6 subjects with four additional random horoscopes, I got equivocal results (1 person picked their own horoscope, 5 people chose other people’s horoscopes, but with the small sample size, I couldn’t test the hypothesis). I’d love to see other people replicate the test and post their results.

The nice part about this is that it gets around all the tired ideological debates (“it’s not science” vs. “keep an open mind”) and looks at whether printed horoscopes have any perceived relevance to the people who requested them.

What I learned about horoscopes is that, when you read your own horoscope, you tend to focus on the bits that are relevant and ignore the rest. Horoscopes are written to favor this habit: they have a bunch of generally applicable advice mixed very nicely together, much like Forrest Gump’s box of chocolates. However, when you read other people’s horoscopes, what you find is that their horoscopes are also applicable to you. In fact, you may well like someone else’s horoscope better than you like your own. Five of the six people above found that, and one person even preferred a randomly generated horoscope over his own.

Most divination methods work this way: it’s not what is displayed by the cards, planets, coins, whatever, it’s what the person reads into them. This is not necessarily a bad thing, but I think it is better to understand how such a method works, rather than uncritically accept it.

Try it out, and tell me what you think.