Thursday, February 05, 2004
Flashback
The problem, for me, with taking advanced classes, was always that I'd forget the basics. So I have to remind myself.
Chi-squared is a measure of how much observed cell counts in a 2-way table diverge from the expected cell counts.
Thus, a large chi-squared value indicates a big difference, and provides evidence against a null hypothesis.
The chi-squared distribution is what's used to test the difference for significance. Use the chi-squared statistic, with df = (#rows-1)(#columns-1)
Pretty cool, huh?
By the way, Chi-squared was developed in 1900 by Karl Pearson, and is the oldest inference procedure still used in its original form. Pearson's work really set stats off as a separate discipline.
Chi-squared is a measure of how much observed cell counts in a 2-way table diverge from the expected cell counts.
Thus, a large chi-squared value indicates a big difference, and provides evidence against a null hypothesis.
The chi-squared distribution is what's used to test the difference for significance. Use the chi-squared statistic, with df = (#rows-1)(#columns-1)
Pretty cool, huh?
By the way, Chi-squared was developed in 1900 by Karl Pearson, and is the oldest inference procedure still used in its original form. Pearson's work really set stats off as a separate discipline.
