Statistics
posted by Kevin on .
This is a goodnessoffit question that I've been trying to wrap my head around for sometime, but I don't understand it since there are two values instead of one. So far, the teacher has only taught us how to deal with one set of observed values.
UCLA conducted a survey of more than 263,000 college fresmen from 385 colleges in fall 2005. The results of student expected majors by gender were reported in The Chronicle of Higher Education (2/2/06). Conduct a goodness of fit test to determine if the male distibution fits the female distribution.
Maj  Women  Men
Art  14.0%  11.4%
Bio  8.4%  6.7%
Bus  13.1%  22.7%
Edu  13.0%  5.8%
Eng  2.6%  15.6%
Phy  2.6%  3.6%
Pro  18.9%  9.3%
Soc  13.0%  7.6%
Tec  0.4%  1.8%
Oth  5.8%  8.2%
Und  8.0%  6.6%
The final answer must be completed to follow this chart...
O E (OE)^2 (OE)^2/E
   
If I can just get a step in the right direction of how to calculate both sets of percentages I hopefully will be able to understand and do the rest. Thanks in advance, I appreciate it greatly.
Kevin

I think I see your issue. You are testing whether the male distribution fits the female distribution. So, treat the female distribution as the expected distribution (E), and the male distribution as the observed (O).
(Chi squared) = sum[ (OE)^2/E ]
10 degrees of freedom.