Re: comparing ratios

Any help on the following much appreciated, please:

I have 6 ratios such as:

Ratio 1) 0.056 : 0.088 : 0.075 : 0.122 : 0.119 :
0.116 : 0.060 :
0.086 : 0.094 : 0.185
Ratio 2) 0.081 : 0.117 : 0.079 : 0.090 : 0.117 :
0.103 : 0.089 :
0.094 : 0.083 : 0.148
Ratio 3) - 6) etc.....

i.e. All ratios have 10 parts; all parts add up to 1.

I am interested in the difference between each ratio
i.e.

Is ratio 1 significantly different from ratio 2?
In the wider group of 10 ratios, which are more
similar; which are more different (i.e. looking for
patterns of similar versus different ratios).

What test do I need to run for this?
(I am using SPSS).

Many thanks in advance for your help
Claire


What you have is a 10 by 6 contingency table. Assuming that you have the six N's, you can test that the six multinomial distributions are equal by a chi-square test with 9*5 = 45 degrees of freedom.

Jack
.

Relevant Pages

  • comparing ratios
    ... All ratios have 10 parts; all parts add up to 1. ... In the wider group of 10 ratios, which are more similar; which are more different (i.e. looking for patterns of similar versus different ratios). ... (I am using SPSS). ...
    (sci.stat.math)
  • Re: Faktoenanalyse mit "Rohwerten"
    ... Am 05.11.2005 15:14 schrieb Oliver Guenther Stangl: ... > Mittelwerte und Streuungen über sämtliche Items konstant sein ... > FA für Ratios ableiten? ... ich habe das mit SPSS ...
    (de.sci.psychologie)
  • Re: logistic regression on spss
    ... the binary logistic regression function on spss? ... education are on different levels (less HS, HS grad, college grad). ... Will this give me odds ratios for each of my levels? ...
    (sci.stat.consult)