Is this a legitimate use of a chi-squared test?

I would like to compare two geographical areas to determine if a mosquito control method had any impact on the number of positive containers found holding mosquito larvae.

Prior to the treatment both areas were surveyed and the Breteau index (# of Aedes aegypti positive containers per 100 yards surveyed) calculated.

Area 1 : 10878 yards surveyed
66 positive containers
Breteau index =66/10878*100 = 0.61

Area 2 : 13,241 yards surveyed
86 positive finds
Breteau Index =86/13241 * 100 = 0.65

I would like to show that prior to the treatment there was no statistical difference in the number of positive containers found.

I considered using a Chi squared test 2 x 2 contingency table to compare the two areas


AREA 1 66 10878

AREA 2 86 13241

Chi-square with Yates correction, df=1, χ2=0.111, two tailed P value =0.7389.

P>.05 therefore no significant difference between two sites.

The problem I have with this is that as far as I am aware a Chi-squared test compares two outcomes whilst I am comparing two totally different entities (positive containers and yards surveyed).

The questiona I have are :

1/ Can I use a Chi-squared test to compare the Breteau indices of the two areas?

2/ Is there an alternative test I could use to compare the two areas? I have heard that a z test for proportions may be the correct test to use however I think that this is a parametric test and the data I have is nonparametric.

Thanks in advance for your help.


Alan Wheeler

Cayman Islands Mosquito Research & control Unit