I think it can be applied. Because there isn't anyone prerequisite such as random sampling chi-square for apply it. For more information you can see http://www.statisticshell.com/html/apf.html

thanks all

Hi,

Kindly see the following:

http://www.planta.cn/forum/files_planta/2_statistical_tests_as_inappropriate_tools_for_data_analysis_performed_on_non_random_samples_of_plant_communities_188.pdf

AbdelRahman -

Using a p-value is easily misinterpreted even in the best of situations, as it is a function of sample size, and needs further analysis.

With nonprobability samples you may still find a great deal of useful information, especially if stratified/grouped appropriately, and if you have continuous data variables, by first looking at some scatterplots comparing variables where one or more x-values can be used to 'predict' the y-value, and you can estimate the variance of the prediction error for each such regression model using, for example, its square root as estimated in SAS PROC REG as STDI. I am not aware of what data you would have that would be continuous though, but you might look for similar correlations, remembering that correlated does not necessarily mean there is causation.

Descriptive statistics and graphical displays could lead you to a follow-on study that may provide better results from a statistical science point-of-view.

Cheers - Jim

Article Practical Interpretation of Hypothesis Tests - letter to the...

thanks all

Not sure. I am still reading statistic research methods.

Randomization is an important assumption for use of chi square test