Which statistic test is best for this hplc measurment ?

[Newchromatographer]

Hi , i have standards ( 1-100 ppm ) and in each concetration my analyte was treated at four different treatment

so concetration 1 has treatment 1,2,3 and 4
concentration 2 has also four treatment

which test can i use because if i use ANNOVA that's mean that i will test only one concentration at different four treatments but what i need to do is to test all conc. not only one and say that all four treatments show no different in the four treatment with all concentration ? how ? which

please help me thanks

[aceto_81]

ANOVA is the way to go, but you need a 2-way ANOVA, and check only for one effect.

Ace

[Jumpshooter]

Conduct an F-test (ANOVA) on the four treatments for each conc'n Class separately. You will end up with an F-statistic that says either "there is no significant difference between the variance in values due to treatments-- trt 1=trt2=trt3=trt4" or "at least on treatment gives a significantly different response compared to the others". If so, Then inspect the values and variance---more than likely you will see a noticeable in the response values diff between treatment 1 vs treatment 4. Then, simply followup the comparison of trt 1 vs trt 4 with a Post Hoc test to substantiate the visual difference in a quantitative manner.

[Newchromatographer]

Guys can illustrate more please because i still don't get it

it is easy to say at 1ppm the four treatments show no different but do i need to test each concentration individually ? is there is any examples in the internent or some books i will be very gratefull

thanks

[aceto_81]

Guys can illustrate more please because i still don't get it

it is easy to say at 1ppm the four treatments show no different but do i need to test each concentration individually ? is there is any examples in the internent or some books i will be very gratefull

thanks

Just google: "2-way anova example", the first hit:
http://www-stat.stanford.edu/~susan/cou ... xanova.pdf

The paper shows how to do it by hand or by using R (www.R-project.org).

Ace