Thread: Statistically different??

Hello all,

I am in the process of deciding if I need to clean some samples that could be contaminated by organic material. I have tried to clean aliquots of two samples by two processes (A and B) and if they were contaminated, after their cleaning I should get higher than the original sample values. (value±2sigma).

Sample 1 without cleaning treatment gives 55.70±3.42 and after being cleaned through PROCESS A gives 60.48±3.56

Sample 2 without cleaning treatment gives 17.52±2.29, after being cleaned by PROCESS A gives 12.92±2.21 and after being cleaned by PROCCES B gives 14.57±2.24.

Sample 2 clearly indicates that cleaning is not necessary, as I'm getting lower values (instead of higher ones) after the cleaning process. Sample 1 gives higher values after cleaning, but both values are too close for what an efficient cleaning should produce (higher by 5 or 10).

I’m trying to show that this difference is not statistically significant and that samples don’t need to be cleaned. I tried F-test in excel and got a 0.84 but I’m not sure how I must make the arrays and how to interpret this number.

Some help would be greatly appreciated.

Thank you very much

Originally Posted by huju

Hello all,

I am in the process of deciding if I need to clean some samples that could be contaminated by organic material. I have tried to clean aliquots of two samples by two processes (A and B) and if they were contaminated, after their cleaning I should get higher than the original sample values. (value±2sigma).

Sample 1 without cleaning treatment gives 55.70±3.42 and after being cleaned through PROCESS A gives 60.48±3.56

Sample 2 without cleaning treatment gives 17.52±2.29, after being cleaned by PROCESS A gives 12.92±2.21 and after being cleaned by PROCCES B gives 14.57±2.24.

Sample 2 clearly indicates that cleaning is not necessary, as I'm getting lower values (instead of higher ones) after the cleaning process. Sample 1 gives higher values after cleaning, but both values are too close for what an efficient cleaning should produce (higher by 5 or 10).

I’m trying to show that this difference is not statistically significant and that samples don’t need to be cleaned. I tried F-test in excel and got a 0.84 but I’m not sure how I must make the arrays and how to interpret this number.

Some help would be greatly appreciated.

Thank you very much

Assume that cleaning does not change the value. Then the difference in
values DV, has a normal distribution with mean 0 and SD=sqrt(SD2^2 + SD2^2).

So in this case DV=(55.7-60.48) = - 4.78, and SD = sqrt ( (3.42/2)^2 + (3.56)^2) = 2.47

As |DV|<1.96 SD, you can say that the data does not allow you to reject the hypothesis
that cleaning has had no effect in this case at 95% level of confidence (if the hyposthesis
were true the probability of rejection with this test < 5%)

RonL