statistical methods for normal distribution
Firoz Ahmed
Thursday, 11 June 2009 17:20 UTC
Hello!
I want to know which statistical method can be used to check the mean of a single variance. Suppose I have carried out an experiment (biological) 4 times. Got different values eg 8.5, 10, 7, 6. How could we know that the average 7.87 is statistically significant (with p-value). That means the conclusion drawn based on these experimental data is reliable.
-
Replies
-
What do you mean by “statistically significant”? What is the question you’re trying to ask?
-
Thanks, we generally replicate experiment (four-five times) and took the average of results. Based on the average the conclusion is made. suppose there are five output values of experiment: 10, 30, 15, 12, 11. Now I just want to know whether the mean value 15.6 is reliable to make conclusion. How could we know the confidence of these data.
-
Describing the mean as “reliable” id a bit vague. But you could simply calculate the standard error: that will give you a summary of the amount of variation you have in the statistic.
-
I think things would be clearer if we understood the question asked of the experiment. For instance, do you a priori hypothesize some sort of expected observation? For instance, if you expect to see an observation of 5, but your experiments leads you to an average observation of 7.87, with a standard deviation of 1.75, a one-sample permutation test (two sided, 5,000 permutations) testing the null hypothesis that the data are equal to 5.0 (alternative that the true mean is not equal to 5) would lead to a test statistic of 11.5, and a p-value of 0.1276. Or did you intend for there to be some other sort of inference?
-
Thanks Wayne for your valuable suggestion. You have described T test for two means. As in my case there is no prior hypothesis or no comparison between two means. However, I got the solution from STATISTICA package by using T.test for single mean; mean =7.87, stdev= 1.75, std error=0.875, t-value=9.0, p=0.002896).
In two sample T-value is calculated by [(mean of A- mean of B)/std error of difference]. In this case T-value is calculated by [mean/std error].
Thanks all of you
-
