Maths/Statistics question. (Q-Test / T-Test)

[Kristoph]

Hey, I'm doing some statistics work here at the moment, and wikipedia has explained it fairly well but I need some more help understanding it.

Q-Test...

With the Q-test, my values for one set are:

0.2, 0.2, 0.2, 0.3

Well for the Q test it says...

Q = Qgap/Qrange

Where Qgap is the absolute difference between the outlier in question and the closest number to it. If Qcalculated > Qtable then reject the questionable point.

Well my outlier in my set of data is 0.3, although on wikipedia's example it uses the lowest value in the data set as it's outlier.

If I put my values into the equation...

Q=gap/range
Q=(0.2-0.3)/(0.3-0.2)
Q= -1


So referring to the table for Q...

If Qcalculated > Qtable then reject the questionable point.

Well I'm doing 95% significance, so the value for Q95% for 4 values, is 0.829.

Q= -1

So thats screwed me right up... Can anyone explain?

 

[D.P.]

I don't really get what you are doing but from that data set of onlz 4 numbers you will never ever be able to say 0.3 is an outlier in any statistical distribution.

 

[Kristoph]

Above you can see I've got my Samples, 1-10.. and for each sample I have 4 replicates. I'm trying to use the Q-test to find anomalies within my 4 replicates. If thats even possible.

The values are of absorbance of light in a colourimeter, we're testing for nitrates in a sugar sample thats been diluted in water.

This is getting confusing.

 

[Kristoph]

Also, from my table I've got my standard deviation value, which was worked out using the 4 replicates for each sample. I'm not too sure what thats telling me though?

Do I apply the standard deviation value to the mean? +/- 0.050 for example, to my mean value for Row 33? Meaning 0.075 +/- 0.050 would make B33 (0) an anomaly?