Hello,
Just have a few experimental design questions which I've had no luck finding answers for :/
For titrations, how does taking the average of concordant titres inherently minimise the effect of random errors? Furthermore let's define reliablity as being 'whether a different person can achieve the same results for the same experiment under the same conditions.' So why exactly does reliabilty increase in tandem with precision?
Any help would be greatly appreciated. Cheers
1) If you have done enough trials, non-concordant results should reflect trials affected by random errors, thus by eliminating them from the calculated average, you are minimising their effect on your final results.
To elaborate on this: assume you obtain the titres 11.01ml 11.25ml, 11.05ml, 10.97ml, and 10.80ml.
in VCE chemistry we define concordant titres as grouped within 0.1ml of each other, thus we can say 11.01ml, 11.05ml, and 10.97ml are concordant (as they are all within 0.1ml of each other). The results 10.80ml and 11.25ml are not concordant, as they are not within 0.1ml of the other data, and their significant deviance from the concordant results indicates that errors occurred within the trial such to influence the results. As a result, when calculating the average, we eliminate these nonconcordant results such to attempt to lower the influence of any random errors on our final data.
2) If your experiment has a high reliability, these other people can follow you experiment and obtain the same results consistently, whereas a low reliability will mean these other experimenters will obtain varying results. Since precision refers to the grouping of one's data (consistent results = precise) the higher the reliability of the experiment, the more precise the data that will be obtained.