Uncertainty and Significantly Different Means - Buffer Solutions



Hi there

I have discovered an "interesting" (meaning I do not know how to proceed) problem.

I am estimating an uncertainty budget for a measurement of buffer solutions. We measure the buffer solutions with two pH electrodes in the same solution at the same time. Temperature is controlled (25°C +/- 0.2). The electrodes are calibrated using primary standard buffer solutions with pH values either side of the solution being measured. Around 300 measurements are collected over 15 minutes after which the mean of the measurements of each electrode are combined to deliver the total mean, which is attributed to the pH of the buffer.

Now my problem. The method that I have just stated, I believe, is valid IF there is no significant difference between the measurements from the individual electrodes. Unfortunately, there is a significant difference.

In practical terms, the differences are small (i.e. less than the uncertainty of the primary standard buffer) but never the less statistically significant.

My question is can I use distribution of the difference between the buffers to calculate the measurement uncertainty of the electrodes or because of the statistical significance, must I go back to the drawing board and re-examine my measurement method.

Any ideas are much appreciated.


Guardbanding? You will have to help me more here, I have never heard of guardbanding. Could you post a link please?
Top Bottom