STATISTICS IN THE LABORATORY continued
Figure 3 is the engineer’s egotistical view of the world. In this universe, the industrial process has zero variation (engineers know how to control things!), and all the observed variation is attributed to those untrained, inept, sloppy analytical chemists and their strange methods that they call measurement.
This view of the world is also unrealistic (industrial processes don’t have zero variation), but it does illustrate a very important point for analytical chemists: excessive variation in the measurement process can prevent the engineers from understanding what their industrial process is actually doing. All they “see” are the graphs in the upper right. They don’t get to “see” the two graphs above the industrial process, the graphs that show what the industrial process is actually doing. As we discussed in a previous column, this can lead to “tampering” of the industrial process—observing changes that are actually a result of variation in the measurement process, thinking that these changes represent the behavior of the industrial pro- cess, and then taking inappropriate action on the industrial process. (We could devote a whole future column to a discussion of all the problems tampering causes.)
Figure 2 – An analytical chemist’s view of the world.
chemists take great pride in our precision!), and all the observed variation is attributed to those darned engineers and their flaky industrial process.
Although this view of the world is unrealistic (measurement processes don’t have zero variation), it actually represents what engineers would like us analytical chemists to provide—perfect measurements! If the mea- sured values we gave them really did represent their industrial process exactly, then the engineers would have a good understanding of their process and its behavior, and they’d be able to see clearly the effects of any changes that they made. Hold that thought.
Figure 4 represents reality: industrial processes have variation, and mea- surement processes have variation. I’ve chosen this particular figure to illustrate two points.
First, in my judgment, this is a picture of what a lot of biopharmaceutical companies are faced with. Biological measurements are often inherently highly variable—microorganisms (often the basis of many biological methods) exhibit exponential growth, and this makes the results fragile with respect to the timing of different steps in the analysis; very small volumes are used for analysis, and micro-pipettes don’t always perform as precisely or accurately as advertised; blunders are often made when samples are transferred to the top wells of 96-well plates, etc. I’m not
Figure 3 – An engineer’s view of the world. AMERICAN LABORATORY 24
Figure 4 – The real world. JANUARY/FEBRUARY 2017
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68