Comparing Pipettor Precision and Accuracy
Every now and then, someone will voice the opinion: "I would never buy Brand X pipettor because it isn’t as accurate as
Brand Y." We all have our favorite brands but are there really some pipettors that are inaccurate and imprecise being marketed
today? Let’s start by reviewing what the terms precision and accuracy mean.
What is Precision and Accuracy?
This question applies in many scientific disciplines. The following explanation was voted "best explanation" in Yahoo
Answers:
The two have nothing to do with each other; keep that in mind.
Precision is a measure of how close together a number of values are. The closer together they are, the better the precision.
Example:
Data Set 1: 2.0, 3.0, 3.4, 2.7, 3.9 avg = 3.0
Data Set 2: 2.8, 3.2, 3.0, 2.9, 3.1 avg = 3.0
In both sets, the average value is 3.0, but in the second set, the precision is MUCH better since the values are closer
together. One way to measure the precision of a set of results is called the standard deviation.
Accuracy is how close a result comes to the "true" value.
Data Set 1: 2.3, 2.2, 2.5, 2.0, 2.0 avg = 2.2
Data Set 2: 2.6, 3.0, 3.6, 3.4, 2.9 avg = 3.1
Suppose that the actual (true) value is supposed to be 3.2. The average for Data Set 2 is very close to that target value
so we say that the result is accurate. Note that Data Set 1 has better precision than Data Set 2 (i.e. the data Set 1 numbers
are closer together), but the average result is NOT accurate (2.2).
What you like to see in a set of results is good precision AND good accuracy.
And this is especially true in comparing pipettors. However precision and accuracy are often expressed as CV or coefficient
of variation. What’s this? Here is the definition taken from Wikipedia:
Definition of 'Coefficient Of Variation  CV
The coefficient of variation (CV) is defined as the ratio of the
standard deviation
to the mean
:
It shows the extent of variability in relation to mean of the population.
This adapts well in pipetting. If you know the target value ()
for example 100µL, you pipet n number of replicates and you can compare one pipettors performance to another right? Let’s
first take a closer look at whether CV or SD is the better metric.
Comparison to Standard Deviation
Advantages
The coefficient of variation is useful because the standard deviation of data must always be understood in the context
of the mean of the data. Instead, the actual value of the CV is independent of the unit in which the measurement has been
taken, so it is a dimensionless number. For comparison between
data sets with different units or widely different means, one should use the coefficient of variation instead of the standard
deviation. [This is the case in variable volume pipettors!]
Disadvantages
When the mean value is close to zero, the coefficient of variation will approach infinity and is hence sensitive to small
changes in the mean. This is often the case if the values do not originate from a ratio scale.
Comparison of Three Pipettors Using Published Specifications
One of the most unusual pipettors on the market is the Vistalab Ovation pipettor. It’s designed to be ergonomic and reduce
repetitive stress injury. It has the added advantage of not needing a rack as it stands upright and ready for reuse on the
bench. Maybe because it looks so different, some people think that it is less accurate than a more traditionally designed
pipettor.
Figure 1: Precision and Accuracy Data for Vistalabs Ovation™ Manual Adjustable Volume Pipettor
Figure 2: Precision and Accuracy Data for Thermo Labsystems Finnpipette Novus™ Electronic Pipettors
Figure 3: Precision and Accuracy Data for Gilson Pipetman™ Classic pipettors
**ISO 86551:2002
specifies the general requirements for pistonoperated volumetric apparatus. It is applicable to piston pipettes, piston
burettes, dilutors and dispensers. It furthermore defines terms for the use of pistonoperated volumetric apparatus and
gives a list of equivalent terms. It also gives user recommendations.
So what conclusions can you draw from this data?
First, that there is a published ISO specification for the maximum allowable error in a pipettor. However ISO introduces
yet another new statistical term, systematic error and random error. Let’s return to Wikipedia for a comparison of the two:
Systematic Versus Random Error
Measurement errors can be divided into two components: random error
and systematic error. Random error is always present in a measurement. It is caused by inherently unpredictable fluctuations
in the readings of a measurement apparatus or in the experimenter's interpretation of the instrumental reading. Random errors
show up as different results for ostensibly the same repeated measurement. They can be estimated by comparing multiple measurements,
and reduced by averaging multiple measurements. Systematic error cannot be discovered this way because it always pushes
the results in the same direction. If the cause of a systematic error can be identified, then it can usually be eliminated.
Because random errors are reduced by remeasurement (making n times as many measurements will usually reduce random errors
by a factor of √n), it is worth repeating an experiment until random errors are similar in size to systematic errors. Additional
measurements will be of little benefit, because the overall error cannot be reduced below the systematic error.
The parallel here is that systematic error is a measure of accuracy. Random error is a measure of precision.
How the manufacturers data is displayed
Vistalabs – displays the measurements in percent volume measured at the lowest, middle and highest volumes
of each pipettor. For the 100uL pipettor at 100µL, the Accuracy (Systematic Error) is +/ 0.8%. The precision (random error)
is +/ 0.15%. ISO 8655 is not referenced.
Finnpipette – displays the measurements at unspecified volumes. Instead, they provide the extent of percent
variation over the range of the pipettor. For the 100µL pipettor the accuracy is a minimum of 0.8% and a maximum of 3.0%
and the precision is 0.2% to 2%. ISO 8655 is referenced but only in the context of how the pipettor was calibrated prior
to the testing.
Gilson – displays measurements taken at three or four different volumes over the range of the pipettor.
They provide the maximum allowable error and to their credit, is the only vendor to publish the ISO 8655 spec for allowable
errors. At 100uL, their P100 had an accuracy of +/ 0.8 µL (or 0.8%) and a precision of less than or equal to 0.15 µL (or
0.15%). Gilson references the ISO 8655 maximum allowable limit to be +/ 0.8µL (0.8%) accuracy and less than or equal to
0.3µL (0.3%) precision.
So which is the most accurate pipettor among these three?

Accuracy (Systematic Error)

Precision (Random Error)

Vistalabs Ovation

+/ 0.8 µL

+/ 0.15 µL

Finpipette Novus

+/ 1.6 µL*

+/ 0.9 µL*

Gilson P100

+/ 0.8 µL

+/ 0.15 µL

ISO 8655 Maximum

+/ 0.8 µL

+/ 0.3 µL

* Because Thermo expresses the minimum to maximum percent over the entire volume range, there
is no way to equally compare these pipettors. For example the accuracy of +/ 1.6 is calculated as follows: 3.8%  2.2%
/2 = 1.6. However this is not valid as the variation in µL at the lowest range, 5 µL at 3.8% = just 0.19 µL. So without
knowing what the values were at specifically measured points, it is impossible to make a fair comparison solely based
on this data.

Conclusions
You have now seen why we have chosen not to display precision and accuracy data on PipetTipFinder. Because:
 There is a standard – ISO 8655 – which all reputable manufacturers adhere. If you stick to respected brands, you will get
a pipettor that exceeds ISO 8655 minimums
 Because of the many ways of expressing accuracy and precision, there is no commonality between vendors on how the data is
expressed
 All of these tests are done using water. If you are pipetting anything other than water you will not get the same results
 These tests assume pipettors that are in perfect calibration using perfect pipetting techniques with tips manufactured specifically
for the pipettor. In other words, nothing like the results you will ever get
So our advice is that it is fair to assume that all name brand pipettors are under the ISO 8655 minimum specifications
and the likelihood is that many brands you don’t recognize are too. Keep in mind, this is very mature technology. And out
final piece of advice is that you discard the notion that brand X pipettor is "too inaccurate compared to brand Y." The
only way you will ever know this is to try that pipettor on YOUR samples using YOUR tips using equally well calibrated pipettors
performing 6 – 10 replicates on at least the high, mid and low volume range of the pipettor with perfect pipetting technique
and without bias. This shouldn’t take more than a few day out of your busy schedule!