Chemistry Reference
In-Depth Information
Chapter 8
Processing and Analysis
What are your data trying to tell you?
Herb Hultin
Now that the data have been collected, they must be analyzed and processed.
Computers are wonderful instruments to process data. We take them for granted
now, but some major professors did not have a computer as an undergraduate. It is
only since the 1980s that computers have become common fixtures on professors'
desks and only since the 1990s that we had the Internet and email. Ask an old-timer
about punching cards and floppy disks, and watch them smile.
Computers are great at organizing and sorting data and at making calculations.
We use them to conduct statistical analysis, prepare graphics, in simulation model-
ing, for word processing, and many other ways. We need to familiarize ourselves
with some statistical package. SAS is a common program used in statistics classes
and many labs. It is a versatile program, although not that user friendly. Find out
what is used in the lab and learn how to use it. In addition to understanding the
mechanics of performing a statistical procedure, we must also understand the under-
lying assumptions of the tests we select and the limits of interpretation of the data.
As we become more reliant on computers, there are dangers that we do not always
appreciate. If a computer becomes a black box in which all we do is put data in and
receive results out, we may draw inappropriate conclusions. One thing many scientists
are losing today is the ability to make careful observations. Find a journal article that was
written before 1930. It will not be online but somewhere in the library stacks in old
musty, yellowing journals. These articles are characterized by detailed observation with
very little quantitative data. Also, we need to be careful of computer-generated graphs.
They tend to distort data patterns leading to misunderstandings and misinterpretations.
Plots where the origin is not 0,0 can visually exaggerate small differences. Broadening
or narrowing the
X
- or
Y
-axis can provide very different perspectives (see Fig.
8.1
).
Hand-plotting on old-fashioned graph paper can help us make sense of our data.
