It is important to understand how data are produced. Only by understanding
how data are produced are we able to make informed judgements about the
quality of that data. Data can come from a number of sources. We can
use available data (anecdotal evidence and pre-existing data); and/or
generate data from observations and/or experiments.
Before we look at using available data or generating your own data, however,
let’s quickly look at another source of data: making up
or fabricating data! As you have probably assumed, this
is not considered the right thing to do. However, it
has not stopped some very famous researchers from doing exactly that.
Let’s look at a famous example.
If in the example above you thought it was suspicious that Burt's results
were the same, even though he had supposedly conducted these studies over
a period of 11 years, then you are starting to show that you understand
how important variation is to data and how impossible
it is to remove variation completely. The absence of variation between
tests conducted in 1955, 1958, and 1966 should immediately make you suspicious.
The absence of variation is even harder to accept when you remember that
Burt was trying to measure complex variables such as each twin's
intelligence and the environment the twins experienced while growing up.
Amazingly, other scientists did not question Burt's results for
another 10 years - possibly because he was such a famous researcher!
SCENARIO
Sir Cyril Burt conducted a lot of research on whether intelligence
in humans was inherited, i.e. passed from parents to child,
or whether it was due to the environment in which humans
grow up. To investigate this question he conducted research
on identical twins who were separated at birth and compared
the intelligence of one twin with the intelligence of the
other by getting them to complete a test of intelligence.
He then compared these results with the results of identical
twins who were reared together. Below are results from his
studies.
Table: Correlations for Identical
Twins, "Group Test" of Intelligence[1]
NOTE: The results for twins
reared apart showed that there was a high association between
intelligence and heredity (although the association was
even higher for twins reared together).
