Civil Engineering Reference
In-Depth Information
EXAMPLE
The Hawthorne studies constitute a good example of evaluation research. These studies,
which were undertaken around 1930, are famous since they supposedly illustrate that a
caring supervisor can increase workers' job motivation and productivity at work.
There were seven female workers in a plant. They manufactured transformers.
Researchers thought that the productivity could be increased if the illumination level at
the workstations was improved. The illumination level was increased in several steps
over some time. Each time the productivity increased, and the conclusion was that
increased illumination improves productivity. To check the reliability of the findings the
investigators lowered the illumination level. The surprising result was that productivity
improved even more. The investigators therefore arrived at a new conclusion: the reason
for the improved productivity was not the illumination level, but rather the concern that
the supervisors had for the well-being of the workers. This increased job motivation and
thereby productivity. This group of workers was indeed treated differently from other
workers in the plant and there were many conversations concerning the effect of the
illumination and how the workers liked it.
Many years later Parsons (1986) reanalyzed the findings. He pointed out that the
women were paid piece rate for each transformer. Behind the workstation was a counter
showing the number of finished transformers, so that each worker knew exactly how
many she manufactured each day. There was also a competitive spirit among the workers
and the payment system also provided incentives for the women to work fast. Some of
the improved productivity was probably a training effect; workers learn a job and they
work faster (see Chapter 15). Neither illumination nor job motivation were very
important in this case—although they could have played a minor role. The feedback from
the counter and from the salary increases were more important.
This example demonstrates the difficulties in drawing firm conclusions from evaluation
studies. An experimental study, which compares different groups of workers who
received different “treatments,” would have been more informative. There could, for
example, have been three illumination levels and three groups of workers. In real life it is
very difficult to arrange these studies, as there is usually little incentive for the owners of
a plant to participate.
3.5
SELECTION OF DEPENDENT VARIABLES
In selecting a dependent variable, there are three main criteria that must be fulfilled: it
must be reliable, valid, and sensitive.
RELIABILITY
Here we refer to the consistency or stability of a measure over time. For example, to get
entrance to a university a student ay need to take the Scholastic Aptitude Test (SAT). If a
student takes the test many times, the test score should be about the same each time; SAT
