|
Developing a
Research Plan
A procedure provides a description of how to organize a
scientific experiment to test a hypothesis.
The idea for research can be refined through library
investigations, including the use electronic information such
as e-mail and the internet, retrievals and reviews of
scientific literature, and through peer feedback obtained from
review and discussion of the problem, hypothesis, and
procedure for experimentation before it is conducted.
Development of a
research plan involves researching background information and
understanding the major concepts in the area being
investigated. Recommendations for the methods used for
studying the problem, methods
of study, technology selection and use, choice and use of
proper equipment, and appropriate safety precautions should be
included in the design of the experiment.
Statistical
Analysis of Results
It
is important to use
statistical techniques to attempt to determine whether chance
alone has resulted in the data obtained in a scientific
experiment. This allows the investigator to assess
correspondence between the predicted result contained in the
hypothesis and actual result, and reach a conclusion as to
whether this hypothesis was supported.
One
technique is to obtain an average or mean value for the
data obtained. The more trials of an
experiment that are conducted, the more acceptable or valid
that data becomes. A minimum of three trials
should be conducted of any quantitative
investigation. The average value of this
data allows for the minimizing of any wildly different results
in a particular trial.
Representing
the Data
The type of graph used to represent the data will depend on
the kind of data collected. Some types of graphs and
techniques for the construction of these appear below.
|
Pie Graphs |
| Pie graphs
are great for displaying parts of a whole. The pie
graph below displays the percentage of gases in our
present day Earth atmosphere.
 |
|
Line
Graphs |
| Line graphs are used to compare different sets
of related data or to "predict" data that is not
directly measured. This may involve the extrapolation
by extending the graph along the slope above or below
the plotted data on the graph. Line graphs may
be also used to predict data between two plotted points on
the graph. Some samples uses of these
techniques are illustrated in the graph below.
One could use the graph above to determine the relative
rates of activity of gastric protease and intestinal
protease are equal at pH 5.0. One could
also determine that the activity of intestinal protease
stops at a pH of 12. |
|
Technique for Constructing a Line Graph |
1. Identify the variables to be
plotted
independent variable -- the variable manipulated by the
experimenter
--is plotted on the x-axis (horizontal axis)
dependent variable -- the factor responding to changes
in
the independent variable
-- is plotted on the y-axis (vertical axis)
2. Determine the scale of the axes
-- determine each axis individually
-- may easily be determined by taking the largest value to
be plotted
and dividing by the number of blocks
and then rounding up to the
nearest convenient number
-- the graph should be spread to occupy the most available
space
3. Number and label each axis
indicating the appropriate units.
4. Plot each data value on the graph
with a point. 5. Draw a line that best fits the
data points.
-- do not connect
data points to the origin unless there is data
to support this
-- if possible, do
the graph of experimental data as a "best
fit" line for the points which have been plotted.
6. Provide a title which clearly
indicates what the graph is about.
7. If the graph has more than one set of
data, provide a key
to indicate what is
represented by the different lines. |
|
Bar
Graphs |
| Bar graphs provide another way of organizing
data. It allows the taking of several
measurements of different items and then making a
comparison of them. A histogram is
similar to a bar graph, but only involves one variable
which is compared. A histogram allows the
sorting of data by categories, as in the example below:
 |
Lab Drawings
Many lab reports will involve a requirement for some type of
drawing. A drawing should convey information
about the experiment that the formal writing in the report
does not. The drawing must be large enough
and detailed enough to convey this additional information.
Assessing
Experimental Results
The results of a
scientific experiment must be presented to the public and
peers (other scientists) before they can be
accepted. An assumption of science is that other
individuals could arrive at the same explanation if they had
access to similar evidence. The procedure used in
conducting the experiment must be stated precisely enough to
allow other scientists to perform the experiment and determine
whether the results obtained are repeatable. The
written report for public and peer study should describe the
proposed explanation, literature reviewed, the research
carried out, its result, and any suggestions for further
scientific research.
Scientists use peer
review to assess the results of scientific investigations and
explanations proposed by other scientists. It is important the
other scientists critique original research conducted by
scientists in this manner. They analyze the
experimental procedures, examine the data obtained in the
experiment, and identify faulty reasoning in assessing the
data. Peer review also leads to scientists pointing out
any conclusions that go beyond the evidence obtained.
They also may suggest alternative explanations for the same
observations.
This peer and
public discussion may lead to revisions of the
explanation provided by the research and lead the
scientist to additional research related to the original
problem being investigated. Therefore,
hypotheses are valuable, even if they turn out not to be
true, because frequently they lead to further
investigation.
Questioning Claims
The claims made in a scientific investigation should be
questioned if the data are based on very small
samples. Claims made by individuals having bias
must be questioned. Bias means to
have an opinion about the experimental results before
the investigation which will lead to the
misinterpretation or manipulation of the data obtained
in the experiment. An improperly
controlled scientific experiment must be
questioned. The experiment should contain at
least one clear control and one independent
variable. The conclusions obtained in an
experiment must be questioned if they are based on the
faulty, incomplete, or misleading use of numbers.
Fact and opinion must not be intermingled in a proper
scientific experiment. The results of the
experiment must cite adequate evidence and have its
conclusion following logically from this evidence.
|
