Important Parts of the Experiment.
To begin a scientific investigation a scientist must first pose a question that he/she wants to answer.
Scientists base these questions on observations they have made, information gathered in their previous
research or the previous research of others or a combination of these two. Discuss the following
questions with your lab team and decide which of them are questions that can be answered using the
Are rapists evil by nature?
What is the cause of food poisoning?
Why do trees lose their leaves?
What is the best recipe for Chicken Marsala?
When will a big earthquake hit Charleston, S.C.?
Does watching TV cause children to have shorter attention spans?
How can the productivity of a corn crop be increased?
How did you decide which ones could be answered with the scientific method?
While a scientific investigation may have its origins in the form of a question what is used to generate
the experiment is usually a hypothesis. As discussed in class this is an educated guess as to the answer
to the question. The hypothesis must be testable and falsifiable. If the question that generated the
hypothesis can be answered no then the hypothesis will be usable. While we can prove a hypothesis is
false we cannot prove a hypothesis is true we can merely support the hypothesis.
Scientific knowledge is not regarded as absolute truth but rather an accumulation of evidence in support
of the hypotheses.
Can you think of a scientific study you read about that was later overturned?
Which of the following would be useful as scientific hypotheses? Justify your answers with a reason.
Plants absorb water through their leaves as well as through their roots.
Cats are happy when you feed them steak.
The more intelligent an animal the easier it is to domesticate.
The universe was created by an all-powerful being.
Bird flu can be transmitted by cat fleas.
An active volcano can be prevented from erupting by throwing a young girl between the ages of 12 and
16 into it during each full moon.
Types of Variables
In order to test a hypothesis one must design an experiment. An important step in doing this is to
identify the variable(s) that will be part of the process. A variable is that part of the experiment that can
be expected to change. Three types are of interest t o us: Dependent variables, independent variables
and standardized variables.
These are what is measured or counted or recorded. So for example if an investigator wanted to study
how plants grow they might chose to measure increase their height of plants to determine the effects of
an independent variable (the researcher has chosen to study) on plant growth. What are some other
aspects of plant growth (besides height) that could be measured?
The investigator might try to measure all of these. If so they would all be dependent variables in the
experiment. It is up to the researcher to determine how many dependent variables to measure.
Independent variables are what the researcher varies as part of the experiment. So in the above
example it would be amounts of nitrogen. What might be some other independent variables that one
might want to test in determining how plants grow?
As we determined when looking at dependent variables there can be data collected for more than one
dependent variable. But a researcher must choose one independent variable per experiment. Why is a
researcher limited to one independent variable in an experiment?
What was the independent variable in your black box investigation?
What was (or were) the dependent variable(s)?
What follows is a list of variables. Your group needs to determine what are dependent variables and
independent variables and then label each one.
Different groups of seeds are soaked in solutions with different pHs and their germination percentage is
determined for each group.
Height of corn plants is measured and recorded every day for 4 weeks.
Percent weight gain is recorded for a group of dogs fed different types of dog food.
The different species of snails is counted for a coastal area before and after a chemical spill.
Standardized variables are the factors that are kept the same in all of the treatments. This allows any
changes in the dependent variables to be directly related to variations in the independent variable that
the researcher manipulated. If the researcher does not keep all the other variables standardized then
he/she will be unable to clearly demonstrate that the results were the result of the independent
variable and not something else. Going back to the plant growth experiment list some of the other
variables that would need to be standardized:
Levels of Treatment
After the researcher has determined what is/are the dependent variable(s), what is the independent
variable and what is/are the standardized variables then the researcher must determine how to vary the
independent variable. So in our previous example that would mean the researcher would need to
determine the amount of nitrogen to use for each plant. The number of different treatments will
determine how many plants will be needed. These are usually based on prior knowledge of the system
either from previous research or background information.
A control treatment is where the independent variable is either not present or set as a standard value.
The results that one gets with the control treatment is compared with the results from the experimental
treatments. For example in the plant growth example the control could be the plants that had no
nitrogen added. Sometimes you can’t eliminate the independent variable. Let’s say one wanted to look
on the effects of increased oxygen levels on rats. You can’t eliminate oxygen it would have to be set a
standardized level to keep the rats alive.
For each of the following examples, tell what an appropriate control treatment would be.
1. The effect of light intensity on photosynthesis is measured by determining the levels of CO2
being used by the plant. What would be the control treatment?
2. Subjects are given papers soaked in a bitter tasting chemical. The researcher records whether
each person can taste the bitter chemical or not. What would be the control treatment?
3. The effect of Saccharin on bladder cancer development in laboratory rats is studied. What
would be the control treatment?
4. A solution is mixed up that mimics stomach acid and thus has a pH of 2. An antacid is added in
to this solution in small amounts, and the pH is measured after each addition. What would be
the control treatment?
To replicate an experiment is to repeat it many times with the exact same experimental methods and
conditions to see if the results are the same. Because biological systems have a certain amount of
variation so you should not expect to get exactly the same results every time. But by repeating the
experiment we can see how much variation there is in the system and we can average the results to
help equal out the variation.
A researcher also must take into consideration sample size. If you only have one plant in each treatment
with nitrogen and 1 of the plant dies in a particular treatment you can’t be sure it was the treatment
that caused the death of the plant. What other factors could have caused the death of the plant?
Go through the examples passed out in class and as a group answer the questions that follow each
experiment and turn in 1 copy for your group.