Drawing Conclusions and Identifying Bias and Errors

1
Drawing Conclusions and Identifying Bias and
Errors
2

• A good experiment is designed so the experimenter
  can reach a valid conclusion. A valid conclusion
  is one the experimenter and other people can
  trust.

3

• Conclusions can be trusted if no error occurred
  during the experiment. Errors may include
  misinterpreting the cause of an effect, or
  carrying out a faulty procedure.
• Bias can also lead to error.
• Bias is a wish, conscious or unconscious, to have
  an experiment lead to a certain conclusion.

4
Drawing Conclusions

• An experiment is conducted to test a hypothesis.
• A conclusion states whether the results of an
  experiment support the hypothesis. The conclusion
  should take the independent and dependent
  variables into account. In other words, the
  conclusion should state whether and how the
  dependent variable changed as the independent
  variable was manipulated.

5
Example

• You test a hypothesis that increasing the
  temperature of a gas changes the volume of the
  gas. In your experiment, you measure and record
  the volume of a certain amount of gas at
  different temperatures.

6

• Lets look at the graph

7

• By studying the data presented in the graph, you
  can conclude that the data support your
  hyypothesis. As temperature increases, the volume
  of the gas also increases. This conclusion
  relates how the dependent variable (volume
  changed as a result of the changes in the
  independent variable and the dependent variable
  is a cause and effect relationship. Increasing
  temperature causes the volume of a gas to
  increase.

8

• From this information, you can also make
  predictions. If you look at the graph, you can
  predict the volume of the gas at temperatures in
  between the points you have measured. You simply
  draw a new point on the line. Then you read
  directly down to find the new temperature.
  Although you have not measured volume t that
  temperature, you can be fairly confident of what
  the volume should be.

9

• When scientist publish the results of their
  research, they do more than present their data.
  It is important to describe procedures
  accurately. This allows other scientists to
  repeat the experiment that were conducted by
  someone else, they are checking the first set of
  results, all the results support the same
  conclusion.

10

• Experimental Error
• In another scientist repeats the experiment and
  gets different results, the scientists have to
  figure out why.
• They will examine the methods they used in order
  to see whether differences in results. One o the
  scientists may have made mistakes. It is
  important to assess experiment critically and to
  be as free of bias as possible. Carefully
  considering the ideas presented by other
  scientists will allow for a greater understanding
  of the processes that were followed and the data
  that resulted.

11

• An experiment can also produce false results.
• These results are often cause by the experimental
  error.
• Experimental error is a mistake in an experiment
  that can lead to false results.

12

• Experimental error can be a mistake made by the
  experimenter.
• These errors include measuring incorrectly or
  following the procedure incorrectly.
• Experimental error can also be caused by an
  independent variable that was not controlled
  during the experiment.

13

• The design of the experiment might have
  introduced two independent variables.
• In the experiment described before, lets say
  both temperature and pressure changed during the
  experiment. The experimenter could not tell which
  of the independent variable caused the change in
  volume. It could have been either variable, or
  both of them.

14
Guarding against bias

• Bias is not always conscious. People may expect
  or want their ideas to be valid, because they do
  not like being wrong.
• In some cases there are other reason

15
(No Transcript)
16
For example

• A team that invents and tests a new drug wants it
  to work. A cure for a disease can help many
  people. It may also make its inventors rich or
  famous.

17

• Scientists who are aware of their biases can be
  particularly careful to check their results.
• Hypotheses testing and using a control group help
  fight bias.
• Careful observation is also important.

18

• When testing a drug, the researchers will not
  just note that a patient said he felt better.
  They will also measure and record the patients
  symptoms and lab results.
• Another way to fight bias is double-blind
  testing. In those experiments, neither the
  patients nor the researchers know who is in the
  experimental group.

19

• The control group gets pills or injections that
  do nothing and look like the real drug. This
  avoids some experimental errors.
• People who know they are getting a drug are
  likely to feel better. Also, researchers may
  think a patient who they know got the drug is
  healthier than one who did not.
• After conducting the experiment and recording the
  results, the scientists look up which patients
  got the drug. This helps them reach a valid
  conclusion.

20

• Suppose you have a cold and someone tells you to
  drink this tea. If you wake up in the morning and
  feel better, you might tell everyone it was
  because of the tea, however, this is not a valid
  conclusion.

21

• What should you do if your conclusion does not
  support your hypothesis?

22
(No Transcript)