1. Experimental Design

2. Tips for
Points of emphasis for AP Biology experimental design essays 1. State a hypothesis. This is usually stated as an expectation of results based on the known effects of the independent variable. 2. Design and identify a control group for comparison. 3. Tell how you will hold at least two other experimental variables constant. 4. Indicate how the independent variable will be manipulated (varied) -X axis. 5. Describe how the dependent variable will be measured quantitatively (Y axis). 6. Verify your results through multiple trials or repetition on the same procedure. 7. Analyze the results statistically (means, etc.). 8. Explain why you are doing the various procedural steps. This is more important than how many grams of this or milligrams of that you are using.

3. A Case Study in Scientific Inquiry: Investigating Mimicry in Snake Populations
Many poisonous species are brightly colored, which warns potential predators
Mimics are harmless species that closely resemble poisonous species
Henry Bates hypothesized that this mimicry evolved in harmless species as an evolutionary adaptation that reduces their chances of being eaten
© 2011 Pearson Education, Inc.

4. This hypothesis was tested with the venomous eastern coral snake and its mimic the nonvenomous scarlet kingsnake
Both species live in the Carolinas, but the kingsnake is also found in regions without venomous coral snakes
If predators inherit an avoidance of the coral snake’s coloration, and artificial snakes resembling the coral and kingsnake are placed in both regions, then the colorful kingsnake will be attacked less often in the regions where coral snakes are present
© 2011 Pearson Education, Inc.

5. North Carolina South Carolina
Figure 1.25
Scarlet kingsnake (nonvenomous)
Key
Range of scarlet kingsnake only
Overlapping ranges of scarlet kingsnake and eastern coral snake
North Carolina
South Carolina
Eastern coral snake (venomous)
Figure 1.25 The geographic ranges of a venomous snake and its mimic.
Scarlet kingsnake (nonvenomous)

6. Field Experiments with Artificial Snakes
To test this mimicry hypothesis, researchers made hundreds of artificial snakes:
An experimental group resembling kingsnakes
A control group resembling plain brown snakes
Equal numbers of both types were placed at field sites, including areas without poisonous coral snakes
© 2011 Pearson Education, Inc.

7. (a) Artificial kingsnake
Figure 1.26
(a) Artificial kingsnake
Figure 1.26 Artificial snakes used in field experiments to test the mimicry hypothesis.
(b) Brown artificial snake that has been attacked

8. After four weeks, the scientists retrieved the artificial snakes and counted bite or claw marks
The data fit the predictions of the mimicry hypothesis: the ringed snakes were attacked less frequently in the geographic region where coral snakes were found
© 2011 Pearson Education, Inc.

9. Percent of total attacks on artificial snakes
Figure 1.27
RESULTS
100
Artificial kingsnakes
83%
84%
Brown artificial snakes 80 60
Percent of total attacks on artificial snakes 40
Figure 1.27 Inquiry: Does the presence of venomous coral snakes affect predation rates on their mimics, kingsnakes? 20
17%
16%
Coral snakes absent
Coral snakes present

10. Experimental Controls and Repeatability
A controlled experiment compares an experimental group (the artificial kingsnakes) with a control group (the artificial brown snakes)
Ideally, only the variable of interest (the effect of coloration on the behavior of predators) differs between the control and experimental groups
A controlled experiment means that control groups are used to cancel the effects of unwanted variables
A controlled experiment does not mean that all unwanted variables are kept constant
© 2011 Pearson Education, Inc.

11. Example #1 95% Confidence Interval Standard Error of the Mean
Note: By looking at the bar graphs, you can see that the means for the light and dark treatments are different. Can you tell if the differences are statistically significant? If the standard error bars and 95% CI bars overlap, then it is likely that the differences are not statistically significant. In this case, the standard error bars do not overlap; however, this does not necessarily mean that the difference between the means is statistically significant. On the other hand, the fact that the 95% confidence interval error bars do not overlap strongly suggests that the two sample means are significantly different. An additional statistical test, the sStudent’s t-test, is required to be sure.
Application in

12. Example #2 Standard Error of the Mean 95% Confidence Interval
Note: By looking at the bar graphs, you can see that the means for the light and dark treatments are different. Can you tell if the differences are statistically significant? If the standard error bars and 95% CI bars overlap, then it is likely that the differences are not statistically significant. In this case, the standard error bars do not overlap; however, this does not necessarily mean that the difference between the means is statistically significant. On the other hand, the fact that the 95% confidence interval error bars do not overlap strongly suggests that the two sample means are significantly different. An additional statistical test, the sStudent’s t-test, is required to be sure.
Application in