Learning Objectives for Chapter Nine – The Simple Experiment

 

 

Pages 243-254

 

1.   Define¹ simple experiment. Define¹ independent random assignment. Explain² why independent random assignment enables researchers to use the simple experiment to make cause-effect statements.

2.   Compare⁴ and contrast⁴ the following:

a.   Experimental hypotheses and hypotheses that do not postulate a cause-effect relationship.

b.   Experimental hypothesis and null hypothesis

3.   State¹ what conclusions can be drawn when

a.   The null hypothesis is rejected

b.   The null hypothesis is not rejected

4.   Distinguish⁴ between

a.   Independent variable and dependent variable

b.   Control group and experimental group

5.   Explain² the implications of the need to keep observations independent on 

a.   How participants are selected, and

b.   How participants are treated.

6.   Explain² why the need for random assignment means that some hypotheses cannot be tested in a simple experiment.

7.   Distinguish⁴ between what statistically significant results mean and what they do not necessarily mean.

8.   Explain² why null results do not prove the null hypothesis.

 

 

Pages 254-261

 

9.   Define¹ Type I error, describe² what can be done to reduce the risk of making Type I errors, and explain² the tradeoff involved in trying to reduce Type I errors.

10.                 Define¹ Type II error. Define power. Then, discuss² the relationship between Type II error and power. Finally, contrast⁴ Type II errors with Type I errors.

11.                 Explain² why reducing the extent to which random error affects your study will not reduce your risk of making a Type I error.

12.                 Explain² why reducing the extent to which random error affects your study will reduce your risk of making a Type II error.

13.                 Describe² how the risk of Type II errors can be reduced. In your answer, be sure to address steps that will (a) reduce random error, (b) allow random error to balance out, and (c) allow the effect to be detected even when there is a lot of random error in your data.

 

 

Pages 261-264

 

14.                 Explain² how steps you take to increase your experiment’s power could harm your experiment’s external validity. Then, explain² how steps you take to increase your experiment’s external validity could harm your experiment’s power.

15.                 Explain² how using placebo treatments and double-blind procedures could improve the construct validity of a simple experiment.

16.                 Explain² how steps you take to increase your experiment’s power could harm your experiment’s construct validity. Then, explain² how steps you take to increase your experiment’s construct validity could harm the experiment’s power.

17.                 Explain² how steps you take to make your experiment more ethical could harm the experiment’s power. Then, explain² how steps you take to increase your experiment’s power could make your experiment less ethical.

18.                 Rank⁶ the following in how important they should be in designing a simple experiment: construct validity, ethics, external validity, and power. Justify⁶ your rankings.

 

Pages 264-272

 

 

19.                 Defend⁴ the following comment: “The average score of the experimental group is an estimate of what the average score would have been had all your participants received the treatment.”

20.                 Explain² why the following statement is true: “The mean for the treatment group could be higher than the mean for the no-treatment group even if the treatment had no effect.”

21.                 Examine⁴ why, in the simple experiment, each of the following is true:

a.   random error causes scores within each group to differ from one another.

b.   random error may cause the experimental group means to differ from the control group mean.

22.                 The means of your experimental and control groups differ.  Outline³ the factors that a statistical test would use to determine whether the difference between the means is large enough to be due to the treatment rather than to chance alone. Examine⁴ the role of the following in your answer:

a.   The size of the difference between your two group means,

b.   The amount of variability within each of your groups, and

c.    The number of participants in each of your groups.

23.                 Choose¹ which of the following experiments illustrates a difference between the experimental and control groups that is most likely to be due to more than chance alone. Defend⁴ your choice.

 

Experiment A                         Experiment B

 

Control           Experimental               Control       Experimental

4                         8                                  8                12    

5                         10                                10              15    

6                         10                                12              16    

3                         10                                6                13

4                         14                                8                18

5                         16                                10              21

 

 

Pages 272-276

 

24.                 State¹ the basic idea behind the t test.

25.                 You find the following on a computer printout of a t test analysis:  “df = 12, t = 5, p <.05, treatment group mean = 11, control group mean = 1, standard error of the difference between means = 2.”

a.   Determine³ the number of scores the computer analyzed.

b.   Determine³ whether the results were statistically significant.

c.    Show³ the numerator (top part) and the denominator (bottom part) of the t ratio.

d.   Compute³ an index of effect size (see Box 9-2). Using  the effect size you calculated, assess⁶ whether the treatment had a large, medium, or small effect.

 

26.                 Referring to the concept of effect size, explain² the difference between statistical significance and practical significance (having an important effect).

 

Box 9-3 (pages 277-82)

 

27.                 List¹ the three general types of replies that advocates of statistical significance testing use in response to attacks on significance testing. Defend⁶ the statement: “Psychologists should continue to use statistical significance testing.”

28.                 List¹ three objections to statistical significance. Rank⁶ them in terms of how serious you think they are. Justify⁶ your rankings. Defend⁶ the statement: “Psychologists should no longer use statistical significance testing.”

 

 

Pages 276, 283-285

 

29.                 List¹ the two most essential assumptions that must be met to compute a meaningful t test.

30.                 List¹ two less serious assumptions of the t test. Then, explain² why these assumptions are usually not a threat to conducting a meaningful t test. In your answer, be sure to refer to the central limit theorem.

31.                 List¹ six questions to ask when results of your simple experiment are not statistically significant.

32.                 List¹ two questions to ask when results of your simple experiment are significant.