How do we gather evidence to evaluate an hypothesis

1
How do we gather evidence to evaluate an
hypothesis?

• SAMPLING
• random samples
• representative samples
• convenience samples

2
How do we design a study?

• experimental studies vs. correlational studies

3
Experiment

• random assignment of subjects to condition

4
Experiment

• Manipulation of INDEPENDENT VARIABLE

5
Lang et al.

• The effect of alcohol on aggressiveness

6
The effect of alcohol on aggression

• HYPOTHESIS
• Increased alcohol consumption leads to more
  aggressive behavior

7
Operationally defining variables

• INDEPENDENT VARIABLE
• Alcohol consumption
• Number of ounces of alcohol swallowed per hour

8
Procedure

• Subjects carried out complex task
• Subjects insulted by research assistant
• Subjects given opportunity to punish assistant
  for poor performance on a task by administering
  electric shock

9

• DEPENDENT VARIABLE
• Aggressiveness
• How to operationally define?

10
Results

• Group 1
• Thought they were drinking alcohol (but they
  werent)
• ____________________
• No Alcohol
• EXPECTATION

• Group 2
• Thought they were drinking plain tonic
  water(they werent)
• ____________________
• ALCOHOL
• No Expectation

11
Voltage Administered 1(EXPECTATION) 2 (ALCOHOL)

• 0
• 75
• 150
• 75
• 320
• 270
• 270
• 450
• MEAN 201.25

• 0
• 0
• 15
• 75
• 150
• 75
• 15
• 0
• MEAN 41.25

12
Do the groups differ in the amount of voltage
administered?
13
Do the groups differ?

• t-test
• How likely is it that the observed difference
  between group means occurred merely by chance?

14
t-test

• takes into account
• size of the difference between means
• number of observations (scores, people)
• variance of the scores

15
Why does the size of the difference between means
matter?

• Larger differences are less likely to occur by
  chance
• 201.25 vs. 41.25 as opposed to 201 vs. 195

16
Why does the number of observations matter?

• The larger the number of observations, the more
  likely it is that the mean is representative of
  the sample

17
Voltage Administered 1(EXPECTATION) 2 (ALCOHOL)

• 0
• 75
• 150
• 75
• 320
• 270
• 270
• 450
• MEAN 201.25

• 0 mean0
• 0 mean0
• 15 mean5
• 75 mean22.5
• 150 mean48
• 75
• 15
• 0
• MEAN 41.25

18
Why does the variance matter?

• Variance spread-out-ed-ness of scores
• The more spread out the scores, the less likely
  it is that the mean is a good representation of
  the sample

19
Amount of voltage administered High variability
Mean 172.5
450
15
75
150
20
Low variability
Mean 45
15
45
75
21
Correlational studies

• Predicting one variable from another

22
Perfect Positive Correlation
r 1.0
23
r -1.0
24
No correlation r0
25
Limitations of Science

• probabilistic evidence
• facts theory-laden
• theories underdetermined by fact