Hypothesis%20Testing

1
Hypothesis Testing

• Chapter 6

2
Steps for Conducting Scientific Research
Step 1 Formulate a hypothesis
Higher temperatures cause people to be more
irritable
Step 2 Design a study
Step 4 Report the results
Step 3 Collect analyze data
3
Study Design

• This study will examine the effects of
  temperature on irritability. Subjects will be
  randomly assigned to one of three temperature
  conditions (low 70o-72o, moderate 80o-82o, or
  a high 90o-92o temperature room).
• While in the room an assistant will irritate the
  subjects.
• Subjects will then complete an irritation scale.

4
Hypotheses

• The Study Hypothesis
• Ha The amount of irritation subjects report will
  be significantly different among the three
  temperature conditions.
• The Null Hypothesis
• Ho There will be no significant differences in
  the amount of irritation reported by the
  subjects among the three temperature conditions.

5
Sampling from the Population
Population
6
Operationalize and Measure Variables

• Independent Variable Temperature
• 3 Conditions
• Low (70 - 72) Condition 1
• Medium (80 - 82) Condition 2
• High (90 - 92) Condition 3
• Dependent Variable Irritation
• Operationalized as ratings of irritation in
  response to assistant
• Scale included in Appendix of paper

7
Irritation Scale
During the experiment, how much did the assistant
make you feel the following
Not at All A Little Bit Moderately Quite a Bit Extremely
irritated or annoyed? 1 2 3 4 5
furiously angry? 1 2 3 4 5
mad? 1 2 3 4 5
so angry you felt like hitting the assistant? 1 2 3 4 5
8
Creating Scale Scores
Subject Condition Item 1 Item 2 Item 3 Item 4 Mean Score
1 3 3 4 5 3 3.75
2 2 2 3 2 3 2.5
3 1 1 2 1 2 1.5
Subject Condition Mean Irritation Score
1 3 3.75
2 2 2.5
3 1 1.5
9
Describing the Data
10
Hypothesis Testing

• What could cause the difference in ratings
  between condition 1 and condition 3?
• Chance
• error
• individual differences not evenly distributed
• Confounds
• The independent variable (treatment)

11
Did the Results Occur by Chance?
The treatment effect

Between groups variance
Within groups variance
F-ratio

Within groups variance
Variation due to chance (error). Typically caused
by individual differences in participants
By controlling within-group variation, scientists
can isolate the effects of the independent
variable on the dependent variable
(between-groups variance).
12
Inferential Errors
True state of affairs True state of affairs True state of affairs
Chance is responsible Chance is not responsible
Null hypothesis decision Fail to reject Correct acceptance Type II error
Null hypothesis decision Reject Type I error Correct rejection
13
Presenting Tabular Results
Table 1 Analysis of Variance for Irritation
Source df SS MS F p
Irritation(I) 1 640 640 25.41 .001
S within group error 36 94 2.61 __ __
14
Presenting Graphic Results
(F 5.0, p lt.05)