Inferential Statistics

Question 1

What is the purpose of a Pearson product moment correlation?

Answer 1

• Measures: linear association between two variables
• Measures of: covariance (how variables change together)
• Tests for a statistically significant relationship between two interval/ratio variables

Question 2

Example of Pearson product moment correlation Hypothesis:

Answer 2

• Ex: Is there a relationship between time students spend texting during the exam and their grade?

Question 2

What is a type I error?

Answer 2

The chance of falsely rejecting a true null

Question 3

What has to be true of the data (assumptions) in order for a correlation to work properly?

Answer 3

• Independent random pairs of cases
• Interval/ratio variables
• Normal distribution
• Bivariate normal distribution (pairs of xy – must be normal on distribution)

Question 3

What is a type II error?

Answer 3

The chance of mistakenly rejecting a true null.

Question 4

How do you interpret the magnitude of a correlation coefficient (r)?

Answer 4

• Closer (r) is to 1 or -1, STRONGER it is

Question 4

How much of a type I error should researchers allow a communication study to have?

Answer 4

5% or less

Question 5

How do you interpret the direction of a correlation coefficient (r)?

Answer 5

• Sign of (r) determines inverse or direct
• (-) inverse
• (+) direct

Question 5

How much of a type II error should researchers allow a communication study to have?

Answer 5

20% or less

Question 6

What is an effect size?

Answer 6

• % of shared variance between x and y variables

Question 6

In inferential statistics, what is the difference between systematic and error variance?

Answer 6

Systematic = true group differencesError Variance = chance differences between groups

Question 7

How is effect size calculated?

Answer 7

r^2 (r-squared)

Question 7

What is the purpose of an independent samples t-test?

Answer 7

To test for a statistically significant difference between the means of two independent groups

Question 8

Correlation v. Causation

Answer 8

• Correlation does not = Causation

Question 8

What is an example of a hypothesis that a independent samples t-test could be used to test?

Answer 8

Students who have taken speech class will be better speakers than students who have not taken speech class.

H1: Mu class> Mu no class

H0: Mu class= Mu no class

Question 9

3 conditions of Causation:

Answer 9

• Time: (IV) must precede (DV)
• Covariance: IV and DV must go together in meaningful way
• Control: Changes in DV must be caused by changes in IV, not some other variable

Question 9

What must be true of the data (assumptions) in order for an independent samples t-test to work properly?

Answer 9

1. Independent Random Sample
2. Interval or Ratio dependent variable
3. Normal Distribution
4. Homogeneity of Variance

Question 10

May be correlated because…

Answer 10

• Y causes X (opp. direction)
• X causes Y which causes X (non-recursive)
• 3rd Z factor • Coincidence: spurious relationship

Question 11

If a researcher mistakenly concludes there is a significant difference or relationship, what type of error has been committed?

Answer 11

Type I Error

Question 12

What is a “normal” distribution?

Answer 12

A distribution (histogram) with no skew or kurtosis

• Symmetrical (the right and left halves look the same)
• Mesokuric (not too peaked or flat)
• Perfect bell-shaped curve

Question 13

True or False?

There is a strong positive correlation between the amount of birthday cake a person has eaten and near-sightedness.

Therefore, eating cake causes poor eye sight.

Answer 13

False

Correlation does not equal causation.

Being old probably causes both. Older people have had more birthdays and tend to have near-sightedness. (Third factor z)