Hypothesis Testing

Question 1

What is reliability

Answer 1

Reliability is how consistent a process is — you get the same results if repeated under the same conditions.

Question 2

What is validity?

Answer 2

Validity is how accurate a process is — it correctly measures what it claims to measure.

Question 3

Ways to improve reliability

Answer 3

• Increase sample size
• Repeat tests (test-retest or parallel forms)
• Use representative samples

Question 4

Ways to improve validity

Answer 4

• Use random sampling to avoid bias
• Ensure no external factors affect what’s being measured

Question 5

What is content-related validity?

Answer 5

Content-related validity checks how well a process measures all aspects of a variable.
• It requires expert knowledge of the variable.
• Valid example: A calculus test covering differentiation, integration, and applications.
• Not valid example: Asking people to report their own happiness levels (too subjective).

Question 6

What is criterion-related validity?

Answer 6

Criterion-related validity checks how well one variable predicts the outcome of another variable (the criterion).
• Valid example: Mock exam results used to predict real exam results.
• Not valid example: Using meerkat heights to predict squirrel heights.

Question 7

What are two ways to test for validity?

Answer 7

1. Content-related validity (covers all aspects of a variable)
   
   2. Criterion-related validity (uses one variable to predict another)

Question 8

What is test-retest reliability?

Answer 8

Test-retest reliability repeats the same process with the same sample at a later time.
• If results are positively correlated, the process is reliable.
• Differences may occur due to time gaps or learning effects.

Question 9

What is parallel forms reliability?

Answer 9

Parallel forms reliability gives the same sample a second set of similar questions or tasks.
• If results are positively correlated, the process is reliable.
• It may be difficult to create two equally similar tests.

Question 10

What are two ways to test for reliability?

Answer 10

1. Test-retest (repeat same process with same sample later)
   
   2. Parallel forms (use a second, similar set of tasks/questions)

Question 11

What is a type I error?

Answer 11

Probability of rejecting H0 when H0 is true (incorrectly rejecting H0)

Question 12

What is a type II error?

Answer 12

Probability of accepting H0 when H0 is true, (incorrectly accepting H0)

Question 13

Ways of reducing chance of Type I and Type II errors:

Answer 13

Type 1: decrease sig. level
Type 2: Increase sample size / increase significance level

Question 14

Determine when we reject or accept H0 regarding:
1) p-value & sig levels
2) values inside or outside critical region
3) number being larger or smaller than critical value

Answer 14

1) p-val > sig level , don’t reject H0 / p-val < sig-level = reject H0
2) when value OUTSIDE of critical region DON’T REJECT H0 / when INSIDE c.r. = REJECT H0
3) if number LARGER than a critical value = REJECT H0, if number SMALLER than critical value = ACCEPT H0

Question 15

Benefits for spearman’s rank coefficient rather than persons product moment

Answer 15

• less sensitive to outliers
• coefficient of determination unlikely to be linear (for some data)

Question 16

What is the sum of square residual

Answer 16