Deck 10: Statistical Significance

Question 1

Statistical significance does not mean that something is important.

Answer 1

True

Question 2

Which of the following is NOT true?

A) A probability of .001 means that there is a very strong relationship
B) All inferential statistics require random samples
C) Your alpha level does not have to be set at .05
D) All of the above are NOT true

Answer 2

A

Question 3

An inferential statistic cannot be interpreted unless you have:

A) a random sample
B) a knowledge of calculus
C) a sample size of at least 500
D) interval level data

Answer 3

A

Question 4

If you don't have the probability associated with an inferential statistic's value and degrees of freedom

A) the probability level will have to be calculated by dividing the statistic value with the degrees of freedom
B) the probability level will have to be calculated by dividing the statistic value by 2
C) you will have to find the probability in a table with the sampling error distribution
D) you will have to determine the degree of error yourself

Question 5

Random error distributions can be used to estimate the amount of error likely to occur in a sample.

Question 6

If a statistical program's probability output doesn't say differently, it is almost always a 2-tailed probability.

Question 7

A probability level of .05001 with alpha level of .05 is

A) significant
B) substantially significant
C) not significant
D) almost significant

Question 8

Under which condition below would you reject a null hypothesis?

A) the calculated probability level is larger than alpha
B) the calculated probability level is smaller than alpha
C) the calculated probability level is within one standard deviation of alpha
D) the direction of the relationship is negative

Question 9

It is a more conservative decision to

A) limit the number of statistics tests you use
B) set your alpha level as high as possible
C) accept difference than to reject it
D) reject difference than to accept it

Question 10

A Chi-square test with an associated probability level cannot be interpreted unless you have the actual Chi-square value.

Question 11

If your probability level is lower than your alpha level, you should

A) accept the null hypothesis
B) reject the alternative hypothesis
C) propose another hypothesis
D) none of the above

Question 12

A probability level is just another name for the likelihood of random error.

Question 13

An alpha level is the name for dividing probability by two.

Question 14

If a zero association exists, which of the following hypotheses would be supported?

A) one-tailed
B) null
C) alternative
D) two-tailed

Question 15

Statistical significance means

A) the results are highly important
B) there is a reasonable level of certainty that results are not due to chance
C) there is a reasonable level of importance that can be attached to the results
D) you have chosen the correct statistic and it was powerful enough to provide accurate results

Question 16

An .05 level of significance means that

A) there is only a 5 percent chance that a statistic's value could be obtained as a result of random error
B) one is 50 percent certain that the sample value is representative of the population
C) there is only a 5 percent chance that the values tested are really different from each other
D) the difference found is not just significant, it is highly significant

Question 17

If an inferential statistic's calculated value is statistically significant, that means:

A) the result is very important
B) there is a strong association between two variables
C) the alpha level was met
D) all of the above

Question 18

When you have a one-tailed relationship, what would you normally do with the associated probability level?

A) divide it by 2
B) compare it directly to a pre-established number
C) determine if the probability level is either positive or negative
D) b and c above

Question 19

If the alpha level is .05, the confidence level is .95.

Question 20

While it is common to use p = .05 as the criteria for significance, you don't absolutely have to.

Question 21

For two-tailed hypotheses, you must remember to divide the calculated probability by 2.

Question 22

If an inferential test result is significant, you no longer have to worry about the direction of a hypothesis.