Lectures 6-7 - Comparing Means

Question 0

The t-test is the most extensively used statistical test. What are the three main versions?

Answer 0

• independent samples
• paired samples
• one-sample

Question 1

Define statistically significant

Answer 1

It means that the result is unlikely to occur by chance and is reliable

Question 2

What does the t-statistic depend upon?

Answer 2

• the size of the difference between the conditions

- the s.e.m. of that difference, which in turn depends on the variability of the data and number of measures)

Question 3

What do you need to calculate p when doing a t-test?

Answer 3

• the t-statistic

- the degrees of freedom

Question 4

What is the calculated p value compared to?

Answer 4

Alpha, usually set at 0.05, which is the level p must be below in order to reject the null hypothesis.

Question 5

Degrees of freedom is usually closely related to… but depends on…

Answer 5

N, the number of participants

The statistical test used

Question 6

Degrees of freedom change…

Answer 6

The probability distribution on test statistics - p is lower for many degrees of freedom than for few degrees of freedom

Question 7

When is an independent samples t-test used?

Answer 7

With between-group designs

Question 8

What are the other names for an independent samples t-test?

Answer 8

Unpaired means t-test
Unrelated samples t-test
Between-groups t-test

Question 9

How is t calculated in an independent samples t-test?

Answer 9

The mean difference is divided by the standard error of the difference.

Question 10

What assumptions does an independent-samples t-test make?

Answer 10

• data must be on a ratio or interval scale
• data is normally distributed
• data in each group is independent of the other
• data in the two groups must have similar variance

Question 11

How do you do an independent samples t-test in SPSS?

Answer 11

Analyse/compare means/independent-samples

Remember to check Levene’s Test for Equality of Variance, it should be >0.05, in which case you use the top row of data

Question 12

If Levene’s Test for Equality of Variances is >0.05, what do you do?

Answer 12

Use the lower row of data

Question 13

How do you report the results of a statistical test?

Answer 13

A … test showed that the data were(n’t) significantly different (t=… , df= … , p</>0.05).

Question 14

In what situation might you report p as more or less than a different value to 0.05?

Answer 14

Wen particularly confident that there is a statistically significant difference, you might quote p as <0.01

Question 15

The actual value of p is sometimes quoted, especially if…

Answer 15

P is near alpha

Question 16

When is a paired-means t-test used?

Answer 16

For within-subject designs

Question 17

A within-subjects t-test is also known as…

Answer 17

• related-sample t-test

- within-subject t-test

Question 18

Why is a paired-means t-test more powerful than an independent samples t-test?

Answer 18

Because it uses subjects as their own control, eliminating individual differences (a source of noise which can ‘crowd out’ what you’re measuring)

Question 19

What does power mean?

Answer 19

A more powerful test is more likely to find a difference/effect

Question 20

How is the t value calculated for a paired-means t-test?

Answer 20

The mean of the differences for the pairs is divided by the standard error of the differences

Question 21

In the paired means t-test output window, what does the paired samples correlations indicate?

Answer 21

Whether participants acted consistently across conditions