statistics problem

[als20]

Hello

Could you please help me with this problem on statistics? Thanks so much

It has been suggested that showing the face of the speaker in a computer package that teaches a foreign language leads to improved memory of words and phrases. To investigate this, a study was undertaken whereby 24 post-beginners to the Polish language were divided randomly into one of two groups. Members from the two groups were each asked to sit in front of a computer which then presented them with 20 words and phrases and gave their English translation together with a sentence that put them into context. The words and phrases were each presented twice.

The participants who were in Group 1 saw the words on screen and heard them being read out, but no face was shown on screen.

The participants who were in Group 1 saw the words on screen and heard them being read out by a speaker whose face was shown.

Both groups took a 15 minute break after their exercise and were then given a piece of paper with the words and phrases typed and participants were given 5 minutes to write down the English translation. The experimenters were interested in comparing these two methods of learning foreign language words.

Number of correctly translated words for the members in each group:

Group 1 (no face) Group 2(face)
12 16
8 10
11 13....
3
.
.
.

(a) Use SPSS to perform an appropriate t-test for this experiment.

(b) Use SPSS to perform and appropriate non-parametric test for this experiment.

(c) You are asked by the experimenters to advise them on which test would be the most suitable to be applied to these data Discuss how you would advise them.

What are the appropriate tests for q(a), (b), (c)? I have some thoughts but I am not sure I am right...

Thank youu!!

[als20]

Could you please explain to me how we can assess normality using a dotplot...

[terzi]

Studying normality in a dot plot is similar to what you would do with a histogram, look for a bell shaped dataset, but the idea of the dot plot is to look for possible outliers that may influence in the distribution. Now, there would be more appropriate to use a Normal Probability Plot to assess for normality.

As you may have guessed, if your data doesn't follow the assumptions required for a t-test, then you should use the non-parametric alternative.