Statistics and Experimental Design, Lab 9, 16 March 2000

NOTE: EXAM PARTICIPANTS: IF YOU HAVE NOT RECEIVED YOUR TAKE-HOME ASSESSMENT, ASK ASTRID FOR A COPY.

Tutors: Max Louwerse, Anna Babarczy, Astrid Schepman, Technical Support: Mike Bennett

Purpose: Test or extend your SPSS skills: Practice for the take-home exam.

Reading/Reference (for all SPSS-related matters) Kinnaer, P.R. and Gray, C.D. (1997). SPSS for Windows made simple (second edition). Hove: Psychology Press

------------------------------------------------------------------------

A. Getting started:

1. If you have a login name that does not start in 'stats', then log in to pisa for surnames A-M, or log in to babel for surnames N-Z.

If you have a special stats login, then please log in as follows:

noah stats2 alton

yoko s stats7 alton

isabel stats8 alton

laura stats10 pisa

daniel stats11 pisa

elina stats12 pisa

jennifer stats17 babel

kate stats18 babel

rob stats19 babel

We are doing this because there were problems with these logins last week, which may have been partially due to an overload on one of the UNIX segments.

If you have forgotten your username + password, please see Mike.

2. Before you start up spss, you might want to check whether there is a file called core in your directory. Such a file is created if you don't exit an application in the prescribed way.

Type: ls -l

If a file called 'core' shows up, and you're sure you don't have a file called core for a valid reason, then you might want to delete it, as it takes up a lot of space. To do this:

Type: rm core

At the prompt, say y for yes.

3. Then:

xrsh machinename: where machinename is the machine to which you logged in originally (e.g babel, pisa, alton).

4. Then type: spss &

After a while, SPSS should start up. If you get an error message, give us a shout. If all goes well, you'll see a new window called 'Newdata'.

 

------------------------------------------------------------------------

PLEASE DO NOT FEEL SHY ABOUT ASKING FOR HELP.

THIS IS YOUR LAST CHANCE BEFORE THE EXAM!!!

IF YOU NEED TO,USE THE DECISION CHART (INSIDE FRONT COVER) FROM

GREEN AND D'OLIVEIRA TO DECIDE WHICH STATISTICAL TEST TO USE.

------------------------------------------------------------------------

B. A researcher has gathered some rating data, in which subjects rate the grammatical correctness of two types of sentence: 1. with split infinitives (e.g. To boldly go...) 2. With unsplit infinitives (To go boldly...). The researcher predicts that type 2 will rated more highly than type 1. She has asked 20 subjects to rate 10 sentences on a scale of 1 - 5 (1 = unacceptable, 5 = acceptable). Subjects performed in only one condition, and were allocated to conditions at random.

Below are the means for each subject, separated by condition, with subjects as the random factor:

Split

3

2

2

3

3

4

1

2

4

2

Unsplit

4

3

4

3

2

4

5

2

3

5

 

Analyse these data, using the appropriate statistical test, and evaluate whether the experimental hypothesis is supported by the data.

------------------------------------------------------------------------

C. A researcher has run an eye-tracking experiment, examining the effect of infinitive type (split - unsplit) on reading times. He predicts that unsplit infinitives will be read faster than split infinitives. The design is within items and within subjects, using a Latin Square. Ten subjects and ten items were used. Below are reading times per character, averaged over three test words, namely: "To (adverb) (verb)" or "To (verb) (adverb), averaged over all subjects for each item, with item as the random factor:

Item...

01

02

03

04

05

06

07

08

09

10

Split..

33

30

33

32

34

29

30

31

33

31

Unsplit

31

29

28

30

31

27

26

27

29

30

 

Analyse these data, using the appropriate statistical test, and evaluate whether the experimental hypothesis is supported by the data.

------------------------------------------------------------------------

D. A student runs an experiment asking head-injured patients which animal is doing something to which other animal in passive sentences such as 'The cow was kicked by the horse'. The patients either get this wrong or correct. The researcher has tested patients with one of two types of brain damage: type A and type B, and hypothesises that there will be a difference in patients' ability to classify the passive sentences correctly as a function of type of brain damage. Data are as follows:

 

TypeA

TypeB

Correct

28

5

Wrong..

35

15

 

Analyse these data, using the appropriate statistical test, and evaluate whether the experimental hypothesis is supported by the data.

Note of realism: It is unlikely that you'll ever have this many clinical patients with two distinct categories of brain damage to test.

------------------------------------------------------------------------

E. A researcher hypothesises that words show an association between being frequently used and containing frequently used letter trigrams (sets of three consecutive letters). Below are data for 15 words, showing their frequency, and the mean frequency of the trigrams contained in them:

Word.....

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Frequency

5

2

10

23

7

40

16

9

9

14

12

30

3

9

10

TrigramFr

6

3

9

16

10

21

7

4

6

4

3

12

3

5

02

Analyse these data, using the appropriate statistical test, and evaluate whether the experimental hypothesis is supported by the data.

------------------------------------------------------------------------

F. A researcher tests the effect of age and syntactic structure on reading times. 20 subjects, 10 aged 8 and 10 aged 10 read both active and passive sentences, and their eye-movements are monitored. The researcher predicts that young subjects will be slower overall, but particularly on the passive sentences. The data, expressed as reading times per character are as follows:

Subject......

01

02

03

04

05

06

07

08

09

10

Age8,active.

45

50

49

48

48

47

51

39

45

43

Age8,passive

55

59

49

53

47

53

57

49

54

55

Subject......

11

12

13

14

15

16

17

18

19

20

Age10,active.

40

38

37

39

40

42

31

37

45

35

Age10,passive

41

37

34

42

46

40

36

40

41

39

Analyse these data, using the appropriate statistical test, and evaluate whether the experimental hypothesis is supported by the data.

------------------------------------------------------------------------

G. Bye

That's it for today, and that's all the teaching for this term.

Good luck on the exam, and have a lovely Easter break.