1.

A researcher studies the effects of oral intake of vitamin D3 on the vitamin D serum blood level. 80 subjects from Groningen and Friesland were randomly selected for the study. 40 subjects were daily administered 2000IU over the course of 2 months. The other 40 were given a placebo. Serum levels were measured after the dosages were administered. The researcher is particularly interested in differences between sexes and races. The following four plots are from the SPSS output of an ANOVA analysis. What can you tell about the vitamin D3 requirements among the different groups? Describe differences between the groups? Bear in mind that scientists recommend that the serum level should be in 30-50ng/ml range.

Figuur 1: Y axis: serum level in ng/ml

2.

A researcher is investigating the effects of L2 speakers’ pronunciation on the L1 speakers’ comprehension. She pays a special attention to the effects of incorrect stresses and phonemic realizations. She records a number of English sentences pronounced with stresses on the incorrect syllables, and with incorrect phonemic realizations. For example:

  1. A sentence with stresses on initial syllables of words. (Check and Serbian speakers of English tend to make these types of errors in their English, as they tend to transfer the rules from L1)
  2. A sentence with stresses on final syllables of words (e.g. French speakers of English tend to make such errors)
  3. A sentence with incorrectphonemic realizations. (e.g. Dutch speakers tend to pronounce the English vowel /a/ as in ‘trap’ as /e/ as in ‘bed’. Dental fricatives were pronounced as alveolar stops: /t/ and /d/)

Comprehension of the pre-recorded sentences was tested in 45 native monolingual speakers of English. Each sentence contained exactly 25 syllables. There were six different types of sentences. For example, sentences with:

  1. Incorrect word initial stress
  2. Incorrect word final stress
  3. Incorrect word initial stress +10 incorrectly realized phonemes.
  4. Incorrect word final stress + 10 incorrectly realized phonemes.
  5. Proper English stress with 10 incorrectly realized phonemes.
  6. Proper English stresses with proper phonemic realizations.

All sentences were played one-by-one to the L1 listeners. Listeners were asked to “push the button” once they understood each of the sentences. For the dependent variable the researcher measured the length of time from the point when a sentence was completely spelled out until the point when a subjects “pushed the button”.

Which statistic(s) is(are) the most appropriate for this study?

3. Explain or give the formula of how SST, SSG and SSE are calculated.

4. What are the advantages of repeated measures ANOVA over One-Way or Factorial ANOVA? Illustrate your answer with an example.

5. The following two plots are from SPSS outputs of two different Factorial ANOVA tests. What can you tell about the differences between the independent variable interactions in the two?

6. Name and briefly elaborate on the basic assumptions of ANOVA.

7.What is the H0of Factorial ANOVA? Make sure that your answer covers both dependent and independent variables.

8. What do Q-Q plots illustrate? What does the following plot indicate?

9. A researcher is investigating the pronunciation of /l/ and /r/ phonemes in 60 Chinese L2 speakers of English. She is interested to see whether exposure to spoken English affects pronunciation. Half of the subjects will watch BBC news 4 times a week for a full hour. The experiment will last a full year and the subjects’ progress will be assessed six times a year. The questions that the researcher is interested in are:

- Is there a significant difference between the two groups?

- Is there a significant difference between different age groups and sexes?

How would you operationalise pronunciation of /l/ and /r/ into a dependent variable? Based on the interests of the researcher, what would be the most appropriate statistic to test the results?

10. The following table comes from an SPSS output of a Pos Hoc analysis. What can you tell about the differences between the different levels of the independent (age) variable?