Assignment II for Biostats Course VHM 801 at AVC - Fall semester 2016

The assignment is worth 15% of the final course mark.

The assignment uses information from a study of dependence on caffeine and symptoms associated with its withdrawal. Study participants were initially recruited through newspaper ads seeking volunteers who regarded themselves as psychologically or physically dependent on caffeine, and met a number of demographical and medical criteria. A total of 99 applicants were further screened by telephone interviews, and the eligible subjects were invited to the first part of the study. This first part comprised 26 participants and involved a physical examination, a screening blood test and a standardized psychiatric interview targeted towards substance dependence. Of those 26 persons, 15 were diagnosed as being caffeine dependent. Eleven of these proceeded to the second part of the study on caffeine withdrawal.

Each of the 11 participants underwent two 2-day treatment periods, one week apart, with a restricted diet wherein the participant's intake of caffeine would be limited to capsules provided. These capsules had different content in the two periods; specifically, in one period they contained an amount of caffeine equivalent to each individual's average daily caffeine consumption, and in the other period they contained no caffeine. For each participant, the order of the two treatment periods was determined randomly. In the instructions provided to participants, the contents of the non-placebo capsules was described as including several compounds, including caffeine, commonly found in food and beverages. At the end of the second day of each treatment period, the participant was asked to fill out several standardized questionnaires and complete some tasks.

For the home assignment, you are required to answer Questions 1-8 below. In general, the assumptions of every statistical procedure used should be stated (formally or informally) and checked (where possible), and every statistical analysis should be summarized in a conclusion.

A dataset for the second part of the study is available in Minitab format and as a comma-separated file, for import into Stata and other statistical software. In addition to subject id and some demographic information, the data include scores for fatigue, depression, vigour and mechanical motor skills for both treatment periods; C=caffeine period, NC=no-caffeine, or caffeine-free, period. Further detail about some of the variables is provided in the questions below.

1. As described above, 15 out of 26 participants in the first part of the study were classified as suffering from clinical substance (i.e. caffeine) dependence. Use this information to estimate, with a suitable interval, a population proportion of people with such substance dependence. Characterize the population this inference could be considered as representative for, as well as you can from the information provided.

2. The second part of the study was described as a "double-blind study design". Explain what this would mean in the specific study, and suggest reason(s) why such a design was used.

3. Motivate why the order in which the two types of capsules were administered to participants in the second part of the study was random, and indicate briefly how you could carry out the randomization (without actually carrying it out).

4. Discuss the advantage(s) of having the two study periods one week apart instead of using two consecutive two-day periods? Additionally, why might it be preferable to have a full week between the start of the two periods rather than, for example, 5 days or 10 days?

5. A vigour withdrawal symptom was defined to have occurred if the quetionnaire's vigour score for the caffeine-free period was at least two standard deviations below the norm for college students (obtained from a large sample of college students). What is the (approximate) probability that the score of a randomly selected person (from the population) will be at least two standard deviations below the mean? List the assumptions your calculation is based on.

6. The study reported 5 out of the 11 subjects having had vigour withdrawal symptoms. Compute the probability of observing 5 or more out of 11 persons sampled from the general population with vigour scores at least 2 standard deviations below the norm. (Hint: Use your value from Question 5 for the population probability.) What does the computed probability tell you about the performance of the study participants in the caffeine-free period?

7. Carry out a statistical analysis to compare motor skill scores (MotC and MotNC) of the partipants in the two periods. These scores were averages of three trials done in immediate continuation of each other, wherein the participants were instructed to press a button 200 times as fast as they could; the scores are in beats per minute. Include a statistical test for the null hypothesis that there is no difference between mean skill scores during the caffeine and caffeine-free periods against any difference between mean skill scores in the two periods.

8. Carry out a statistical analysis to compare the partipants' scores in the two periods for either fatigue (FatC and FatNC), vigour (VigC and VigNC) or depression (DepC and DepNC), as of your own choice. These scores were obtained from the questionnaires filled out by the participants after the study periods. The expectation prior to the study from the available literature on withdrawal symptoms was that fatigue and depression scores might be higher, and vigour scores lower, in the caffeine-free period. Include a statistical test for the null hypothesis that there is no difference between mean scores during the caffeine and caffeine-free periods against a suitable alternative.