130501P - DATA ANALYSIS: ANALYSIS OF CLASS DATA

Presentation at a Training Program on Biostatistics for physician managers working in Public Health Administration, Qassim Province on May 1, 2013 by Professor Omar Hasan Kasule Sr MB ChB (MUK), MPH (Harvard), DrPH (Harvard) EM: omarkasule@yahoo.com

Data Collection

The following data shall be entered by participants into an excel file that shall be made available at a convenient place. Each participant will enter the data with no personal identifiers. The last column is an experiment, put half a teaspoonful below the tongue and close the mouth. Using a stop watch determine how long it takes for the sugar taste to ‘die away’. Stop the exercise at exactly 1.0 minute. Record the time and the status 1=taste disappeared 0=taste did not disappear.

Age	Gender	BMI	Eye glasses self	Age-father	Eye glasses-father	Age-mother	Eye glasses-mother	handedness	Pre-score	Post score	Time to end of sugar taste
1
2
3

Var 1 Age:                                                      1=<40, 2=40+

Var 2 Gender                                                            1=male, 2=female

Var 3 BMI                                                     1=<30 2=30+

Var 4 Eye glasses-self                                              1=yes, 2=no

Var 5 Age father                                           1=<70, 2=70+

Var 6 Eye glasses-father                             1=yes, 2=no

Var 7 Age mother                                        1=<70, 2=70+

Var 8 Eye glasses-mother                         1=yes, 2=no

Var 9 Handedness                                       1=right handed 2=left handed

Var 10 Pre-course biostat score                1=High 2=Low

Var 11 Post-course biostat score               1=High 2=Low

Var 12 Time end of sugar taste in seconds or minutes or time at termination

Var 13 Status at time Var 12 above                      1=taste ended 0=taste did not end

A. Assignments (descriptive)

1. Check the data for each case/observation for consistency eg a 40-year old with a 30-year old father is inconsistent
2. Compute the frequency with percentages of each of Var 1-Var 11
3. Compute the mean of Var 13

B. Assignments (inferential) on association for independent samples: Ho: prop eye classes in females – proportion of eye glasses in males =0

1. Use cross tabulations and the chisquare test statistic to test for association between Var 2 and Var 4
2. Using cross-tabulations and the chi square test statistic of association, determine which variables are likely to be confounders of the relation between Var 2 and Var 4
3. Use the Mantel-Haenszel procedure to adjust the relation between Var 2 and Var 4 for each of the potential confounders
4. Use the logistic regression procedure to adjust the relation between Var 2 and Var 4 for all potential confounders simultaneously

C. Assignments (inferential) on effect measures for independent samples

1. Using cross tabulations compute the odds ratio for the relation between Var 2 and Var 4
2. Adjust the odds ratio above for each of the potential confounding factors using the Mantel-Hanszel procedure
3. Adjust the odds ratio  above for all potential confounding variables simultaneously using the logistic regression procedure

D. Assignments (inferential) on association and effect measures for paired samples

1. Test for association between Var 10 and Var 11 using the McNemar chi square statistic of association
2. Compute the odds ratio for Var 10 and Var 11

E. Assignments (inferential) on survival analysis

1. Using the Kaplan-Meier procedure, compute a survival curve for disappearance of the sugar taste
2. Recompute the curve above by gender, age, BMI, and handedness