METHODICAL INSTRUCTION

FOR STUDENTS OF THE 4 COURSE

LESSON № 2 (PRACTICAL – 6 HOURS)

Theme 1. Dependent and independent samples. Testing significance difference of two samples and differences of their central tendencies.

Theme 2. Testing the significance of the difference of a few samples oh and differences of their central tendencies.

Theme 3. Analysis of variance. Kraskala-Wallis analysis. The median criterion.

Aim. Learn the main principles of the Biostatistical methods. Also the general Testing the significance of the difference of a few samples oh and differences of their central tendencies must be learned too. Shown using of the Analysis of variance for solving medical and pharmaceutical task.

Professional motivation.

It is important first to understand that STATISTICA is a modular package: different parts of the package are loaded for different statistical procedures. For example, if one wants to compute descriptive statistics first, and then conduct a 2 Way ANOVA on the data, this will entail loading the module ‘Basic Statistics/Tables’ for the first step, and the module ‘ANOVA/MANOVA’ for the second step.  Some aspects are common across all modules, though, such as Graphs and the STATISTICA BASIC language.

There are 4 t-test modules offered by STATISTICA:

·         t-test, independent by groups

·         t-test, independent by variables

·         t-test, dependent samples

·         t-test, single sample

These are all contained in the Basic Statistics/Tables module.

Background

1.      Skills of working with the operation system Windows.

2.      Skills of working with tables in the Statistica.

3.      Skills of working with WWW service via Web-browser.

4.      Skill of working with presentations in the MS PowerPoint.

I. Practical work 9⁰⁰ – 12⁰⁰ (4 hours)

·         Illustrative materialss:

Metodical instructions that allocated on the university web-site. Computers that is runned under the MS Windows operational system with installed STATISTICA application.

·         Methods of practical work:

Practical exercise 1. Statistical hypothesis testing.

Problems. The scientists conducting the medical experiment checkup for 15 practically healthy men and women by age 17-18 years. For the each of them defined the concentration of haemoglobin and amount of red corpuscles (erythrocytes). Results was stored in table (table 1). It is necessary to check up the next hypothesis:

Are there statistically meaningful differences in the parameters of the blood for the people of different sex?

Table 1. Results of the blood paramenters checkup

Steeps for the work performing:

1. Formulate the task by the mathematical statistics point of view (define which one statistical hypothesis it is need checking). Your answer must be grounded.

Solve. Reed lecture notes to do this.

2. Define the type of the variables distribution.

Solve. The Coefficient of skewness and coefficient of kurtosis need compute for each sample in your task for this. Then, compare calculated results with special tables of the criticals value and get answer about the distribution type.

3. Choose method of the statistical hypothesis check in according of the variables distribution type. Your answer must be grounded.

Solve. Reed lecture notes to do this.

3.1. Test homogeneity of the Variance in selected samples.

3.2. Comparing Sample Means.

Solve. Select Statistics Basics Statistics from the main menu bar. Seclect apropritate module to perfotm testing:

·         t-test, independent by groups

·         t-test, independent by variables

The t-Test dialog box will be open. In the Variable 1 Range and Variable 2 Range fields, type (or select from the spreadsheet) the cell ranges containing the dataset corresponding to mens and womens blood parameters, respectively. Remember, that you must perform different test for comparing hemoglobin and erythrocytes means! Left the Alpha field (confidence field) at the default 5%. Finally, choose your output options. You can use any empty cell bellow your table data on the same spreadsheet. Click OK to compute. The results include computed means and variances for the two input datasets, along with other information such as the number of observations, the hypothesized mean difference, and the degrees of freedom. The t-statistic is also reported, along with both the one-tailed and two-tailed probabilities and critical values. Use this data, get answer about rejection or acception hypothesis.

4. Make a scientifically grounded conclusion on task.

5. Save STATISTICA document in the folder D:\Users\ your_LastName, file name: hypothesis_1_test

Practical exercise 2. Statistical hypothesis testing

A pharmaceutical firm owns by the seven drugstores in the city. Advertising was placed in a public transport for two months to promote sales volume of some drug. In table 1 present sales volume for this drug before the advertising presents and when it present. It is necessary make the decision – continue a publicity advertising campaign, or halt it.

Table 1. Volumes of the drug sales before and during advertising campaign.

Steeps for the work performing:

Formulate a task by the mathematical statistics point of view (define which one statistical hypothesis it is need checking).

Define the type of the variables distribution.

Choose method of the statistical hypothesis check in according of the variables distribution type.

Check up the additional conditions for the statistical hypothesis check method (if need).

Calculate appropritate statistical test and check result by comparing with the critical value.

Make a scientifically grounded conclusion on task.

Save STATISTICA document in the folder D:\Users\ your_LastName, file name: hypotese_test

Practical exercise 3. Dot Blot Analysis

Students should be able:

·         Use STATISTICA to calculate descriptive statistics

Use STATISTICA to performing statistical hypothesis testing.

·         Use STATISTICA to conduct Analysis of Variance

II. Seminar discuss 12³⁰ – 13⁰⁰ (2 hours)

Special attention should be paid to the following questions:

Types of mathematical models of pharmaceutical technology.

Deterministic and stochastic models in the pharmacy.

Organization and conduct research to create drugs.

Types of experimental design in pharmaceutical technology.

Using univariance analysis of variance in research to develop drugs.

Using two-factor analysis of variance.

Using multivariate analysis of variance.

III. Individual work 14¹⁰ – 15⁰⁰ (1 hours)

Test task examples:

1.      Many people describe hypothesis testing as COUNTERINTUITIVE because

A.     we test whether something happened in order to conclude that nothing happened.

B.     we test whether nothing happened in order to conclude that something happened.

C.     we can only conclude that nothing happened when we are 100% sure that something did not happen.

D.     we can only conclude that nothing happened when we are 99% sure that something did not happen.

E.we test whether something happened but can still conclude that nothing happened.

2.      If “going to the doctor” is used as an analogy, then STATISTICAL POWER is

A.     your doctor missing a real illness.

B.     your doctor stating you are not sick when there is nothing wrong.

C.     your doctor is absent.

D.     your doctor confirming that you are really sick.

E.getting scared for nothing.

3.      We would FAIL to reject the null hypothesis when the test statistic value is smaller than.

A.     standard error.

B.     degrees of freedom.

C.     critical test score

D.     mean.

E.variance.

4.      Whereas the null hypothesis is symbolized as Ho, the alternative hypothesis is symbolized as Ha or __ .

A.     H1

B.     HA.H.

C.     Hz

D.     H0

E.A.H.

5.      A Type II error occurs when

A.     Right answer not listed there

B.     we incorrectly fail to reject a false null hypothesis.

C.     we incorrectly reject a true null hypothesis.

D.     we correctly reject a false null hypothesis.

E.we correctly fail to reject a false null hypothesis.

 

References

General

1.      Practical classes matherials

2.      Schwartz, Russell. Biological modeling and simulation : a survey of practical models, algorithms, and numerical methods / The MIT Press, – 2008, 403p.

3.      Biostatistics for oral healthcare / Jay S. Kim, Ronald J. Dailey. – 1st ed., Blackwell Publishing Company, – 2008, 344p.

4.      Information Technologies in Biomedicine / Ewa Pietka, Jacek Kawa (Eds.), Springer-Verlag Berlin Heidelberg, – 2008, 569p.

5.      Medical statistics: a guide to data analysis and critical appraisal / by Jennifer Peat and Belinda Barton. – 1st ed., Blackwell Publishing Ltd, – 2008, 338p.

6.      Medical statistics from scratch : an introduction for health professionals / David Bowers.—2nd ed., JohnWiley & Sons Ltd,, – 2008, 302p.

7.      Bioelectrical signal processing in cardiac and neurological applications / Leif sornmo, Pablo laguna, Elsevier Academic Press, – 2005, 685p.

8.      Probabilistic modeling in bioinformatics and medical informatics / Dirk Husmeier, Richard Dybowski, and Stephen Roberts (eds.)., Springer-Verlag London Limited, – 2005, 510p.

9.      Biomedical Informatics Computer Applications in Health Care and Biomedicine / Edward H. Shortliffe (Editor), James J. Cimino (Associate Editor), Springer Science+Business Media, LLC, – 2006, 1060p.

10.  Information technology solutions for healthcare. – (Health informatics) / Krzysztof Zielinski, Mariusz Duplaga, David Ingram, Eds., Springer-Verlag London Limited, – 2006, 368p.

11.  PACS and imaging informatics : basic principles and applications / H. K. Huang., John Wiley & Sons, Inc., – 2004, 691p.

12.  MEDICAL INFORMATICS: Knowledge Management and Data Mining in Biomedicine / Hsinchun Chen, Sherrilynne S . Fuller, Carol Friedman, William Hersh, Eds., Springer Science+Business Media, Inc., – 2005, 655p.

Additional

1.      Functional informatics in drug discovery / edited by Sergey Ilyin., Taylor & Francis Group, LLC, – 2008, 160p.

2.      Bioinformatics Volume II Structure, Function and Applications / Edited by Jonathan M. Keith, PhD., Humana Press, a part of Springer Science+Business Media, LLC, – 2008, 497p.

3.      Web Mobile-based Applications for Healthcare Management / Latif Al-Hakim, Idea Group Inc., – 2007, 450p.

4.      Evaluating the Organizational Impact of Healthcare Information Systems / James G. Anderson, Carolin E. Aydin, eds., Springer Science+Business Media, LLC, – 2006, 225p.

5.      Informatics for the clinical laboratory : a practical guide / editor, Daniel F. Cowan, Springer-Verlag New York, Inc., – 2003, 331p.

6.      Handbook of Medical Imaging: Processing and analysis / Isaac N. Bankman, PhD, Editor, Academic Press, – 2000, 910p.

7.      BIOINFORMATICS: A Practical Guide to the Analysis of Genes and Proteins / Andreas D. Baxevanis, B. F. Francis Ouellette, Eds., John Wiley & Sons, Inc., – 2001, 489p.

 

Authors:                    as. A.V.Semenets

 

Authorized on the department meeting

“14” June 2013. Report № 12

Reconsidered on the department meeting

“____” _________________ 200__. Report № ___