3. Analysis of data
It is the final step of a research investigation
Reaching towards the end of the ladder of research process
Analysis, interpretation and generalization becomes a highly important piece of
research project
For the successful and meaningful analysis raw data must be properly compiled ,
edited , coded and presented with the use of texts, tables and figures
Accurate and objective interpretation conveys the readers a coherent and meaningful
understanding of the findings
4. Errors in Data
• Random error
• Systematic error- Faulty instruments, wrong technique, wrong study
design
• Bias- consistent difference between “recorded value” and the “true
value”
- Bias at data collection, analysis, interpretation, calculation and
publication
• Confounding errors- interpretation of a study result may be affected
by an external variable
• Attrition Bias
5. Avoiding Errors
• Lay down correct diagnostic criteria and observe them rigidly
• Encourage patients to report so that they are not “ lost to follow up”
• All subjects must be under similar conditions
• The questions to be asked to all subjects must be similar, nondirective
• Training of observers and blinding the hypothesis with them
• Study should be accurate and reproducible to get consistent results
6. Avoiding Errors
• Standardized equipment
• Inclusive and exclusive criteria
• Large sample size- Confounders get equally distributed in both the groups
• Restricted entry of subject at any point in study
• Match both experiment and control groups by many variables
• Randomize the study to receive either of the treatment
7. Definition
• Analysis referred as a method of organizing data in such a way that
research questions can be answered, and hypothesis can be tested
• Analysis is defined as the process of systematically applying statistical
and logical techniques to describe , summarize and compare data
8. Purposes of Data Analysis
• To reduce data to intelligible and interpretable form
• The relationship between the research variables can be studied, tested,
established or rejected
• E.g. Effect of acupressure on progress of labour
• To assess the significance of the difference between the means
• E.g.- comparing 2 groups
• To evaluate the degree of correlation between the variables or the
characteristics
9. Steps of Processing the Quantitative data for
Analysis
Compiling
the data
Editing the
data
Coding the
data
Selecting the
software
Entering the
data
Data
cleaning and
managing
the missing
value
Classification
of data
Tabulation of
data/
Presentation
of data
Analysis of
data
11. 2. Data Editing
Monitoring the data for quality and comprehensiveness
Check for accuracy and completeness
Lack of uniformity
Unanswered questions
Some information needs reconstruction into some different category
12. 3. Coding the Data
Coding is translating items and answers into numerical value or assigning numbers
/alphabets to the various categories of a variable that can be entered into the database and
for easy analysis
Symbols compatible to computer assisted analysis
Alternatives must be mutually exclusive( used in one concept only/ specific to one kind of
information)
Coding decision can be taken at the designing stage of the questionnaire
Eg- hypertensive/non hypertensive
Master coding sheet to be prepared
Coding manual with precise instructions
13. • Coding of all the forms should be done on daily basis in order to
manage the work
• Open ended questions- answers to be categorized and coding to
be done
• Missing data and ‘Don’t know’ answers should be coded
• Every subject to be coded
14.
15. B. Organization of Data
1. Selecting a software:
- Using computer for the analysis of data
- Software system-
- SAS, SPSS, EPI info, Microsoft excel, Stata, Minitab, Mathematica, Milwin ,
Acastat software,statplus2009,Medicalc etc
- -EPI Info- (CDC): http:/www.cdc.gov/epiinfo
16.
17.
18. 2. Entering the data:
- Direct entry from questionnaire
- Data into coding sheet and then into selected software
- 2 persons- to prevent errors
19. 3. Cleaning the data/ Managing the missing value
• Printing the sheet
• Checking for errors
• Random checking for wrong entries
• Counterchecking of small data with original sheet for precision
• Missing data- it is important not to put zero to the category, this will lower the mean value
• If a particular item answers are consistently missing- exclude the question /item form all
subjects
• Extensive data is missing, subject should be completely excluded from the study
20.
21. 4. Classification of data
• The collected raw data is categorized into common groups having
common features
• Helps in making comparison among the categories of observations
• Numerical characteristics can be grouped into class intervals
22. 5. Tabulation of data/ Presentation of data
• Tabulation is an orderly arrangement of data in rows and columns
• It conserves space
• Facilitates process of comparison and summarization
• Basis for various statistical computations
23. 6. Analysis of Data
• Descriptive statistics- Frequencies of descriptive variables
• Inferential statistics: To test the hypothesis/ research questions/objectives
24. Steps of data analysis
1. Deciding the
purpose of
data analysis
2. Reorganising
the data
3. Reformulating
the hypothesis
In terms of
statistical
hypothesis
4.Setting
the level
of
significa
nce
5.Choosing
an
appropriate
statistical
test
6.Performing
the
statistical
test
7.Evaluating the
test
8.Interpreting the
research
hypothesis