1 of 10

Data analysis Part II: οΏ½Characterizing study population οΏ½(Data summarization)

​

Tigist W. (MD, MPH)

December, 2022

2 of 10

Research

  • A systematic collection, analysis and interpretation of data to answer a certain question or to solve a problem.

In other words, research has the following flow

​

  • Observing Question/ problemπŸ‘ͺ designing a way to answer/solve the Question/ problem (proposal) πŸ‘ͺ To implement the proposal we need dataπŸ‘ͺ systematically collect the data πŸ‘ͺ Analyze the collected dataπŸ‘ͺ interpret the result of the analysis (thesis/ manuscript)πŸ‘ͺ Apply the result to the practice

2

3 of 10

Data analysis: Definition

  • It is a systematic process of applying statistical techniques and logic (scientific knowledge) to a well managed data to answer our research question.

OR

  • Making a meaning out of a raw data.

​

  • How do we make a meaning??? πŸ‘ͺ Data analysis levels/steps
    • Describing the data
      • Characterizing the study population based on all the studied variables (exposures and outcome/s).
      • All research types (Qualitative and Quantitative)
      • Descriptive analysis
    • Comparing groups

​

    • Building relationships

4 of 10

Data analysis tool selection criteria: οΏ½For all analysis levels

  • Selecting the appropriate data analysis tool for all the three levels
    • Always depends on the scale of measurement of the variable to be analyzed.

​

​

​

5 of 10

Summarizing a categorical variable

  • E.g., Disease outcome (Recovery/complication) of participants in 400 study population
    • 250 recovered and 150 complicated πŸ‘ͺ Simple count
    • The ratio of recovered to complicated is 1.7 πŸ‘ͺ Ratio
    • 62.5% of the patients recoveredπŸ‘ͺ Proportion
    • The prevalence/incidence of recovery was 62.5% during the 02 months follow up period πŸ‘ͺ Rate

​

N.B. A single summary measure will not be a representative result of the studied population unless it is accompanied by a range of values.

    • Confidence interval (90%, 95%, 99%)

​

  • Outcome variable πŸ‘ͺ Rate (prevalence or incidence rate)

E.g., The prevalence of recovery was 62.5% during the 02 months follow up period.

  • The prevalence of recovery was 62.5% (95% CI= 54.2% - 69.5%) during the 02 months follow up period.
  • Exposure variable πŸ‘ͺ Simple count (proportion)

E.g., Among the study participants 250 (62.5%) were females.

​

​

6 of 10

Summarizing a numerical variable

  • E.g., Length of hospital stay
    • Methods to find the central value πŸ‘ͺ Mean, median, mode
    • Methods to assess dispersion πŸ‘ͺ Standard deviation (SD), interquartile range (IQR), variance …

N.B. A single summary measure will not be a representative result of the studied population unless it is accompanied by a range of values.

​

  • Mean (SD)
    • When the variable has a normal distribution
  • Median (IQR)
    • When the variable has a skewed distribution

How to check distribution???

  • Using plots (subjective method)
  • Using tests of normality with p-values (objective method)
    • tests of Kolmogorov-Smirnov test or Shapiro-Wilk test

7 of 10

​

​

SPSS demonstration

8 of 10

Practice

​

https://forms.gle/Bt1LRyWWdFPAr9tr7

​

​

​

9 of 10

Become an independent researcher

@ Medical Research Lounge

​

Join our telegram group πŸ‘‰ https://t.me/Medical_Research_Lounge

​

Subscribe to our you tube channel πŸ‘‰ https://youtube.com/@medicalresearchlounge

​

Website πŸ‘‰ http://www.medicalresearchlounge.com/

​

Follow us on TikTok πŸ‘‰ https://vm.tiktok.com/ZMFDhH4Ta/

​

Follow us on LinkedIn πŸ‘‰ https://www.linkedin.com/company/medical-research-lounge?lipi=urn%3Ali%3Apage%3Ad_flagship3_company_admin%3BmwZnimq%2BQmG%2Bt8bUXcSjLw%3D%3D

​

9

10 of 10

​

End

​