I Hate PSM

Epidemiology: Tests of Significance

BiostatisticsEpidemiology

Commonly used tests of significance are

For qualitative data: (comparing two proportions)

  • Standard error of difference between two proportions
  • Chi – square test

For quantitative data: (comparing two means)

  • Unpaired – t test (the t – test)
  • ‘Z’ – test
  • Paired t – test

Chi – square test

  • Applied on qualitative data
  • Equally applicable on small and large samples
  • Can be applied if there are more than 2 categories

Std. error of diff between proportions test

  • Applied on qualitative data
  • Applicable on larger samples only (≥30 participants)
  • Can compare the proportions between two groups only

Unpaired or student’s t – test

  • Calculate the observed difference between the two means
  • Calculate the standard error of difference between the means of the two groups
  • Calculate the t – value as:

t= (observed difference between the two means)/(standard error of difference between the means)

  • Determine the degrees of freedom as:

D.f. = n1 + n2 -2
Where n1= number of subjects in the first group
And n2 = number of subjects in the second group
Refer to t – distribution table

  • Look against the calculated d.f. and observe the t value under p=0.05
  • If the calculated value exceeds this value of t, the p is

The Z – test

  • Applicable only in larger samples
  • Z is calculated instead of t

And:

  • If Z=1.56, p value is 0.10
    • If Z=1.96, p value is 0.05
    • If Z=2.3, p value is 0.02
    • If Z=2.58, p value is 0.01

Higher the value of Z, lower is the p – value

Paired t – test:

Used when:

  • Single sample but each subject gives two values (paired values) of the same variable
  • Quantitative data
  • Random sample

Examples of where paired t – test is applicable:

  • Value before and after treatment e.g. systolic BP of each subject before taking treatment and the systolic BP of the same subject after treatment
    • Comparing effects of two different drugs on the same individual for example blood sugar level after 2 hours of drug A and the same after drug B
    • Checking the accuracy of a new measuring instrument against a standard one, for example Hb estimation by an new method against the same by colorimeter

For more details on the teststests of significance
References:
Tiwari P. Epidemiology Made Easy. New Delhi: Jaypee Brothers; 2003
Gordis, L. (2014). Epidemiology (Fifth edition.). Philadelphia, PA: Elsevier Saunders.
Bonita, R., Beaglehole, R., Kjellström, T., & World Health Organization. (2006). Basic epidemiology. Geneva: World Health Organization.
Schneider, Dona, Lilienfeld, David E (Eds.), 4th ed. Lilienfeld’s Foundations of Epidemiology. New York: Oxford University Press; 2015
K. Park. Principles of Epidemiology and Epidemiologic Methods. In Park’s Textbook of Preventive and Social Medicine. 24th Ed. Jabalpur: Banarasidas Bhanot, 2017: pg 58 – 145

Steps in Investigation of an Epidemic: Epidemiology: Steps in Investigation of an Epidemic
Tests of Significance: Epidemiology: Tests of Significance
Monitoring and Evaluation: Epidemiology: Monitoring and Evaluation
Advantages and Disadvantages of Case Control Studies: Epidemiology: Advantages and Disadvantages of Case Control Studies
Advantages and disadvantages of cohort study: Epidemiology: Advantages and disadvantages of cohort study
Basic Concepts in Epidemiology: Epidemiology: Basic Concepts
Types of Epidemiological Studies: Epidemiology: Types of Epidemiological Studies
Differences between Case – control and cohort study: Epidemiology: Differences between Case – control and cohort study
Uses of epidemiology: Epidemiology: Uses of epidemiology
Blinding in Experimental Studies: Epidemiology: Blinding in Experimental Studies
Evaluation of a Screening Test: Epidemiology: Evaluation of a Screening Test