Hi, I was reading the paper (Article MicroRNA Expression, Survival, and Response to Interferon in...
). In Table 1, I'm wondering how did the authors calculate a p-value of 0.29 for Age?The age in Table 1 is about age median and range (see below). Any suggestion is welcome. Thank you!
Age- year Cohort1 Cohort2
Median 50 50
Range 13-83 20-77
Differences between two medians can be evaluated by Fisher's exact test.
(the supplementary appendix to this paper indicates it was done by an "unpaired test' presumably a t test on means, but this may be an error).
Well, I think that p-value for age was calculated with unpaired t-test, where high p-value means that there would be no age difference between Cohort 1 and 2.
Mean and range are just one of the descriptive statistical indicators for Cohorts' populations, and the authors fitted age by Cohorts' samples.
Sang-Min Paik James Leigh Thank you! If we only compare the median age in the two cohorts, then it is not a 2x2 contingency table. How could I perform chi-square or Fisher exact test? I agree with Sang-Min Paik that the authors might use the ages of all the patient in the two cohorts to perform the two sample unpaired t-test. Am I right?
Authors used unpair t-test (with or without correction for equality of variances) if data in both samples were normally distributed, else if data in both samples were not normally distributed authors used non-parametric analogue of the unpair t-test - Mann-Whitney U-test.. Interval descriptive statistics are rather unusual (usually mean and standard deviation are used for normal distribution of data, median and quartiles for abnormal).
PS The use of the minimum and maximum as descriptive statistics is rather unusual (usually, the mean and standard deviation are used for the normal distribution of the data, and the median and quartiles for the abnormal).
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Nucleic Acids
- RNA
- Small RNA
- microRNA