It looks near normal. You can also look at the Q-Q plot (where you want data to resemble a 45 degree line) and there are statistical tests for examining data normality.
Hello,
I am a newbie in Biostatistics field... I would be so thankful if you tell me if below chart has a normal distribution or not and why ?
Chart title : The distribution of the weight of our donor population
Thank you very much
It looks near normal. You can also look at the Q-Q plot (where you want data to resemble a 45 degree line) and there are statistical tests for examining data normality.
Hello!
You can use Shapiro-Wilk test for normality (H0: distribution is normal). I think that is more useful then visual analysis of graph.
Well, as Aregel mentioned there are tests for normality. However these tests can have limitations, and have the ability to reject the null hypothesis of normality with small departures from the normal distribution when sample sizes increase. I personally feel there is a bigger push to use graphs to get past those types of limitations for tests and to better tackle the end use of data. If the assumption of normality in the proposed analytics is not overly stringent, say as in the t-test with larger samples sizes, then interpreting histograms and Q-Q plots may be good approach. Though other options such as tests exist. Though we must remember either approach could be construed as arbitrary at times (graphs visually interpreted or selected levels of significance used in tests).
Last edited by hlsmith; 12-10-2015 at 11:34 AM. Reason: typo
Stop cowardice, ban guns!
Tweet |