Talk:Documentation/Calc Functions/PERCENTILE
SF comments
- (1) Summary. I think that this paragraph is insufficient – many readers might not understand the concept of an alpha-percentile. Suggest adding something like the following – “Percentiles can be used to understand the significance of particular values within a data set. The nth percentile of a data set is the value below which n% of the data set lies. This function uses an alpha argument lying in the range 0 to 1 (inclusive), which corresponds to a percentage in the range 0% to 100%. For example, an alpha of 0.5 gives the 50th percentile, also known as the median.”.
- (2) Returns. In the light of the proposed improvement to the Summary section, suggest simplifying this statement to read – “Returns a real number which is the scale value for the specified data set and alpha.”.
- (3) Arguments, 2nd bullet. No mention of what "n" is. Would it be worth adding an example of this in action? Perhaps assume a data set with five values, then alpha values of 0, 0.25, 0.5, 0.75, and 1 will pick out the values in the data set in ascending order?
- (4) Additional details, 1st paragraph. Maybe re-word slightly as – “alpha = 0.25 gives the first quartile, alpha = 0.5 gives the median, and alpha = 0.75 gives the third quartile.”.
- (5) Additional details, procedure. I am not sure that this is necessary, and do we know for sure that it reflects exactly what our software does? Suggest deletion.
- (6) Additional details, last bullet. Suggest changing “PERCENTILE” to “percentiles”.
- (7) Examples, 2nd example. Insert “the” before “range”.
- (8) Examples, general. Suggest including an example for which no interpolation is carried out. Maybe combine this with specifying the data set as an array in {} brackets?
--Stevefanning (talk) 2020-10-31T11:59:37 (UTC)
Ronnie
all points addressed.
--Ronnie rg8888 (talk) 2020-11-27T17:37:55 (UTC)