"Zexu Ding" [email protected] wrote:

Hello Prof,

When I calculate the Person correlation, there are data as follows:

Burundi	BDI	2014	831.29486	10494920	5.3740325	59.9	2.905
Burundi	BDI	2015	781.5793	10727153	4.2975073	60.2	2.905
Burundi	BDI	2016	764.3366	10903329	5.82391	60.8	2.905

As we can see, the Cantril is all the same, thus the denominator will be zero. Should I ignore this country and report the error?

Cheers Zeke

Hi Zeke, This is not an error. The data is valid; the Cantril values just happen to be the same. (Perhaps they asked the same person each year??😉). In any case, from the point of view of data cleaning, it is clean data, and for this analysis, all you can do is ignore it for the calculation, but perhaps comment on stderr as it is unusual.

Cheers MichaelW