Good evening,
using "corr" function I have noticed that it does not return the sign correctly. In particular in cases where the correlation should be negative I get a positive value.
I show below an example of data
| Provincia |
Valore ascisse |
Valore ordinate |
Correlazione |
| Chieti |
0,00346216 |
0 |
0,13534543 |
| L'Aquila |
0,0057456 |
0 |
0,13534543 |
| Pescara |
0,00406408 |
0,00081282 |
0,13534543 |
| Teramo |
0,00511682 |
0,00051168 |
0,13534543 |
In this case I used corr ([Valore ascisse]; [Valore ordinate] for report) obteining the value 0,13534543 instead of right value -0,13534543. If I calculate correlation using formula specified in details of correlation covariance-pop
([Valore ascisse]; [Valore ordinate] for report)/(standard-deviation-pop ([Valore ascisse] for report)*standard-deviation-pop ([Valore ordinate] for report))
I get right value with negative sign.
There is any reason about this behaviour?
Best regards
Domenico
------------------------------
Domenico Pagano
------------------------------