Page 1 of 1

Correlation coefficient and impact

Posted: Wed Dec 14, 2011 10:07 pm
by Tomico
Could you tell, why correlation coefficient in sensivity view and impact for risk matrix are different?

Re: Correlation coefficient and impact

Posted: Tue Dec 27, 2011 10:14 am
by Intaver Support
Correlation coefficient in Sensitivity View and risk impact in Risk Matrix or Risk Register are different numbers. Correlation coefficient is calculated first as a correlation between impact of the risk (e.g. increase of duration) and changes in project schedule (e.g. increased project duration). The coefficient will be equal 1 if even very small risk "task duration increase" is associated with increase in project duration. For example, we have project with one task (10 days duration) and two risks:
- increase task duration on 1 day
- increase task duration on 10 days
Correlation coefficient in both cases will be equal 1, however risk impact is greater for the second risk.
Therefore in RiskyProject we assume that 100% of risk impact is associated with increased duration two times. There is a similar assumption for the cost risks. As a result:
- impact of the first risk is 0.1
- impact of the second risk is 1 (two times increase of the duration)
These impacts will be multiplied on correlation coefficients obtained from Sensitivity view. The final number will be shown as an impact in Risk Matrix or Risk Register.

Re: Correlation coefficient and impact

Posted: Sun Jan 05, 2014 4:56 pm
by Kathy
I have a risk which has two alternatives:
- increase duration on 10% - 5% chance
- increase duration on 30% - 1% chance
do these alternatives have the seed?

Re: Correlation coefficient and impact

Posted: Sun Jan 05, 2014 4:58 pm
by Intaver Support
Yes. Different alternatives for the risk have the same seed by definition.

Re: Correlation coefficient and impact

Posted: Mon Jun 16, 2014 4:40 pm
by Tomico
I tried to perform analysis of a small schedule with two tasks. Each task has one risk assigned to it. I did it with and without correlations. I found that results are very close. Do you have similar results?

Re: Correlation coefficient and impact

Posted: Mon Jun 16, 2014 4:43 pm
by Intaver Support
Yes. We have similar results. We recommend to have to define correlation of risk have large probability and impact. Since these risks will have significant effect on project schedule it is important to define if these risk are correlated or correlated.