Need Responses to the 2 Prompts-2 Hours at the Max

Prompt 1:

If an organization cannot statistically "prove" a hazard control is effective in reducing injuries, should it take focus away from measuring the hazard control?  Why or why not?

Prompt 2:

Some authorities on safety performance measurement believe safety metrics can be predictive of incidents/loss events.  In essence, the metrics can be somewhat "crystal ball" in predicting incidents.  Others believe leading safety performance metrics are more akin to the gauges in the dashboard of an automobile that show how certain injury prevention systems (i.e., cooling, oil pump, etc.) are performing at a certain time.  Which of these lines of thought do you believe is most correct, and why?