The following is an excerpt from a workshop developed by the Church Council on Justice and Corrections which examines the roles and limitations of statistics in media reports. 

Community and Personal Safety
by the Church Council on Justice and Corrections
Republished with permission

What We Did and How We Did It

1. The role and limitations of statistics: The facilitators began by reviewing with the group some of the conditions under which surveys are carried out and statistics are "counted": 

...whose answer gets counted can depend on who picks up the telephone, and who even has a telephone... How a person answers the question can depend on who is asking, what I think their purpose is, what I may think I or my community may stand to gain or suffer as a result... How the answer is interpreted and used can depend on the purpose of the person "writing the story"...

...for statistics can be used as a "weapon" - who stands to benefit? 

All kinds of groups have an interest in perceptions of rising crime (burglar alarm companies, get tough politicians, some police wanting more in budgets, some media wanting to keep the public's and advertisers' attention,...) - who else?...

2. The role and limitations of media reporting: Human beings tend to be fascinated by violence. We appear to be drawn to watch what we fear the most. It can make us feel better about our own lives - and this becomes a major selling point for much of the news and entertainment media. Saturation coverage of one incident (Bernardo, a high school stabbing, a conflict with racial overtones), or of a highly selective range of incidents (sexual assaults, a home invasion, a swarming) - repeated daily over a long period - can escalate fear in the community in a way that erodes quality of life far more than what real risks warrant. Media reporting can also string incidents together to create the perception of a "trend", which becomes a "problem", which may not be there at all. "The temptation to weave a pattern in crime writing is formidable" -Timothy Appleby, Globe and Mail: 

"Staff Sergeant Burgemeister, is there any connection between this murder and the one last week...

"We have absolutely no evidence of that, but at this stage we're not ruling out anything...

Hence:
"Staff Sgt Burgemeister said yesterday he was not ruling out a connection between the two murders, which have left this small community reeling."

The policeman said exactly the right thing and it was reported accurately. The clear, and wrong impression though, is that there is a better than even chance they are connected. 

3. Two exercises in reading statistics carefully: