Q and A for crime story assignment

After speaking to a few students after class, it occurred to me that people are still struggling with how to research a story once you've received a tip. The trend that you' ve discovered from your data work is like a tip that you or your newsroom receives. Now you must follow it up. So here's a general Q and A that should help set you on your way.

Q: How do I know if it's a significant trend?

A: If the number is a lot bigger than all the others. For instance, if one ward has a rate for a particular crime that is way above the others, say 20 times greater, then the question is why. Information in databases tell us what, not why. As well, look at the actual numbers for that crime. To do this, go back to your master table and filter it. For instance, if you want to check the actual fraud numbers for Somerset ward, filter the master table by selecting only Somerset in the Ward column and fraud in the Offence column. This will give you a table with two rows of numbers for 2007 and 2008. If the numbers are really small, say an increase from 3 to 4, then the percent change may not be statistically significant. That being said, you can't go wrong with sticking with the information in the pivot table.
NOTE: If a number the "Actual" column in the master table is really big in one year and small in the other -- say 200 in 2002 and 2 in 2008 -- be immediately suspicious for it could be an error or a statistical blip. Before using the number, check with the police, but make it clear the query is for background purposes only.

Q: How do I put the trend into context?

A: Compare it to other wards that could be side by side, or at the other end of the region. Or compare the rate in your pivot table to the national or provincial average you'll find in this link from Statistics Canada that contains summary tables which break down crime stats. The one you want is the 5th title from the top called "Crimes by offences, by province and territory"
http://cansim2.statcan.gc.ca/cgi-win/cnsmcgi.pgm?Lang=E&SP_Action=Result&SP_ID=2102&SP_TYP=5&SP_Sort=1&SP_Portal=2
NOTE: A word of caution. If you want to see how the"assault ( incl sexual assaults )" offence in the Ottawa police table compares to the same offence in the StatsCan table, you must combine the categories "Sexual assault (level 1 to 3)" and "Sexual assault." As well, only compare the subcategories in the Ottawa police tables to the subcategories in the Statitics Canada tables. In other words, don't compare the "Crimes Against The Person" offence in the Ottawa table to the "Crimes of violence" category in the StatsCan table. Instead, it's more accurate to compare "Homicide (Incidents)" offence from the Ottawa police's "Crimes Against the Person" table to the "Homicide" offence in the StatsCan table.

Q: How do I find people without talking to the police?

A: This is where brainstorming comes in. Ask yourself, beside the police, who has a stake in this problem? The city politician who represents ward? The provincial or federal politicians? Or all of them? What about advocacy groups? Business groups?

Q: What sources would I consult to find these groups?

A: In the case of politicians, it's obvious. Pick up the phone. But before you do that, find out what some of these interested parties have said in public. For instance, I suggested that you create a database of all the police service board meetings dating back to 2001. Now that you have all of the minutes in one, huge document, you can use keyword searches to locate instances where people have talked about the problem you're researching. If it's prostitution, perhaps there was a large prostitution sweep a few years back. Of course, you can also consult newspaper articles and stories at CBC to find out if groups, politicians or experts have been quoted talking about an aspect of the trend you're researching. These are all federal offences, so chances are the topic has also come up in the House of Commons. So conduct a Hansard search by going to this link: http://www2.parl.gc.ca/housechamberbusiness/ChamberPublicationIndexSearch.aspx?View=H&arpiD=1&arpiJ=0&Language=E&Mode=1&Parl=40&Ses=2
The same can be done for the provincial level.

Q: How can I be sure about the nature of the crime I'm researching?

A: Some categories are easier than others. Homicide is murder. Simple enough. The only question is was the crime first or second degree. Assault is trickier? Is it sexual? Or something else? Fraud is also tough, as is a term such as mischief. So it's important to define your term. The definition will also help in your research and allow your audience to understand what you're talking about. I will be looking for a definition in the story. The definition should be explained simply and in your own words. No jargon, please!

Q: What happens if I fail to make headway with the trend I've identified?

A: Find another one. The beauty of databases is that there lots of trends that can lead to stories. It's an endless source of tips waiting to be researched in the manner I've just described. So, just to be on the safe side, it might be worth having a plan-B just in case doesn't work out.

Q: Why am I using researching trends for the years 2007 and 2008?

A: Because those are the most recent years available for the public. The Ottawa police force won't make the 2009 figures public until it has presented them to the police services board. And that won't happen until March or April. So just make sure you explain in your stories that these are the most recent statistics.

Q: Is it a good idea to set up a Google Alert?

A: Yes. If you're interested in fraud, for instance, set up an alert using the keywords such as "fraud," "white-collar crime" and Ottawa ( in order to limit the hits to cases in the city of Ottawa.). This is a low-tech way to gather possible intelligence while mining other sources such as Hansard and minutes of meetings. NOTE: And don't forget about social network sites such as Facebook where people may be talking about possible indignities they may have suffered at the hands of criminals.