Build your own database. The phrase strikes terror in the minds of journalists who suddenly picture themselves locked in a padded room typing numbers for weeks on end, never seeing daylight.
As much as we strive to get government offices to provide information electronically, it is not always possible. Sometimes journalists must build their own database from hard-copy documents.
Carolyn Tufts, when she was a reporter for the Belleville (Ill.) Democrat, spent her free time entering information from thousands of traffic tickets into a database.
Robert Imrie, an Associated Press reporter in Wausau, Wis., wanted to analyze deer hunting accidents, but only had hard-copy forms—so he build his own database. Ralph Frammolino, a reporter for the Los Angeles Times built a database from juror questionnaires in the Simpson trial.
Building your own database also can be useful to keep track of information we get into our newsrooms regularly: accidents, homicides or other crime incidents. If you input these items as they come in, eventually you'll have a database you can use to look at trends.
Input options
In cases where large amounts of data must be entered, the reporter doesn't always have the time to become a full-time data entry person. Here are some options:
Use a professional data-entry company. They are usually quite efficient and not expensive. In many cases, you can have these companies "double-enter" the data. This means that two people enter the data, then cross-check the two files. Use a temporary service. This requires that you train the temp on how to properly enter the data or create a simple data-entry form. This also requires regular spot-checks of the data. Do it yourself in your spare time. Especially on your first projects, this will help you interview the data.
Once you get ready to build the database, here are some dos and don'ts to keep in mind:
Dos and Don'ts
Break up names into separate fields for first, last, and middle or middle initial. Plan. Plan. Plan. If your set up your database for U.S. addresses, will there ever be a point when you'll be putting foreign addresses in the database? If yes, create a field for country. In many database programs you can give these fields a "default" value. That means the computer would automatically put USA in the country field unless you specified something different. Will you want to look at the data by day of the week or month? You may want to break up the date into three parts or add a separate field for day of the week. Be consistent. If you're typing in city names or streets, use the same format each time. Use all upper case to avoid problems with case-sensitivity. Enter some of the data, then have a colleague look at the data. This is a good way to test the database to make sure it's easy to understand and they may suggest other information to include. Balance. Don't include so much that it takes you the next 10 years to enter the data, but include that data you think you "might" use: It's easier to input it the first time around. If you're entering the data yourself, have someone else spot check from the original form. Get a summary report based on the forms you're using. This can help you check the totals in your database. Do some basic integrity checks on the data: Check for duplicates, count fields and look for inconsistencies. For example, if you were inputting accidents, does your database use I-70 and I 70 and Interstate 70? In Paradox: check mark these fields individually and CALC COUNT ALL, then sort alphabetically. In FoxPro and Access: group by these fields individually, COUNT them, then sort or order by alphabetically. Backup your database frequently.
Jennifer LaFleur can be reached at (408) 920-5728, or send e-mail to JENSTER@aol.com