The numbers used in this Bangor Daily News analysis originate with the Internal Revenue Service, which tracks county-to-county moves using tax returns.

For instance if John Smith, or more precisely John Smith’s Social Security number, claims Aroostook County as his primary residence on his federal tax return in 1998 and the next year files his return in Hillsborough County, Fla., Smith and any dependents he claimed as exemptions would be counted as having migrated from one county to another.

The numbers used in the stories and graphics in this series represent the combined number of tax returns and exemptions recorded by the IRS for tax years 1995 to 2000.

For several reasons listed below, the IRS numbers are not intended to provide a precise count of those moving from one place to another, but instead show the direction and flow of migration patterns around the country.

The numbers must be viewed with some qualifiers:

. Not everyone files a tax return. That tends to skew the numbers toward those people who have an income level that makes filing a return necessary.

. Because the IRS only records the Social Security number of the primary taxpayer, divorce, deaths and changes in income levels could skew the results.

. On some returns, the address listed may not actually represent the primary address of the taxpayer. Instead the address may be a post office box, business or that of the taxpayers accountant.

. If fewer than 10 taxpayers move from one specific county to another, they are not counted as migrants to a county but are instead lumped into a group of migrants to a given region. For instance, if seven people moved from Penobscot County to Los Angeles County, Calif., from 1997 to 1998 they would be counted only as moving to a different state in the West. The result is a bias toward counties with more popular destinations.

In determining the 20 most popular metropolitan areas for those leaving the northeastern eight counties, The Bangor Daily News used estimates in cases where a metropolitan statistical area, as defined by the U.S. Census Bureau, included partial counties.

To more accurately assess the impact of migration on northeastern Maine as a whole, the NEWS analysis, when applicable, did not include moves within the northeastern eight counties. For instance, if someone moved from Presque Isle to Bangor in 1996, the NEWS did not count that person as either an out-migrant from Aroostook County or as an in-migrant to Penobscot County except in the graphic titled “Comings and goings” on the next page.

Jeff Tuttle, projects reporter