Anyone who follows the media has probably heard many times that the rich are getting richer, the poor are getting poorer, and incomes of the population in general are stagnating. Moreover, those who say such things can produce many statistics, including data from the Census Bureau, which seem to indicate that.

On the other hand, income tax data recently released by the Internal Revenue Service seem to show the exact opposite: People in the bottom fifth of income-tax filers in 1996 had their incomes increase by 91 percent by 2005.

The top one percent -- "the rich" who are supposed to be monopolizing the money, according to the left -- saw their incomes decline by a whopping 26 percent.

Meanwhile, the average taxpayers' real income increased by 24 percent between 1996 and 2005.

How can all this be? How can official statistics from different agencies of the same government -- the Census Bureau and the IRS -- lead to such radically different conclusions?

There are wild cards in such data that need to be kept in mind when you hear income statistics thrown around -- especially when they are thrown around by people who are trying to prove something for political purposes.

One of these wild cards is that most Americans do not stay in the same income brackets throughout their lives. Millions of people move from one bracket to another in just a few years.

What that means statistically is that comparing the top income bracket with the bottom income bracket over a period of years tells you nothing about what is happening to the actual flesh-and-blood human beings who are moving between brackets during those years.

That is why the IRS data, which are for people 25 years old and older, and which follow the same individuals over time, find those in the bottom 20 percent of income-tax filers almost doubling their income in a decade. That is why they are no longer in the same bracket.

That is also why the share of income going to the bottom 20 percent bracket can be going down, as the Census Bureau data show, while the income going to the people who began the decade in that bracket is going up by large amounts.

Unfortunately, most income statistics, including those from the Census Bureau, do not follow individuals over time. The Internal Revenue Service does that and so does a study at the University of Michigan, but they are the exceptions rather than the rule.

Following trends among income brackets over the years creates the illusion of following people over time. But the only way to follow people is to follow people.

Another wild card in income statistics is that many such statistics are about households or families -- whose sizes vary over time, vary between one racial or ethnic group and another, and vary between one income bracket and another.

That is why household or family income can remain virtually unchanged for decades while per capita income is going up by very large amounts. The number of people per household and per family is declining.

Differences in the number of people per household from one ethnic group to another is why Hispanics have higher household incomes than blacks, while blacks have higher individual incomes than Hispanics.

Considering the millions of dollars being paid to each of the anchors who broadcast network news, surely these networks can afford to hire a few statisticians to check the statistics being thrown around, before these numbers are broadcast across the land as facts on which we are supposed to base policies and elect presidents.

Now that the Internal Revenue data show the opposite of what the media and the politicians have been saying for years, should we expect either to change? Not bloody likely.

The University of Michigan study, which has been going on for decades, shows patterns very similar to those of the IRS data. Those patterns have been ignored for decades.

Too many in the media and in politics choose whatever statistics fit their preconceptions.