Comparing income inequality across Census Bureau data bureau may result in conclusion that are not, in fact, representative of the American population by and large, a new research note states. This is because of the large shifts in household wealth, particularly during times of recession. In harder times, people's job status can change more often and shift household wealth levels. This leads to larger swaths of the population frequently sliding between data sets, according to one Federal Reserve official. Thomas Garrett, assistant vice president and economist at the Federal Reserve Bank of St. Louis, said household incomes vary, making comparisons tough, if they "incorrectly assume" each quintile, equal to one-fifth of the population, contains the same households over time. "Comparing income quintiles from different years is a proverbial apples-to-oranges comparison because the households compared are at different stages in their earnings profile," Garrett said. Garrett said nearly 58% of households in the lowest income quintile in 1996 moved higher by 2005, and almost half of the households in the second-lowest quintile moved to a higher quintile over the same period. Conversely, more than 57% of the wealthiest 1% of households in 1996 fell out of that category by 2005, while more than 45% of the wealthiest 5% slid further down the scale, as well. "Income mobility muddies the picture of income inequality derived from a simple comparison of income quintiles from different years," Garrett said. Write to Jason Philyaw.