The next time you read housing statistics in the financial press, make sure you take the information with a grain of salt. Not all indices are equal, but they can tell the same story.

The recent Standard & Poor's/Case-Shilller U.S. Home Price Index (S&P), which features data from either a 10-city or 20-city composite, suggested a harder fall for housing than data gathered for the National Association of Realtors' Existing Home Sales.

The 10-city composite showed the lowest level of home price growth since 1996 at 0.4 percent for 2006. For the same period, the National Association of Realtors found that existing-home sales reached the third highest total on record, with 6.48 million homes sold in 2006. However, home prices rose only 1.1 percent in 2006.

Both seem accurate, but they track home sales quite differently.

The S&P indices compile data of single-family home prices for all nine U.S. Census divisions. In fact, the methodology is considered so accurate, the Census Bureau, a division of the Commerce Department , uses similar methodology in determining its new home price data, with one proviso -- that the data can be revised. (Because the data is derived from building permits, and then goes on to track completions and sales, it doesn't include the many new home sales that take place before the permits are drawn, nor does it account for contract cancellations once the permit is issued.)

Included in the analysis are the cities of Boston, Chicago, Denver, Las Vegas, Los Angeles, Miami, New York, San Franciso, and Washington, D.C.

The 20-city composite includes the above cities along with Atlanta, Charlotte, Cleveland, Dallas, Detroit, Minneapolis, Phoenix, Portland, Seattle and Tampa.

Designed to track the price changes of resale homes, the indices only include homes that have two or more recorded sales transactions at fair market value. How fair market value is determined is unclear, but the data also does not include housing which has gained exponentially in popularity -- condos and co-ops, presumably because of historical volatability. Nor does it include new homes.

NAR's existing home sales, on the other hand, does include condos, townhomes and co-ops, or multifamily homes, and calculates median home sale prices from data in major regions such as the Southwest, Northeast and West. Instead of major cities only, the data can include rural, land-only and small town sales.

"Existing-home sales, which include single-family, townhomes, condominiums and co-ops, are based on transaction closings," explains Walt Molony, spokesperson for the National Association of Realtors. "This differs from the U.S. Census Bureau’s series on new single-family home sales, which are based on contracts or the acceptance of a deposit. Existing-home sales, which generally account for 85 percent of total home sales, are based on a much larger sample -- nearly 40 percent of multiple listing service data each month - and typically are not subject to large prior-month revisions."

The NAR includes multifamily sales in its single-family homes data if the sales are to single-family households. For example, a single woman buying a condo is a household. Another factor is that monthly single family data goes back to 1968, but monthly condo data only goes back to 1999. "From 1982 thru 1988, condos were reported only on a quarterly basis," says Molony. "When the Existing Home Sales (EHS) series was created, condos weren't an important part of the housing market."

But that's changed. With single homebuyers commanding 30 percent of the market, and one in ten homeowners owning a second home, low-maintenance properties are more popular. They are more volatile. The average single-family homeowner owns their home an average of six years, the average condo owner owns only four years.

"Condos account for 12.4 percent of transactions, so they're important in measuring total market activity," explains Molony. "That being said, EHS sometimes can be skewed by changes in the composition of what's being sold. Also, since condos sometimes behave differently than single-family, it's important for us to maintain the separate series, and trends over several months are more important than focusing on individual monthly changes."

That said, Molony cautiously praises the S&P. "Shiller's numbers are based only on select metros, and I understand they're pretty good because he's measuring resales of the same properties over time, but they're not representative of the U.S. as a whole," he says.