The repeat sales methodology for estimating residential price indices is based on actual appreciation of individual properties. On the other hand, the repeat sales method wastes data, typically discarding a large percentage of all sales. This article explores two issues related to the subsample of repeat sales. First, are paired sales representative of the entire population of properties that sold? Second, is there evidence that sample selectivity biases the price trend estimates? Evidence from five metropolitan areas supports a negative answer to the first question and the second question. It appears that a "lemon" or "starter home" effect causes repeat residential sales to be a biased subsample of all transactions. Cumulative price trends for the repeat subsamples can differ from the full samples over periods ranging from two to ten quarters. While short-term price trends can differ widely, there are no systematic differences among the samples over long periods of time (e.g., three years or more).The repeat sales methodology estimates real estate price trends from pairs of transactions for properties with unchanged characteristics. The methodology was originally proposed by Bailey, Muth, and Nourse (1963) and later refined by Palmquist (1979Palmquist ( , 1982 and Case and Shiller (1989). Price indices based on paired sales have been used by Case and Shiller (1989), Hendershott and Peach (1990), Pollakowski andWachter (1990), andDiPasquale andWheaton (1990).The repeat sales methodology is attractive because it is based on actual appreciation of individual properties [see Case and Shiller (1987) and Palmquist (1982)]. Therefore, it may offer better control over quantitative and qualitative characteristics of the property than alternative methods, such as hedonic regression, that control for these characteristics by using variables and parameters estimated with error.The most important criticism of the repeat sales methodology is that it wastes data, typically discarding a large percentage of the sample. For example, Case and Shiller started with a universe of 952,606 transactions in four cities over a 16-year period; they found only 358 JOHN M. CLAPP AND CARMELO GIACCOTTO 39,210 (4.1 percent) repeat transactions with no discernible changes in property characteristics. Thus, while they endup with a lot of transactions per city, the data provide limited ability to disaggregate to the neighborhood level. Abraham and Schauman (1990) found that about 2.5 percent of Federal Home Loan Bank Board sales, during 1970Board sales, during -1989 repeats. On the other hand, Mark and Goldberg (1984) found that about 40 percent of the sales in Vancouver, 1957Vancouver, -1979 repeats. As the time period lengthens for a given market area, it is reasonable to expect an increasing percentage of repeat sales, but a decreasing percentage sold without changes in property characteristics.This article explores two issues related to the subsample of repeat sales. First, are paired sales representative of the entire population of properti...