There are a few adjustments to the data for getting a better impression of the data. For example "real" data attempt to minimize the distortions that come from

There are a few adjustments to the data for getting a better impression of the data. For example "real" data attempt to minimize the distortions that come from changes in prices, while the raw "nominal" data measure economic activity. Another adjustment is for the seasonality of data. Here's a graph of some a few retail data series, where not seasonally adjusted appear with the seasonally adjusted data. [1] Here you can take a look at the most recent release of the Employment Situation.[2] It states that the survey of businesses indicated 217,000 more people working. Now, look at the table of the data.[3] You have May: 139,192 thous Apr: 138,272 thous => Surveys of businesses therefore indicated that 920 thousand more people were working! So, what's going on? It is that no one cares about reporting the seasonal cycles in employment, because it isn't news that helps you understand economic trends. So, people actually look at the seasonally adjusted data. May: 138,463 million Apr: 138,246 million => 217 thousand more people In order to come with the seasonal factors the Census (and BLS and BEA) use an autoregressive integrated moving average. Back to your original question, what does it mean to say that winter sucked? It means that people's prosperity was lowered by the winter, just like it always is, but while the seasonal factors usually remove that effect this time they are part of the seasonally adjusted data because the seasonal factors are derived from past data with less severe winters. Really it comes down to the fact that a single score is inappropriate for an understanding of the economy. GDP measures GDP, and something else with a different methodology would measure something else. You can find the fallacy acted out online with sites that purport to improve the data with adjustments without explaining the tradeoffs that come with adjustments by definition. [1] http://research.stlouisfed.org/fred2/graph/?id=RSCCAS,RSCCASN,RSFSDP,RSFSDPN http://research.stlouisfed.org/fred2/graph/?id=RSCCAS,RSCCAS..., [2] http://www.bls.gov/news.release/empsit.nr0.htm http://www.bls.gov/news.release/empsit.nr0.htm [3] http://www.bls.gov/news.release/empsit.t17.htm http://www.bls.gov/news.release/empsit.t17.htm