View Single Post
Old 03-20-2012, 05:50 PM   #2035  
Home Theater Enthusiast
Kosty's Avatar

Join Date: Aug 2007
Posts: 5,063

Originally Posted by mikemorel View Post
Right. Now we know approximately how many homes have more than one BD player, or a player and a PS3, or a replacement player - 10 million.
Possibly but not necessarily. That may be a poor assumption.

Some of it also can just be an margin of error in two different reporting agencies estimates. Or one is using a different household estimate or including Canada or using shipped versus sold numbers or using survey estimates instead of industry data.

The Centris survey IIRC ( I can't find the original data) has a smaller sample size than the NPD data set we usually see. The DEG data IIRC is based on industry sales data and IIRC adjusted by industry estimates for duplicate households. Even if the NPD was all units sold or shipped in the time period it could be using a different set of assumptions and methodology that biases it in the other direction.

Survey based information tends to undercount CE items as some consumers have difficulty in describing tech electronics in their households as well.

114,235,996 is the 2011 US Census estimate for households. The DEG or NPD or Centris could also be adding in the Canadian households in the calculation as those figures are often North American market based for consumer electronics.

40.0/114.235 = 35.0%

29.7/114.235 = 26.0%

If the NPD and Centris data are using survey numbers for identification and the DEG is using industry supplied data then its not just as simple as comparing the magnitudes of each and subtracting out.

Comparing magnitudes of different data sources is always problematic. Its the trends that match not the magnitudes.

One of the data estimates especially if its survey based could easily be 5% or 10% high and the other 5% or 10% low even if they were trying to estimate exactly the same thing.

You just cannot make that assumption here that one is counting all units and the others are counting non duplicate households, they just both could have error margins in the opposite directions.

When we get the annual NPD household survey in a couple months we will get a better idea as they have a more consistent long term methodology with a greater sample size.
Kosty is offline   Reply With Quote