What the hell does rebasing to 100 or whatver number mean?
Is it some sort of scaling, do you subtract the difference or is it a percentage thing?
Turn on thread page Beta
Rebasing in numerical tests watch
- Thread Starter
- 21-10-2008 13:45
- 22-10-2008 00:23
It's to do with indices. If you're comparing data points over time, one way to make the comparison simpler is to label a particular point as the "base", call it 100, then express all other points as a number relative to 100. Inflation in particular is often done this way. If you've got a basket of goods that cost £20 in 2005, £24 in 2006 and £26.40 in 2007, you might pick 2005 as the base year and call the price index 100, then 2006 has an index of 120 (24 / 20, x 100) and 2007 has an index of 132. It's then really easy to see that prices went up 20% in 2006, and were 32% higher in 2007 than in 2005.
When you get a long way from your base year, though, and you're more interested in recent data than the base year, you can make your numbers more useful by rebasing. That's just tweaking all your indices up or down by the same proportion, such that a different year becomes 100. So, if we wanted to make 2006 our base year, we need to bring it down from 120 to 100, and everything else by the same proportion. So we multiply all the numbers by (100 / 120). 2006 becomes 100 and 2007 becomes 110 (132 x (100 / 120) = 110), so now we can easily see that prices went up by 10% in 2007. 2005, not that we care so much, becomes 83.3.
- 24-09-2012 17:02
great answer, thanks!
- 18-02-2018 21:19
Thanks a great deal