| Forecasts | Data | Research |

help

glossary

Definition of Terms

Although a number of technical terms may be used in this website don't let the long words scare you.  Below are easy to understand definitions of common phrases or words.

Correlation Coefficient--A correlation coefficient is a one number measure of how well something which is being predicted "fits" the underlying assumptions used to predict it.  Or it can be thought of as how strong the relationship is between the number being predicted and the numbers and relationships used to predict it.   The correlation coefficient ranges from 0 to 1.  A correlation coefficient of 0 means there is no relationship and 1 means there is a perfect relationship.  For purposes of comparison   correlation coefficients in the 0.4 to 0.6 range are considered good in psychology and the social sciences where coefficients in the 0.8 to 0.98 range are considered good in the physical sciences and engineering.  In other words the closer the number is to one the better.

Model--A model is a symbolic representation of how something behaves or it is a description of causes and effects.  An example may be to say that interest rates rise when more people want to burrow money.  And then from this simple idea or model interest rates are related to demand for money using a bunch of fancy math and numbers.

Standard Deviation--This is a term from statistics which is a measure of how much a number can expect to jump around the average due to unexplained or dare we say random events. 
For instance if the average is 10 and the standard deviation is 1 then more than half of the measurements are expected to fall within the range of 9 to 11.  It actually works out that over 99% of all points should fall within plus or minus three times the standard deviation.  This means almost all measurements should fall between 7 and 13 in the example above.  Below is a chart showing the chance of a data point falling within so many standard deviations of the predicted value:

+/- 1 Standard Deviation       68.3%
+/- 2 Standard Deviations     95.4%
+/- 3 Standard Deviations     99.7%

NSA-Not Seasonally Adjusted.  No adjustments are made to the numbers because of seasonality or cycling within the year.

SA-Seasonally Adjusted.  The numbers are adjusted based on seasonality or cycling within the year.

SAAR-Seasonally Adjusted Annualized Rate.  The numbers are adjusted for seasonality and then given at a yearly rate.


© 2000 Applied Reasoning Inc.  All Rights Reserved.