United States 1^st Class Postage History and Prediction by Dorothy Evans

My first step in analyzing the postage data was to create a table and corresponding charts in Excel. I was also aware that more recent postage data was available so I looked up changes to the postage rate at http://www.akdart.com/postrate.html from this information I created two charts and developed trend lines in Excel utilizing the trend line function. I chose to utilize exponential trend lines because these resulted in the most accurate models. Each of the 2 data sets resulted in slightly different equations. The first dataset from 1933 to 1996 resulted in the equation y = 1E-250x76.173 where y is the dependant variable, the amount of the postage, and x is the independent variable, the year. The next dataset from 1933 to 2006 resulted in the equation y = 3E-254x77.242 where again y represents the postage in cents and x represents the year.

As you can see from the predicted values in the table utilizing the predicated values from the different data sets results in slightly but very close predications for when the cost of 1^st class postage reaches 64 cents, 1 dollar, and the next 3 cent increase. A summary of my findings and predications based on these two models is below.

	Prediction 1 (based on dataset from 1933 to 1996)	Predication 2 (based on dataset from 1933 to 2006)
64 cents	Late 2021	Early 2022
1.00	2033	2034
Next 3 cent increase	2009	2009
Equation	y = 1E-250x76.173	y = 3E-254x77.242
Error statistic	26.3	26.0

I have also calculated the error statistic between the predicted value and the actual value for all the known years. As you can see the statistic is more accurate for the second prediction equation and I would therefore utilize this equation in any further predictions.

For a look at the entire excel file click here.