00462 ®7b4bbd48af23178e168612bf58f499

A Graphical Aid for Analyzing Autocorrelated Dynamical Systems 467

As a side notÄ™, the two cut points chosen for the models based solely on the apparent autocorrelation pattem changes in the extracts were each within one month of major policy changes that had madÄ™ food stamps easier to access by the recipients. The first cut point coincided with the opening of satellite offices directly in areas that historically had the greatest demand for food stamps. The second cut point corresponded with a major reduction in paperwork and proof for need requirements as well as a reduction in the number of reviews of each case per year. This is possibly the reason for the explosive growth in food stamp demand siÅ„ce the growth was, reportedly, not coupled to statewide economic conditions.

The first segment is the longest. A typical movie extract for this segment is presented in FigurÄ™ 30 below. The segment corresponds with the early years which formed the "blob" in the upper left comer of FigurÄ™ 29. This particular extract represents one years worth of data. Notice that the segment has an extract that presents itself as a long ellipse. From the movie it is elear that this pattem is relatively stable, there is little noise, and that the data is autocorrelated.

FigurÄ™ 30. A representative sample extract from the first segment of the movie. The data in this extract covers the time period from September 1982 to August 1983.

Wyszukiwarka