This report deals specifically with changes made to the survey forms in January 1981 and the resulting changes to the data-time series. Naturally, when a series has changed at some time point, the data after the change are no longer comparable to those before. In many cases, though, comparisons are desired that use pre- and post-intervention data as a series. It is thus necessary to have a methodology for updating the older data so that such comparisons can be made validly. To produce this methodology, the particular intervention must be modeled. However, when attempting to analyze one particular intervention, other types of interventions must be considered also. If effects of the other interventions can be modeled, the overall variability of the series can be reduced and the intervention of interest can be better isolated. Thus, in the following, we discuss (in addition to the format modifications of the forms) the trends and changes noted in the JPRS since January 1976 to December 1982. The year 1976 was chosen since it corresponds to the first year for which microdata are computerized in a universal format in the JPRS master files. We discuss, in particular, changes to the data series for inventories of: (a) motor gasoline, (b) distillate oil, (c) residual fuel oil, and (d) crude oil. These are the series studied in detail in subsequent sections of this report.
