Usage Notes for the U.S. Place-level Reports

Places (cities) must appear on both the 1980 and 1990 Summary Tape File 3C in order for them to be included here. The primary qualification for a city to be on an STF3C file is that it have a population of at least 10,000 at the time of the census. Therefore you will not find any places that were smaller than this in *either* 1980 or 1990. The other obvious requirement for inclusion on STF3C is that a place exists at the time of the census. In some instances cities have incorporated during the 80's and now have populations of over 10,000: for example, Maryland Heights and Chesterfield, Mo. But neither existed in 1980 (there was a Maryland Hts CDP but its defini- tion differed from the incorporated version and it had less than 10,000 population); and neither will appear in this trend report.

Unlike states and counties, which tend to have very stable boundaries over time, places can be very volatile. Their boundaries often change due to annexations, mergers and de-annexations. The "trends" shown in this report may well be as much or more the result of such transactions as they are of what may be happening in a fixed geographic location. A city of 25,000 may turn into a city of 40,000, not because of any sudden increase in the birth rate or in inmigration, but because they annexed some of their formerly unincorporated suburban areas.

Except for the 10,000 population limit these same comments apply to any place-level trend report (i.e. those based on STF3A data, which may include places of any size.)

Standard Trend Report Information

The Basic Demographic Trend Report series presents key demographic items extracted from the 1980 and 1990 Summary Tape Files 3. The items chosen are a subset of those shown in the MSCDC's Basic Tables Report series. Items were chosen where it was possible to get a close correspondence of the item definition across the two censuses. Thus, for example, we did not attempt to present a distribution table of family or household income values, since the intervals used for the two decades were different and inflation would have distorted them even if they had been left the same. So, there was no way to get comparability and thus they were excluded from the report. These reports were generated both as printed reports and as text files for (originally) Gopher directories (now converted for comparable web access, of course). Unlike the Basic Tables reports, the format of the reports for these two media will be identical. With the Basic Tables reports the printed versions displayed four geographic areas per page; for these Trend Reports each geographic area occupies a full page.

Each line of the report after the header lines contains all the information related to one data item for the geographic area identified at the top of the page in the AREA and GEOCODE fields. The data items (variables) are organized into vertical tables of related items. The variable label is shown in the left most column. 1980 and 1990 values are shown along with PCT values (where relevant) for each year. The denominator for the PCT calculations are the values reported in the first line of the table except for tables T2 and T3, where the denominator is total population. The CHANGE columns show the difference between the 1990 value and the 1980 value (AMOUNT) and then the PCT change. The latter is defined as the value of the change amount as a percentage of the 1980 value, i.e. 100*(VALUE90-VALUE80)/VALUE80. Negative values indicate a decrease in the value of the variable over the decade.

The "CHG IN %" column is simply the difference between the 1990 PCT column and the 1980 PCT. It is important to distinguish between "PCT CHANGE" and "CHG in %". The former usually measures the change in a count, while the latter measures the change in that count as a percentage of its universe. For example, the city of St. Louis had a value of -8.8% for the PCT CHANGE in its African- American population over the decade, but a +1.9 value for the corresponding CHG IN %. What this tells you is that while the city was losing black population in absolute terms, the percentage of the population that was black actually increased by almost 2 percentage points.

As noted in the Notes at the bottom of each page, income, average housing value and median rent figures for 1980 have all been adjusted for inflation using the Consumer Price Index inflation factors cited.

The program to generate these reports matches geographic areas based on the value of the GEOCODE variable. If a value appears on only one of the two (1980 STF3 and 1990 STF3) files a record of this is written to a NOMATCH exception file which is then printed as an appendix to the report. This appendix will not appear on web versions of the reports. The important thing to keep in mind here is that there may be cases where the same geographic code does not pertain to the same geographic area over time. The most common example of this are with place (city) codes, where annexations or de-annexations may alter the area comprising the city but the code used for it does not change. ZIP codes are also notorious for changing their definitions while maintaining their "code" names. ZIP codes which were created during the decade will not appear in the report since there will be not 1980 data for them. This can be "fixed" by doing special processing on the 1980 file to apportion data from smaller geographic areas to estimate the 1990 geography but users should not assume that this has been done unless a report has specific notes to that effect.

There is currently no general data set equivalent of these reports as there is with the Basic Tables reports. However, all the items from the 1980-VALUE and 1990-VALUE columns can be captured in a data set as a spinoff from the report generating process and this data set can be saved, converted and/or downloaded. The CHANGE and CHANGE IN % figures are not saved but can be readily recalculated from the original 1980/1990 values. Contact the Missouri State Census Data Center regarding getting machine readable files related to these reports. The SAS (r) source code for generating the reports is stored as member x3trept in our SAS macro library. The code is a bit messy since it was originally written for use in an MVS environment and was then altered to run under UNIX.