In processing the survey data for the current year, each school's responses to comparable questions from last year's survey were used. Thus, for example, if a school did not report on the number of Macintosh computers in classrooms then the corresponding value entered on last year's survey was used. This was intended to reduce the respondent burden and to improve the response rate, since feedback from non-responding schools in the 98-99 survey indicated that a major cause for not responding to the survey was that very little or nothing had changed and they considered it a waste of time. Of course, not all questions were asked last year, so schools that did not fill out the form will have missing data for the new questions on the 1999 survey. As users completed pages of the questionnaire this year, the results were recorded with a time-stamp. There was also a question that asked them about the status of their questionnaire -- was it complete, mostly done, just started or untouched (stored in the completion_status variable).

In creating reports, analysis was limited to a universe of schools for which the data were reliable (i.e. for which there were classrooms where there might be computers, etc.). One measurement of reliability was to identify responses that exist within the current master list of schools from the DESE core data database. To accomplish this, a series of filters was applied to eliminate data from the final reporting data sets. The following is a summary of the criteria used to perform these filters, together with some statistics regarding how many schools were effected by each filter
 

1.  Schools that could not be matched on the DESE core data master list of schools for 1998-99. There were 47 such schools. The enrollment for these schools is unknown, since we were picking up enrollment data from the core data file. Of these 47 schools, 17 actually had a completion_status value indicating that they had completed the survey. There were 23 elementary schools, 11 junior high or middle schools, 3 high schools, 3 vocational schools and 7 administrative centers. (See postoraBx.lst for a detailed listing of the 47 schools or browse the project.noschl data set - see endnote). That report also lists the 23 schools that were on the DESE master school file that were not matched on the survey data base.)

2.  School type. Certain types of schools were removed from the datasets used for analysis. The major categories of buildings excluded were ADM - administrative centers (62 of these); PRE - pre schools (22); and HOS - hospital centers (7). In all 97 buildings were excluded based on their type. These 97 buildings reported 2,575 total computers, 468 internet connected and a total enrollment of just over 6000 students.

3.  No computers and no instructional rooms. It was decided that a building with neither computers nor any instructional rooms should be excluded from the analysis. There were 84 such schools. We do not believe that there really are 84 buildings out there with no instructional rooms or computers -- we think that most of these cases are the result of incomplete response to the survey (as evidenced by the fact that these 84 buildings reported a total of 378 internet connected computers). These 84 buildings had a total enrollment of just over 24,000.

4.  Large school with no PCs. The validity of having a school these days with no computers at all in instructional rooms was debated. Should we assume that a building that has not reported any computers has not really reported accurate data? It was decided that it was indeed possible for schools to still have no computers at all, but that it was extremely unlikely in larger schools. Therefore excluded any school with an enrollment of over 300 students and which reported no computers was excluded. There were 17 such schools with a total enrollment of 9530 students. These 17 schools reported 861 internet connected computers (indicating, of course, that the 0 total computers reported represents a failure to respond to these critical questions.)

5.  No enrollment (except vocational). Eight schools that reported no student enrollment were excluded. Vocational schools never report enrollment so this filter did not apply to them.

Schools Not Omitted. It should be noted that some schools that were considered for exclusion from the final analysis were left retained. These were cases where there was evidence that the survey had not been carefully filled out for the current school year (1998-1999). There were 35 schools that were included in the totals where the completion_status item indicated that the surveys were not "complete". This means that data from the previous year were used for these schools. The decision to include these was based on the logic that using year old data would be better than just excluding all these schools (e.g., a more reliable estimate of values was possible). There were other cases where we felt that districts may have responded to the survey but not with as much enthusiasm and completeness as we would have liked.

The response rates for the 1999 Census of Technology can be calculated in several ways. This summary is provided to allow the reader to ascertain the appropriateness of the census information for their particular uses.

Gross 1999 Response Rate

The census population consisted of the entire list of districts and schools contained in The Missouri Department of Elementary and Secondary Education's (DESE) Core Data file as of March 6, 1999 and records from other data sources, such as MoreNet data from previous years. On August 28, 1999, the census was frozen for the purposes of analysis. At that time, any school with census information and which was not contained in the most current DESE Core Data file were excluded from analysis. This exclusion was done in order to allow for coding and data entry anomalies. Of the remaining 525 districts and 2220 schools, 2156 (97%) of buildings and 503 (96%) of districts responded indicating that they had completed the census. These figures constitute the Gross 1999 Response Rate for the census.

Adjusted 1999 Response Rate

Although the Gross 1999 Response Rate was quite high, there were 22 Districts and 64 buildings that did not complete the census. For these districts, their 1998 census information was brought forward and included as their 1999 census information. This adjustment allowed DESE to make the most reliable estimate of technology parameters possible because the best predictor of 1999 census data is the previous year's data. The resulting Adjusted 1999 Response Rate was 100% for districts and 100% for buildings (both calculated based on 525 districts and 2220 buildings).

The School Universe

Some schools, such as early childhood centers and special education buildings, are irregular in their use of technology; these schools were excluded from the universe of schools used to calculate student to computer ratios and other values. Other reasons for exclusion were lack of enrollment data, no rooms were reported, or because the school had enrollment >600 and did not report having any computers. A total of 206 schools were excluded for various reasons (10.7% of the 2220 schools in the final dataset). The School Universe used for calculating computer:student ratios consisted of all remaining elementary, junior high, middle and high schools in Missouri (89.3% of the 2220 schools in the final dataset). The 1998 Census of Technology used the same criteria for school type, which makes comparison of the two years' surveys possible. A complete report of excluded schools may be accessed at: /computingcensus/1999/excrpt.txt

The files: postoraBx, and excrpt.lst, may be accessed from the census web site, as well.

                                                              TOP