Internal Reliability Checks

The few basic questions on size, growth and perceptions of change provide the opportunity to ask how valid or reliable are the perceptions and judgements of the administrators. We can first examine how reliable are the reports of population growth and of population movements into or out of the city. If there is little reliability, there would be serious question about the utility of using administrators to complete questionnaires such as this. If the reports are reliable, that is, if administrators reports roughly the same thing on different questions, we can have more faith in their perceptions.

A. Population Growth From questions about city size in 1960, 1970 and 1980, we can calculate the average annual rate of increase for each city for the 1960s and the 1970s. The next question in the survey asked specifically what is the current total rate of increase, and how much comes from in-migration or natural increase. This permits us to compare calculated rates of growth for the 1970s with stated current total rate of growth. These need not be the same, of course, since the rate for the 1970s may well have changed in the 9 years since then. Nonetheless, they should be roughly equal, since the growth rates tend to be relatively stable over time. For example, the correlation coefficient for the calculated growth rates in the 1960s and the 1970s is +.51. The correlation coefficient for the growth rate calculated for the 1970s and the reported current growth rate is also +.5. When two outlying cases are omitted the correlation coefficient increases to +.59. Both are statistically highly significant, and show that the two reports are very much the same. This supports the view that the reported rates of growth are relatively reliable.

B. Population Movement We can also ask if the administrators' perceptions of inflow and outflow or closely related to the calculations of the percent of in-migration. This will be done separately for Japan and for the other countries because Japan has given us estimates of growth from 1980 to 1989. Table I.1 shows for countries other than Japan, the minima, maxima, and means of the percent growth from in-migration by the administrators' judgement of the direction and magnitude of the flow. The overall relationship is quite clear. The mean for in- migration is highest where administrators perceive a great inflow. It declines steadily for the other perceptions, and is a negative number where administrators perceive an outflow of the population. The one exception is found in the category where administrators perceive a small outflow. Here the mean is in the correct rank, but note that the maximum is +5.5. This results comes from one city, where the administrator indicated that the in-migration rate was 9.5 percent, but perceived the movement to show a small inflow. If we omit that one outlying city, as is done in the parentheses below the entry, the results are quite consistent. Thus it appears that the administrators' perceptions of the direction and magnitude of the migration flow are quite accurate. The second panel in table I.1 shows the same data for the rate of change calculated from the population figures for 1970 and 1980. These, too, show a very consistent decline in the means from perceptions of Great Inflow to perceptions of Great Outflow. We find the same pattern when we examine the means of reported total increase as well, though these are not shown in the table.

For Japan we show the same figures in Table I.2, except that the second panel shows the means for the rate of change calculated from population figures for 1980 and 1989. These, too, show a close correspondence between the administrators' perceptions and the mean rates of increase both from in-migration and from the calculated total increase. As above, we obtain the same results for the reported total increase.

In effect, these data show a high degree of correspondence between the administrators' perceptions of population movements, and the reported or calculated rates of increase. This gives us confidence that the administrators are reporting the movement very much as it is. Thus we may also have some confidence in their judgements about other conditions as well.

TableI.1
Percent City Growth
by Administrators' Judgements of Change
(Excluding Japan)

Table I.2
Percent City Growth
by Administrators' Judgements of Change
Japan