F. Objective Indicators

For both Korea and Japan, the government collects and publishes a number of objective indicators for individual cities. We can identify some of those indicators that reflect conditions roughly similar to those on which the urban administrators gave their assessment scores. For example, the numbers of automobiles per capita might be roughly equated with administrators' judgements of the problem of traffic volume. For indicators of this type, we can ask to what extent the variance of the indicators by city matches the different judgements of the administrators. The correlation or correspondence between the objective indicator and the administrators' assessments can give us a measure of the validity of the administrators' assessments.

There is, of course, a basic problem we face in attempting to assess validity in this manner. None of the indicators we have really provides information on the quality of the problem or condition. For example, where vehicles per capita is high, we might expect traffic volume to appear as a serious problem for an administrator. At the same time, we know that the perception of traffic volume as a problem depends on some combination of numbers of vehicles and the quality of the road network. With the same number of vehicles, an open city with wide avenues, over passes, and extensive parking facilities will present less of a problem to any observer than a city with narrow streets, many bottlenecks and little parking. That is, the validity of the objective indicator as a measure of the quality or problem posed by that condition is very much open to question. Nonetheless, using the objective indicator provides at least one mechanism for understanding more fully what the administrators mean by the scores they give.

The objective indicators can be used in another way as well, however. The technique is known as deviant case analysis, and is often a powerful research tool in the social sciences. As we shall see, for example, in some cases there is a fairly close correspondence between objective indicators and the administrators' scores. In these cases, we can examine the distribution of the scores and objective indicators, and often find significant outlying or deviant cases. We can identify those outliers to focus attention on the specific problems of specific cities to help us better understand the nature of urban problems. For example, if we see a generally close correspondence between actual tax revenues and the administrators' scores on the tax base, we can look for cities that deviate from this general pattern. We look for cities with low revenues and high administrators' scores, and ask what those cities do to make their lower tax revenues go further. The same could be done for cities with high tax bases but a low score from the administrator. This is known as deviant case analysis.

To proceed with this analysis, we shall treat Korea and Japan separately. In each case, we shall examine the correlations between certain objective indicators and the administrators' scores on roughly similar problem areas. We shall select a few indicators where the correlations are relatively strong. For these, we shall examine the mean of the objective scores for cities at each level of the urban problem scores. From this we can look for deviant cases, which might suggest fruitful avenues for future research.

Korea

The government of the Republic of Korea publishes data on fourteen objective indicators that roughly match the conditions on which the administrators have made judgements about the seriousness of the problem. In four of these the correlation coefficients indicate a fairly strong relationship (although only one, for sewage, was statistically significant at the 1 percent level). These include the following:

1. The traffic volume score shows a -.17 correlation with the number of vehicles per kilometers of road in the city.
   
2. The quality of high school education score in the city correlates +.21 with the ratio of high school teachers to students.
   
3. The score for the size of the revenue base correlates +.24 with a figure of the total city budget per capita.
   
4. The Sewage score correlates +.48 with a score for what is called the sewage diffusion ratio. This measures the extension of sewage facilities to households and establishments (This correlation coefficient is statistically significant at the 1 percent level.)

Table 13 shows the means of the objective indicators for each level of the administrators' perception of the seriousness of the problem in that condition.

Table 13
Mean Values of Objective Conditions by
Administrators' Problem Scores in Korea

a. Traffic Volume. Overall the relation between the administrators' perceptions of the problem of traffic volume is closely related to the number of vehicles per kilometer of road. The mean scores decline almost steadily as the administrators' perceptions become more positive. The one deviant case, Masan, is the only case where the administrator scores traffic volume an advantage for the city despite its having the third highest ratio of all cities in vehicles to road.length, and a relatively high level of density (6636 persons per square kilometer). This suggests it would be useful to examine Masan in greater detail with field observations. The question is whether the administrators has made a mistake in scoring this problem area, or whether Masan has done something to alleviate its traffic problems. When we remove Masan from the calculation, the correlation coefficient rises to +.31, which is statistically significant at the 10 percent level.

b. High School Education. Again the objective indicators move in the predicted direction. The higher the ratio of high school teachers to students, the higher the administrators tend to score the issue of high school education. There is one deviant case, Gumi, whose administrator rates high school education a special problem, despite the fact that it has a higher ratio of teachers to students than any other city covered in the enquiry. It might be that Gumi does nave special problems, because it has an unusually high proportion of young women working in the electronics assembly industry. The problem this raises may need special attention by schools. When Gumi is excluded from the calculation, the coefficient rises to +.37, which is also significant at the ten percent level.

c. Budget. Once again the relationship is not a strong one, but it is in the predicted direction. The higher the city's percapita revenues, the more favorably the administrators score the issue of the revenue base. Here, however, there are three interesting deviant cases. Masan and Uijongbu give a relatively high score to the issue. For them it is only a "problem", while for most others it is an "urgent major" or "serious" problem. Yet these two cities have lower than average levels of percapita revenues. One can ask whether these cities are in some way more efficient in their use of revenues, or whether they have other resources that give them less trouble with revenues. The other deviant case is Yosu, whose administrator sees the problem as a special or serious one, despite the city having the highest percapita revenues of all cities. Does Yosu have special problem, and higher than normal costs that make its administrator score this condition serious, despite the high objective score? Excluding these three cities from the calculation, raises the coefficient to +.32

d. Sewage. Finally, the administrators scores for sewage are strikingly higher when the objective indicator of sewage services increase. There is not much difference between the two lowest scores: the administrators' judgement of urgent or special problem. But above that, the apparent quality and extension of the sewage system rises very rapidly to quite high levels, where the administrators see this to be satisfactory or even an advantage to the city. There are no significant deviant cases in this distribution.

In the Korean case, therefore, there is relatively strong support for the validity of the administrators' assessments of problems, at least for these four problem areas. The analysis also pointed, however, to useful follow-up research activities for urban officials and scholars. Such follow-up might point to special activities or procedures that help address urban problems.

Japan

For Japan we have only four objective indicators, three of which show a fairly strong relationship with the administrators' scores. There is no relationship between administrators' scores for water and the government's diffusion ratio for water. For three other indicators, however, there are fairly strong relationships. These are as follows.

1. The score of sewage correlates +.41 with the cities' diffusion ratios for sewage.
2. The score for quantity of personnel correlates +.59 with the ratio of government officials to population.
3. The score for hospital care correlates +.15 with the number of doctors per 1000 population in each city.

Table 14 shows the mean scores on each of the objective indicators for each level of administrators' problem score in Japan.

a. Sewage. Sewage scores also tend to rise with the objective sewage diffusion ratio, though note that no administrator rates the sewage condition as "advantageous" for the city. Here there are two interesting cases, both of which score the sewage issue merely "a mirior problem". One, Naha, has the highest sewage diffusion score of any city. This raises the question of whether Naha has special sewage problems that are not effectively addressed even with its high objective rating. At the same level of administrator score is Utsunomiya, with a much lower objective score on sewage diffusion. Does Utsunomiya have less problems with sewage, or has the city done something else that makes its moderate level of sewage diffusion work more effectively? Since one of these cases supports the overall relationship, and the other weakens only slightly, removing them from the calculation reduced the coefficient slightly to +.39.

b. Personnel Quantity. One of the strongest correlations we have found is between the administrators' scores on personnel quantity, and the ratio of government officials to population. The means of this ratio rise consistently and dramatically as the administrators' perception scores rise. The overall correlation of +.59 is significant at better than the 1 percent level. There is only one case that is mildly deviant. Nagano has the highest ratio of officials to population, yet the administrator still rates personnel quantity "a minor problem". Does Nagano have special problems that require more officials? Or are the many officials in some way not available to the city government? Excluding Nagano from the calculation raises the coefficient to +.69.

c. Hospital Care. There is only a weak relationship between the score on hospital services and the objective indicator of doctors per population. The relationship is in the right direction, but it is not statistically significant. Here again we have two deviant cases at opposite ends of the spectrum. Oita's administrator rates hospital care advantageous (5) to the city, but it has one of the lowest doctor to population ratios of all cities. At the other end, the Yamagata administrator considers the problem an "urgent" one (1), despite the city's substantially higher than average doctor to population ratio. This raises the possibility of a preliminary comparative analysis that would ask what advantages Oita has that Yamagata does not enjoy in hospital care. Excluding these two form the calculation raises the coefficient to +.33, which is statistically significant at the ten percent level.

Table 14
Mean Values of Objective Conditions by
Administrators' Problem Scores in Japan

Conclusion

This analysis has done two things for us. Where we do have some objective indicators of conditions roughly related to the problem areas on which the urban administrators gave scores on the extent of the problem, the administrators' scores tend to correspond roughly with those objective indicators. In some cases the correspondence is quite close, as in Korea's sewage, and Japan's sewage, hospital service and especially personal quantity issues. We cannot be certain, of course, that this correspondence is a general phenomenon that can be extended to all other problem areas and to other countries. It does, however, give us some confidence that the scores we have from the administrators represent real problems on the ground.

In addition, however, the analysis points to special cases of cities whose problems are both better and worse than expected given the objective conditions under which they live. This can help direct attention toward those special cases where we can learn more about how specific objective conditions affect urban administrators, and also perhaps how administrators have addressed problems in special ways to ameliorate them.