Housing value is based on average housing value for a
block within a census tract as reported by the 1970
Census of Housing; Block Statistics, Topeka, Kansas
Turbanized Area), HC(3)-91, October 1971.

questionnaires began to be returned, primarily due to the enroute time for the mail. By the end of the third week after mail out, about three-quarters of the participants had submitted their questionnaires. By the end of the fourth week, 90 percent had been received. The remaining 10 percent were received in the fifth through the seventh week after mail out. It is important to note that the 90 percent response was obtained in both cities in the same time frame. This finding places an order of magnitude estimate on the duration of Project Conserve. Large scale implementation should recognize this time frame when scheduling resource requirements.

Summary

The pilot test realized an overall participation rate of more than 20 percent. Considering that the pilot test was conducted in mid-summer, this one-out-of-five participation rate indicates a very high desire on the part of the public to have such a service provided. Had the pilot test been conducted in midwinter, it is likely that participation would have been even higher.

Incentives are effective in significantly increasing participation. A follow-up letter encouraging participation is definitely effective in less affluent areas and is likely effective in more affluent areas in the absence of stronger negative incentives. These undefined negative incentives appeared to be operative in the more affluent section of Topeka.

The personal contact of a follow-up telephone conversation to encourage participation and answer questions appears to yield the highest response rate. Of the home owners contacted in this manner, two-out-of-five chose to participate.

Evaluating participation rate as a funtion of affluence indicated that there is a significantly greater tendency to participate on the part of more affluent neighborhoods. In this case, affluence was measured by family income, housing value and educational attainment. It was interesting to note that within the more or less affluent groups, this relationship was not statistically significant. That is, housing value within an affluence group was not significantly correlated with response rate within that group.

-35

TECHNICAL ACCURACY

A basic issue during the development of the elements of Project Conserve was the ability of the questionnaire to capture adequate information about a home to allow the accurate calculation of the costs and savings associated with energy conserving retrofit action.

The approach to resolving this issue began with the establishment of a basic design philosophy: all costs and savings reported to the home owner should be conservative. This philosophy was implemented primarily in the computer program which produced the reports. It was at this point that the data from the questionnaire, the local cost data for fuel and home improvements and the local climatic data came together. The technique employed was to bias all estimates of cost upward by 10 percent and all estimates for savings downward by 10 percent.

The engineering principles that formed the basis for the computer model also provided latitude within which to assume a conservative posture. In all cases where such heat-loss factors as conduction or infiltration displayed a range of values for a specific structural component, the more conservative parameter was selected. Questionnaire Data

The 948 questionnaires received from participants in the pilot test demonstrated the ability of this facet of the system to collect adequate information from home owners. No one question posed consistent difficulty to a detectable degree. The confusion that was created by the presence of codes and punched card column numbers could be easily removed. This confusion manifested itself in question #32 which requested detailed window measurements. A noticeable number of participants attempted to record their window dimensions over the spaces with card column codes of similar magnitude.

Other observed, but unquantified difficulties, were detected in the following areas:

summer cooling costs were frequently omitted or
were found to disagree with the seasonal
utility bills;

seasonal utility bills were frequently left un-
answered or apparently recorded as the monthly
average rather than the five-month total;

numbers of windows and numbers of doors
frequently contained the response "all".