Table 7. NEW PRIVATELY OWNED HOUSING UNITS AUTHORIZED, BUT NOT STARTED, IN PERMIT-ISSUING PLACES AT END OF PERIOD

Note: These backlog data represent the number of housing units authorized in all months up to and including the last day of the reporting period and not started as of the terminal date without regard to the month of original permit issuance. Cancelled, abandoned, expired, and revoked permits are excluded from the backlog.

Table 9. NEW PRIVATELY OWNED HOUSING UNITS STARTED, BY LOCATION AND TYPE OF STRUCTURE: QUARTERLY, 1972 TO 1977

SUMMARY OF FINDINGS

The average time from start of construction to completion of private residential buildings with 5 units or more decreased from 12.0 months for buildings completed in 1975 to 9.3 months for buildings completed in 1976. For buildings with 2 to 4 housing units the average construction time decreased from 7.4 to 6.4 months between 1975 and 1976. Multiunit buildings are considered started when work begins on the footings or foundation and completed when more than half of the units are occupied or available for occupancy. Buildings with 1 unit completed in 1976 took an average of 5.5 months to complete compared to 6.1 months for those completed in 1975.

RELIABILITY OF ESTIMATES

The distribution and average completion times in this report are based on sample survey results and may differ from statistics that would have been obtained from a complete census using the same schedule and procedures. An estimate based on a sample survey is subject to both sampling and nonsampling errors. The accuracy of a survey result is determined by the joint effect of these errors.

The particular sample selected for this survey is one of a large number of similar probability samples that might have been selected under the same specifications. Estimates derived from the different samples would differ from each other. The standard error or sampling error of a survey estimate is a measure of the variation among the estimates from all possible samples and thus is a measure of the precision with which an estimate from a particular sample approximates the average results of all possible samples. Estimates of the standard errors have been computed from the sample data. As derived for this report, the estimated standard errors include part of the effect of nonsampling errors but do not measure any systematic biases in the data.

The sampling estimate and an estimate of the standard error permit us to construct interval estimates with prescribed confidence that the interval includes the average result of all possible samples. For example, Table S-1 shows that in 1975 the average number of months from start to completion of buildings with one unit was 6.1 months and the standard error of that estimate is 0.1 month. This means that we are confident with two chances out of three of being correct, that the

average number of months from start to completion of buildings with one unit from all possible samples is from 6.0 to 6.2 months. Doubling the interval gives us limits of 5.9 and 6.3 months, increasing our confidence of being correct to 19 out of 20 times that the interval contains the average value over all possible samples. We have almost certain confidence by tripling the interval. The average estimate may or may not be contained in any one of these computed intervals; but for this particular sample one can say that the average estimate from all possible samples is included in the constructed interval with a specified confidence: that is, two chances out of three, etc.

The standard errors for estimated percentages in the distributions in Table S-2 and S-3 are shown on the chart below. An illustration of the use of the chart follows: Table S-2 shows that 5 percent of the buildings with 25 to 49 units were completed in the tenth month following the month of start. The standard error of this estimate, as illustrated in the chart, is about four-tenths (0.4) of a percentage point. This means that we are confident with two chances out of three of being correct that the average estimate from all possible samples is between 4.6 and 5.4%; doubling the interval gives limits of 4.2 and 5.8 percent, increasing our confidence of being correct to 19 out of 20 times that the interval contains the average value over all possible samples.