Data on year built are more susceptible to errors of response and nonreporting than data on many of the other items. In most cases, the information was given according to memory or estimates of the occupants of the structure or of other persons who had lived in the neighborhood a long time.

Duration of vacancy.--The length of time a housing unit was vacant was computed from the day the unit became vacant until the day of enumeration. It should be noted that the time period is not the total time a unit remains unoccupied nor the time a unit has been in a particular vacancy status, but it is the duration of vacancy up to the day of enumeration. For newly constructed units, it represents the time period since the date when the unit was considered a vacancy, that is, when construction had reached the point that all exterior windows and doors were installed and final usable floors were in place.

The time intervals used in the tables represent full months, calculated from a date in the month to the same date the following months. For example, if the unit became vacant on October 29 and was still vacant on the day of enumeration, December 20, the time reported would be "1 up to 2 months," meaning that the unit had been vacant for more than one month but less than two months. Or if the unit became vacant on November 25 and was still vacant on December 20, the time reported would be "less than 1 month."

Plumbing facilities. --Housing units "with all facilities" consist of those which have: Both a flush toilet and a bathtub or shower inside the structure for the exclusive use of the intended occupants, and hot running water.

Units "lacking facilities" consist of those which lack one or more of the plumbing facilities or which lack exclusive use of these facilities. For example, included as "lacking facilities" would be a downstairs apartment with hot running water but whose occupants would share the bathroom with the occupants of the upstairs apartment.

Facilities were considered "inside the structure" if they were located within the housing unit or elsewhere in the structure. Facilities are for exclusive use if only the intended occupants of this one housing unit would use them.

A housing unit was considered as having hot running water whether it was available the year round or only part of the time. For example, hot running water may be available only during the heating season or at various times during the week.

Monthly rent. --The monthly rent is the amount asked for the unit at the time of enumeration, regardless of whether it is to include furniture, heating fuel, electricity, cooking fuel, water, or other services. If the amount of rent were to vary during the year, the rent reported was the monthly amount asked at the time of enumeration. As in the 1960 Census, the statistics in this report are for all vacant rental units in urban areas and those in rural areas which are located on places of less than 10 acres; thus rural vacancies on places of 10 acres or more are excluded. The statistics in the reports for 1960 or earlier exclude rental vacancies in rural areas which were classified as farm by the 1950 definition (the classification depending on the reply to the question "Is this house on a farm or ranch?”). Data for such units are excluded because of the difficulty of separating the rental for the housing unit from the rental for the land.

The median monthly rent is the rent which divides the series into two equal parts, one-half of the units with rents higher than the median and the other half with rents lower than the median. In the computation of the median, a continuous distribution was assumed; and the limits of the class intervals were assumed to stand at the midpoints of the 1-dollar interval between the end of one of the rent groups and the beginning of the next. For example, the limits of the interval designated $30 to $39 were assumed to be $29.50 and $39.50.

Inclusion of utilities in rent.--The utilities included in the inquiry were heat, light, cooking fuel, and water. The statistics reflect whether all or not all of these utilities are provided for in the amount of rent asked at the time of enumeration, not what could be provided for more or less rent. Data on the inclusion of utilities are limited to the same units as those for which rent is reported.

Sale price of 1-housing unit structures.--The sale price is the amount asked for the property, including the structure and its land. Statistics on sale price for this report are limited to units in 1-housing unit structures, without business, and with only one housing unit included in the property. As in the 1960 Census, the statistics in this report are for all vacant homeowner units in urban areas and those in rural areas which are located on places of less than 10 acres; thus rural vacancies on places of 10 acres or more are excluded. The statistics in the reports for 1960 or earlier exclude homeowner vacancies in rural areas which were classified as farm by the 1950 definition (the classification depending on the reply to the question "Is this house on a farm or ranch?"). Data for such units are excluded because they would reflect varying amounts for

land, business uses, or for more than one housing unit on the property.

The median sale price is the amount which divides the series into two equal parts, one-half of the units with prices higher than the median and the other half with prices lower than the median. The median was computed on the basis of more detailed tabulation groups than are shown in the tables and was rounded to the nearest hundred dollars.

Urban-rural residence.--The territory classified as urban is the same as that in the 1960 Census. Urban housing comprises all units in (a) places of 2,500 inhabitants or more incorporated as cities, boroughs, villages, and towns (except towns in New England, New York, and Wisconsin); (b) the densely settled urban fringe, whether incorporated or unincorporated, of urbanized areas; (c) towns in New England and townships in New Jersey and Pennsylvania which contain no incorporated municipalities as subdivisions and have either 25,000 inhabitants or more or a population of 2,500 to 25,000 and a density of 1,500 persons or more per square mile; (d) counties in States other than the New England States, New Jersey, and Pennsylvania that have have no incorporated municipalities within their boundaries and have a density of 1,500 persons or more per square mile; and (e) unincorporated places of 2,500 inhabitants or more. The remaining units are classified as rural.

SOURCE AND RELIABILITY OF DATA

Source of data.-- The estimates presented in this report are based on data obtained in connection with the monthly population sample survey of the Bureau of the Census.

The statistics for the fourth quarter 1967 are averages for the three months, Oct., Nov., and Dec. The statistics presented for the year are averages for the four quarters of the year. Beginning with the second quarter 1960, separate living quarters were enumerated according to the 1960 Census housing unit definition. For the preceding periods, living quarters were enumerated according to the 1950 definition of The result of the change in dwelling unit. definition is believed to be too small to affect the vacancy rates.

each month's sample. Of this total, about 35,000 were occupied housing units; about 3,500 were vacant housing units; and the remainder were units which, at the time of enumeration, had been demolished, or for other reasons were not living quarters within the scope of the survey. Detailed information was obtained for each vacant unit in the sample, beginning with the second quarter 1955.

Reliability of the estimates.--Since the quarterly statistics presented in this report are based on a sample, they are subject to sampling variability and may be expected to differ from results that would have been obtained from a complete count in which identical enumeration techniques were employed. The standard error is a measure of sampling variability. The chances are about 68 out of 100 that the difference due to sampling variability between an estimate and the figure that would have been obtained from a complete enumeration is less than the standard error. The chances are about 95 out of 100 that the difference is less than twice the standard error and about 99 out of 100 that it is less than 2 1/2 times the standard error.

The standard errors of a few of the more important statistics are shown in table L. To illustrate: The rental vacancy rate for the United States for the quarter is estimated as 5.6 percent. As shown in table A, the standard error of this estimate is about 0.2 percent. The chances are, therefore, about 68 out of 100 that a complete enumeration would have yielded an estimate between 5.4 and 5.8 percent.

The annual average rates are subject to a smaller sampling variability than the quarterly rates. The variability for the annual average rate can be approximated by taking one-half of the standard error of the rate in table L.

Comparison of characteristics of vacancies for the fourth quarter 1967 with those for previous quarters (Series H-111, Nos. 1 to 50) reveals that many of the differences are small. A small difference, particularly when it is based on a small subclass of vacant housing units, should be interpreted with care. As can be seen from table A, the sampling variability in such cases may be large relative to the difference.

In addition to sampling variability, the data. are subject to errors of response and biases due Factors affecting accuracy of to nonreporting. reporting are the respondent's knowledge of the facts and the enumerator's ability to obtain accurate information and classify the unit with respect to such items as condition, seasonal status, number of rooms, rent, sale price, and the like.