88-004 - Laserfiche WebLink

Home Browse Search

061 Table 1 Indicates that at the time of the 1980 Census, there vas an average of 0.64 persons age $ to 17 per household throughout the City. Analysis of the Census data indicates that it is probable that most of these persona live Ln single famil7 unite or duplexes rather than in larger apartment complexes. Because tbs Census does not provide reports crees referencinB age group and housing type, the housing characteristics of the school age population muir be inferred from other information available. The Census provides detailed abe breakdowns at the Census Tract level and a more limited breakdown at the block level, Housing data is available indicating the number of sinBle units, units in structures of 2 to 9 units, and units in structures with 10 or more unite by either Census Tract or Census Block. Figure 5 on page 6, based on the 1980 Census, indicates the number of persons below abe 18 per dwelling unit as a function of the percentage of dwelling units in structures with I0 or more dwelling units in the structure. The squares on the graph are the individual data points for the 108 Census blocks in the City with more than I00 dwelling units. The heavy line ia the best fit least squares linear regression line to the data points. The data Indicate, as one might expect, that Census blocks with a high percentage of larger apartment units tend to have fever residents below age- 18. As is clear from the graph) not all points lie on the lines lndicatinB ' that some Census blocks have a variety of unit types and some units in large structures have a larger number of persons below age 18. The correlation coefficient (r) for this relation il only -0.42, yielding an r2 value of 0,18. This value indicates that although population below 18 is related to the number of units located In structures with more than 10 unite, only 1az of the variance in residents below abe 18 in these blocks can be explained by this factor. Using Census tract data, one can obtain more precise age data, but one must use a much larger geographic area which is likely to contain a wider variety of housing types. Figure 6, based on census tracts, compares the same statistic tn Table 6 (percent of unite in structures with IO unite or more) for the 52 Census trlCtl in Slats Ann to the average number of people abed 5 to 17 in each dwelling unit. The lea~t .quares rescission l~ne .hoes a much stronser correlation ii this case, wzth a correlation coeff~cient of -0,6~, and afl r2 of 0,41, iodxcatinB that 412 of the variance of theaBe distribution can be explained by the variation in percentage of units in larBe structures, The stronger correlation may be partly a result of the elimination of the 0 to & age group from the sample.