Loading...
HomeMy WebLinkAboutGeneralizability (V2)1 Generalizability It was expected that the overall characteristics of residential households participating in this survey would be similar to the characteristics of all Alaskans households (US 2010 Census) and to households included in the ARIS database. This section summarizes areas of comparison of these three data sources. There appear to be few material differences between the three sources, suggesting that the sample may be representative of Alaska, and that ARIS data can be merged with survey research data to give a better picture of overall energy consumption patterns. Generalizability Residence Size The residential sample was expected to represent the characteristics of housing stock in the three climactic regions. Two different measures of house size were used to compare the residential data with other data sources, including the 2009 Alaska Housing Assessment 1, ARIS 2, and the 2010 US Census 3 . The comparative size of houses in the data sets examined was determined by the total square footage of the homes and number of bedrooms. The 2009 Alaska Housing Assessment divided estimates of housing stock and housing needs by community type. There appears to be a strong match between the Anchorage data and climate zone 7 and Fairbanks data and climate zone 8. While there is no exact match for climate zones in the 2009 Alaska Housing Assessment data assessment, the category of ”Rural 1,” which includes highway connected villages and off-road regional hubs, may be the closest match for Southeast Alaska and climate zone 6. Figure 1 shows mean square footage of houses by residence type in the 2009 Alaska Housing Assessment compared with residential survey data. The figure also includes error bars showing the standard error (+/- 5%) of these mean values. The figure suggests that house size as measured by mean square feet for this study is significantly smaller than house size measured in the 2009 Alaska Housing Assessment. 1 2009 Alaska Housing Assessment, prepared for the Cold Climate Housing Research Center , and the Alaska Housing Finance Corporation, Information Insights, December 2009, Part I 2 The Alaska Retrofit Information System (ARIS) Maintained by the Cold Climate Housing Research Center on behalf of the Alaska Housing Finance Corporation. 3 US census bureau, accessed at http://www.uscensus2010data.com/ 2 Figure 1 Comparison of Mean House Size The number of bedrooms in the house measures was used to determine the extent of similarities between the residential energy use survey and other existing data sets. The error bars were set at +/- 5%. The figure suggests that there are no material differences between the US Census, ARIS data and the residential survey. Generalizability- Number of Bedrooms Figure 2: Number of Bedrooms- ARIS 3 Another crucial characteristic of the dwellings in this study was their age of construction. Older homes, in general, are less energy-efficient than more modern dwellings. This is a product of improvements in building technology and energy efficiency over time. Respondents were asked to estimate date of construction of their homes. However, some respondents were uncertain of the exact year that their house was built, and responded with the decade in which it was built. To assure consistency in the data used in this analysis, all responses were converted to the decade in which the house was built. These age categories were made consistent with ARIS and 2010 US Census data. Generalizability- Decade of Home Construction Figure 3 shows that the residential sample appears to be representative of the Alaska housing stock. The Figure presents a comparison of 2010 US census data on Alaska's residential housing stock, ARIS data and the aggregate residential sample from climate zones 6, 7 and 8. Data in the 2009 Alaska Housing Assessment was not used because it was in an incompatible format. Figure 3: Decade of Home Construction Using a +/-5% margin of error, there appears to be few significant difference between the age of Alaskan homes between the three data sources. Both the ARIS and the residential study have a higher proportion of total homes built during the 1980s than reflected in the 2010 Census. ARIS also appears to over represent homes that were built in the 1980s compared with the Residential Survey and the U.S. Census. Newer homes are underrepresented 4 in ARIS. Notwithstanding these differences, it appears as if the data from this residential sample generally reflects the age of Alaskan homes in the Railbelt and Southeast Regions. Figure 4 compares the residential survey results with those of the 2010 Census. The US 2010 Census presented data on the primary type of heating fuel for all Alaskan households. This summary measure may explain the differences in the fuel use for all Alaskan houses compared with the residential survey sample of climate zones 6, 7 and 8. The census data shows a substantially higher use of natural gas and a lower use of heating oil statewide compared with the residential survey data. This may be attributable to the lack of comparability of the comparison regions. Generalizability- Primary Heating Fuel Figure 4: Primary Heating Fuel Source Opinion on Generalizability This analysis suggests that the sample is fairly close to the U.S. Census. ARIS, on the other hand, appears to underrepresent smaller households and over represent larger ones. This is to be expected, as families who participated in the Home Energy Rebate program would be more likely to have some concern about their overall energy use. This data suggests that the sample represents the housing stock of the regions studied. 5 Mean square footage The figure below shows mean square footage of houses by residence type. Respondents were asked about the total square footage of their residence. For those respondents unable to recall the precise square footage, surveyors asked them to “give a rough estimate of the square footage” of their home. For this analysis, the exact square footage reports were used. When this data was not available, the midpoint of the estimated range reported by respondents was used to estimate the residence size. Table 2 shows the average square footage of the four residence types. Single-family detached homes are the largest of the four categories, followed by single family attached homes. The size difference between multifamily homes and mobile homes varies I climate zone. There appear to be no statistically significant differences between the size of the residence categories between climate zones. Table 1: Residence Size (in Square Feet) by Residence Type and Climate Zone Residence Type Climate Zone Significant Differences 6 7 8 F Significance Single Family Detached 1789.4 2263.4 1964.3 .460 .637 Single Family Attached 1434.4 1906.5 1582.2 .119 .888 Multi Family 1088.1 1077.1 1486.4 .667 .520 Mobile Home 1160.3 1024.2 1108.0 .876 .426 Total 1364.9 1567.3 1528.0 .324 .724 Number of Bedrooms The other dimension that can be used to compare residence size is the number of bedrooms in the residents. The residential survey respondents in the three climate zones appear to have houses of approximately the same size, as measured by the number of bedrooms. Table 2 shows that there are no statistically significant differences between the average number of bedrooms among the three climate zones. 6 Table 2: Number of Bedrooms by Residence Type and Climate Zone Residence Type Climate Zone Significant Differences 6 7 8 F Significance Single Family Detached 3.38 3.33 2.93 2.59 .085 Single Family Attached 2.31 2.55 1.96 1.98 .149 Multi Family 2.0 1.96 2.36 .945 .396 Mobile Home 2.41 2.63 2.40 .723 .491 Total 2.52 2.59 2.65 Residence Age Another crucial characteristic of the dwellings in this study was their age of construction. Older homes, in general, are less energy-efficient than more modern dwellings. This is a product of improvements in building technology and energy efficiency over time. Respondents were asked to estimate date of construction of their homes. However, some respondents were uncertain of the exact year that their house was built, and responded with the decade in which it was built. Insufficient cell size prevented a statistical analysis of the Association between residence type, decade built and climate zone. Therefore, this analysis is limited to the association between residence type and decade of construction and the climate zone and decade of construction separately. Figure 5 shows the decade of construction of the four categories of residences. There appear to be substantially more homes built in the last 20 years than in 1980s and prior decades. However, there appears to be no significant association between the type of residence and the decade of its construction (N=199, No significant Differences, Chi Sq=5.81, p=.759). Figure 5: Decade of Construction of Residential Survey Homes by Residence Type 7 Figure 6 compares the decade of construction and the climate zone in which the residence is located. They are not appear to be any statistically significant differences between the decade of construction and climate zone (N=120, 84 missing, No significant Differences, Chi Sq=5.26, p=.510). Figure 6: Decade of home construction of Residential Homes by Climate Zone Primary Heating Fuel Primary Heating Fuel by Residence Type Interviewers asked homeowners about their primary type of heating fuel. Figure 7 shows that oil is the dominant fuel for all four types of residences, followed by natural gas. Electric heat is favored by residents living in multifamily homes. These differences are statistically significant (N=202, 2 missing, Chi Sq=33.39, p=.015). 8 Figure 7: Primary Heating Fuel by residence type Primary Heating Fuel by Climate zone Figure 8 shows the differences between the climate zone in which the residence was located and the preferred primary heating fuel. Natural gas is the preferred heating fuel in climate zones 7, where electricity is preferred in climate zone 6 (N=203, 1 missing, Significant Differences, Chi Sq=110.21, p=.000). Figure 8: Primary heating fuel by climate zone 9 Characteristics of the Bethel residential sample Residential Projected and Actual Samples The estimated square feet of living space in Bethel is not significantly different than the square footage of homes in other regions (F= .616, p=.764). The number of bedrooms in Bethel is not significantly different than the square footage of homes in other regions (F= 389, p=.926). Table 3: Bethel Sample Size Housing Type Sample Percent Single Family Detached 78 62.9 Single Family Attached 14 11.3 Multifamily 16 12.9 Mobile Home 16 12.9 Total 124 100 Table 4: Residential Projected and Actual Samples Housing Type Size Measure Square Feet Number of Bedrooms Single Family Detached 1522.6 2.76 Single Family Attached 1278.5 2.43 Multifamily 984.4 1.88 Mobile Home 1243.3 2.88 Average 11389.6 2.62 Age of home: decade of construction Table 5shows that 95% of homes surveyed in Bethel were built since 1970, with the highest proportion being built in the decade of the 80s. By contrast, 85% of the entire residential were built since 1970. (N=332, 6 missing, Significant Differences, Chi Sq=32.24, p=.001). The more modern construction of homes in Bethel may be related to improved construction techniques yielding greater energy efficiency. 10 Table 5: Decade Built- Bethel Primary heating fuel Oil is the most common source of heating fuel in Bethel homes. Figure 1 shows that over 87% of all homes are heated with oil. The primary source of heating fuel is significantly different than the overall sample, in which just over 60% (61.2%) of the population uses oil as the principal heating source. The use of electricity as a primary heating fuel in Bethel is also lower than the Railbelt or Southeast regions (N=203, 1 missing, Significant Differences, Chi Sq=110.21, p=.000). Figure 9: Primary Heating Fuel- Bethel Generalizability Figure 10: Decade of Home Construction, ARIS and Bethel Sample 11 Figure 11: Number of Bedroom Comparison