‹ Consumption & Efficiency

Commercial Buildings Energy Consumption Survey (CBECS)

Previous CBECS Survey Data 2003 | 1999 | 1995 | 1992 |

1989

  • Building Characteristics Tables
  • Consumption and Expenditures Tables
  • Microdata
  • Released: January 2009

    The 1989 CBECS Public Use Files are comma separated value (.csv) files that each contain 5,876 records. They represent commercial buildings from the 50 States and the District of Columbia. Each record corresponds to a single responding, in-scope sampled building, and contains information for that building such as building size, year constructed, type of energy used, energy-using equipment, and conservation features.

    The smallest level of geographic detail available is the Census division, of which there are nine in the U.S. No state level data are available.

    1989 CBECS Building Characteristics, Consumption, Expenditures
    Layout Files Data Files Revised Date
    File 1: Summary File TXT CSV 12/08
    File 2: Building Activity TXT CSV 12/08
    File 3: Operating Hours and Weather TXT CSV 12/08
    File 4: Building Shell, Equipment, and Multibuilding Facilities TXT CSV 12/08
    File 5: End Uses of Major Energy Sources TXT CSV 12/08
    File 6: End Uses of Minor Energy Sources TXT CSV 12/08
    File 7: Lighting and Conservation Features TXT CSV 12/08
    File 8: Electricity TXT CSV 12/08
    File 9: Natural Gas TXT CSV 12/08
    File 10: Fuel Oil TXT CSV 12/08
    File 11: District Steam and Hot Water TXT CSV 12/08
    File 12: District Chilled Water and Sum of Major Fuels TXT CSV 12/08
    File 13: Imputation Flags for Summary Data, Building Activity, Operating Hours, Shell and Equipment TXT CSV 12/08
    File 14: Imputation Flags for End Uses TXT CSV 12/08
    File 15: Imputation Flags for Lighting and Conservation Features TXT CSV 12/08
    All Format Codes TXT 12/08
    All Layout Files and Format Codes PDF 12/08
    Questionnaire PDF 2/09

    File Organization

    Because of the size of the CBECS questionnaire, the variables were separated into groups by subject matter. These 15 smaller files make it easier to manipulate the data.

    Several variables are frequently used in the analysis of commercial energy data. These core variables are included in each group of variables:

    • BLDGID4: building identifier, which is the link between files;
    • ADJWT4: adjusted sampling weight;
    • STRATUM4 and PAIR4: variance stratum and pair member which can be used for calculating variances;
    • REGION4 and CENDIV4: Census region and division;
    • SQFT4 and SQFTC4: square footage, both exact and category;
    • PBA4: principal building activity;
    • YRCONC4: year constructed category; and
    • ELSUPL4, NGSUPL4, FKSUPL4, STSUPL4, HWSUPL4: a set of variables indicating whether electricity, natural gas, fuel oil, district steam or district hot water were used in the building.

    For each group of variables, there are two items: a layout file and a data file. The layout file is a text file which gives, for each variable on a file: the variable name, a description, the position on the file, and the corresponding format. The data file is a comma separated value file.

    To determine what the different values for each variable represent, use the text file provided of all the format codes (alternatively, there is also a PDF document containing all the layout files and format codes). The formats are arranged in alphabetical order and are written so that they may be easily turned into a SAS format library.

    Each of these 15 files can be used by itself or be merged with other files for more complex analyses. By merging files together, a new file can be created that contains, for each respondent, variables from two or more files. The variable BLDGID4 should be used to link the files.

    Examples
    To find the national estimate for... Do this... And you should get...
    Total number of buildings Sum ADJWT4

    4,527,655.66 (or 4,528 thousand)

    Total number of office buildings Sum ADJWT4 for cases where PBA4="02" 679,038.55 (or 679 thousand)
    Total floorspace Create a new variable (weighted square footage) by multiplying ADJWT4 by SQFT4 for each case, then sum this new variable 63,534,072,116.37 (or 63,534 million square feet)
    Total floorspace in buildings with air conditioning Sum the new weighted square footage variable (see above) for cases where COOL4="1" 52,065,435,213.74 (or 52,065 milion square feet)
    Total electricity consumption in thousand Btu Create a new variable (weighted electricty consumption) by multiplying ADJWT4 by ELBTU4 for each case, then sum this new variable 2,773,064,703,837.93 (or 2,773 trillion Btu)

    The CBECS sample was designed so that survey responses could be used to estimate characteristics of the entire commercial buildings stock nationwide. All published CBECS tables report national estimates.

    In order to arrive at national estimates from the CBECS sample, base sampling weights were calculated for each building (these are the reciprocal of the probability of that building being selected into the sample). Therefore, a building with a base weight of 1,000 represents itself and 999 similar, but unsampled buildings in the total building stock. The base weight is further adjusted to account for nonresponse bias. The variable ADJWT4 in the data file is the final weight. In order to obtain a national estimate, each sample building's value must be multiplied by the building's weight.

    Imputation Flags or Z variables

    Files 13 through 15 contain variables that begin with the letter Z. These "Z variables" are also referred to as "imputation flags." Imputation is a statistical procedure used to fill in values for missing items. Missing values for many, but not all, of the variables were imputed in 1989. The imputation flag indicates whether the corresponding non-Z variable was reported, imputed, or inapplicable. There are no corresponding "Z variables" for variables from the CBECS questionnaire which were not imputed, variables where there was no missing data, and variables which were derived based on other variables.

    Confidentiality of Survey Respondents

    Several variables have been modified to protect the confidentiality of respondents. This note escribes the procedures used.

    Square Footage: The numeric square footage (Question B-1) has been modified in two ways, depending on the size of the building. For buildings over one million square feet, the numeric square footage has been replaced with the weighted average square footage of all responding buildings over one million square feet. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South, and West). For buildings one million square feet or less, the numeric square footage has been rounded to within ten percent of the upper limit of the buildings' square footage categories (Question B-2). However, if the rounded value fell below the lower limit of the category, the value was coded at this lower limit. For example, buildings in the range 5,001 to 10,000 square feet were rounded to the nearest 1,000 square feet (except that buildings rounding to 5,000 were coded as 5,001).

    Number of Workers: For buildings where the numeric number of workers (Question E-11) was 5,000 or more, the reported numeric number of workers has been replaced with the weighted average number of workers of all responding buildings with 5,000 or more workers. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South, and West).

    Number of Floors: The upper range of the number of floors (Question F-4) has been replaced with two categories: 15 to 25 floors (coded as 994 on the file) and over 25 floors (coded as 995 on the file).

    Special Measures of Occupancy: Five special measures of occupancy are included in the 1989 CBECS. They are classroom seating capacity (B-7m) for education buildings, seating capacity (B-7n) for food service buildings, licensed bed capacity (B-7o) for in-patient health care buildings, licensed bed capacity (B-7p) for skilled nursing buildings, and number of guest rooms (B-7q) for lodging buildings. These measures were each rounded in the following fashion:

    Fewer than 25 units
    25 to 49 units
    50 to 99 units
    100 to 249 units
    250 to 499 units
    500 to 999 units
    1,000 to 2,499 units
    2,500 to 4,999 units
    5,000 or more units

    no rounding performed
    rounded to nearest 5
    rounded to nearest 10
    rounded to nearest 25
    rounded to nearest 50
    rounded to nearest 100
    rounded to nearest 250
    rounded to nearest 500
    rounded to nearest 1,000

    Weather Variables: Heating and cooling degree-days which have a base 65 degrees Fahrenheit (F) are included on the data files. Also included on the data file are the annual mean and standard deviation of daily average temperatures. These can be used to compute approximate degree-days at any base temperature of interest, using a Gaussian (normal) approximation to the distribution of daily average temperatures.

    The heating degree-day variable has been inflated or deflated by a random percentage, normally distributed with mean zero and standard deviation 2.0. The mean and standard deviation of temperature and the base 65 degrees Fahrenheit (F) cooling degree-days have been modified to be consistent with the modified heating degree-days.

  • Survey Forms

1986

  • Building Characteristics Tables
  • Consumption and Expenditures Tables
  • Microdata
  • Released: January 2009

    The 1986 CBECS Public Use Files are comma separated value (.csv) files that each contain 6,072 records. They represent commercial buildings from the 50 States and the District of Columbia. Each record corresponds to a single responding, in-scope sampled building, and contains information for that building such as building size, year constructed, type of energy used, energy-using equipment, and conservation features.

    The smallest level of geographic detail available is the Census division, of which there are nine in the U.S. No state level data are available.

    1986 CBECS Building Characteristics layout files data files revised date
    File 1: Summary File TXT CSV 2/09
    File 2: Building Activity TXT CSV 2/09
    File 3: Operating Hours TXT CSV 2/09
    File 4: Building Shell, Equipment, Energy Audits, and "Other" Conservation Features TXT CSV 2/09
    File 5: End Uses of Major Energy Sources TXT CSV 2/09
    File 6: End Uses of Minor Energy Sources TXT CSV 2/09
    File 7: HVAC, Lighting and Building Shell Conservation Features
    TXT CSV 2/09
    File 8: Electricity TXT CSV 2/09
    File 9: Natural Gas and Fuel Oil TXT CSV 2/09
    File 10: District Steam and Hot Water TXT CSV 2/09
    File 11: Propane and District Chilled Water
    TXT CSV 2/09
    File 12: Imputation Flags for Summary Data, Building Activity, Operating Hours, Shell and Equipment TXT CSV 2/09
    File 13: Imputation Flags for Energy Audits, "Other" Conservation Features and End Uses
    TXT CSV 2/09
    File 14: Imputation Flags for HVAC, Lighting, and Shell Conservation Features TXT CSV 2/09
    All Format Codes TXT 2/09
    All Layout Files and Format Codes PDF 2/09
    Questionnaire PDF 2/09

    File Organization

    Because of the size of the CBECS questionnaire, the variables were separated into groups by subject matter. These 14 smaller files make it easier to manipulate the data.

    Several variables are frequently used in the analysis of commercial energy data. These core variables are included in each group of variables:

    • BLDGID3: building identifier, which is the link between files;
    • ADJWT3: adjusted sampling weight;
    • STRATUM3 and PAIR3: variance stratum and pair member which can be used for calculating variances;
    • REGION3 and CENDIV3: Census region and division;
    • SQFTC3: square footage category;
    • PBA3: principal building activity;
    • YRCONC3: year constructed category; and
    • ELSUPL3, NGSUPL3, FKSUPL3, PRSUPL3, STSUPL3, HWSUPL3: a set of variables indicating whether electricity, natural gas, fuel oil, propane, district steam or district hot water were used in the building.

    For each group of variables, there are two items: a layout file and a data file. The layout file is a text file which gives, for each variable on a file: the variable name, a description, the position on the file, and the corresponding format. The data file is a comma separated value file.

    To determine what the different values for each variable represent, use the text file provided of all the format codes (alternatively, there is also a PDF document containing all the layout files and format codes). The formats are arranged in alphabetical order and are written so that they may be easily turned into a SAS format library.

    Each of these 14 files can be used by itself or be merged with other files for more complex analyses. By merging files together, a new file can be created that contains, for each respondent, variables from two or more files. The variable BLDGID3 should be used to link the files.

    Examples
    To find the national estimate for... Do this... And you should get...
    Total number of buildings Sum ADJWT3 4,154,007.27 (or 4,154 thousand)

    Total number of office buildings Sum ADJWT3 for cases where PBA3="02" 613,910.08 (or 614 thousand)
    Total floorspace Create a new variable (weighted square footage) by multiplying ADJWT3 by SQFT3 for each case, then sum this new variable 58,488,004,765.12 (or 58,488 million square feet)
    Total floorspace in buildings with air conditioning Sum the new weighted square footage variable (see above) for cases where COOL3="1" 46,347,206.477.23 (or 46,347 milion square feet)
    Total electricity consumption in thousand Btu Create a new variable (weighted electricity consumption) by multiplying ADJWT3 by ELBTU3 for each case, then sum this new variable 2,390,382,715,844.76 (or 2,390 trillion Btu)

    The CBECS sample was designed so that survey responses could be used to estimate characteristics of the entire commercial buildings stock nationwide. All published CBECS tables report national estimates.

    In order to arrive at national estimates from the CBECS sample, base sampling weights were calculated for each building (these are the reciprocal of the probability of that building being selected into the sample). Therefore, a building with a base weight of 1,000 represents itself and 999 similar, but unsampled buildings in the total building stock. The base weight is further adjusted to account for nonresponse bias. The variable ADJWT3 in the data file is the final weight. In order to obtain a national estimate, each sample building's value must be multiplied by the building's weight.

    Imputation Flags or Z variables

    Files 12 through 14 contain variables that begin with the letter Z. These "Z variables" are also referred to as "imputation flags." Imputation is a statistical procedure used to fill in values for missing items. Missing values for many, but not all, of the variables were imputed in 1986. The imputation flag indicates whether the corresponding non-Z variable was reported, imputed, or inapplicable. There are no corresponding "Z variables" for variables from the CBECS questionnaire which were not imputed, variables where there was no missing data, and variables which were derived based on other variables.

    Confidentiality of Survey Respondents

    Several variables have been modified to protect the confidentiality of respondents. This note describes the procedures used.

    Square Footage: The numeric square footage (Question B-1) has been modified in two ways, depending on the size of the building. For buildings over one million square feet, the numeric square footage has been replaced with the weighted average square footage of all responding buildings over one million square feet. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South, and West). For buildings one million square feet or less, the numeric square footage has been rounded to within ten percent of the upper limit of the buildings' square footage categories (Question B-2). However, if the rounded value fell below the lower limit of the category, the value was coded at this lower limit. For example, buildings in the range 5,001 to 10,000 square feet were rounded to the nearest 1,000 square feet (except that buildings rounding to 5,000 were coded as 5,001).

    Number of Workers: For buildings where the numeric number of workers (Question C-3) was 5,000 or more, the reported numeric number of workers has been replaced with the weighted average number of workers of all responding buildings with 5,000 or more workers. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South, and West).

    Number of Floors: The upper range of the number of floors (Question D-3) has been replaced with two categories: 15 to 25 floors (coded as 15 on the file) and over 25 floors (coded as 26 on the file).

    Special Measures of Occupancy: Five special measures of occupancy are included in the 1986 CBECS. They are classroom seating capacity (B-10m) for education buildings, seating capacity (B-10n) for food service buildings, licensed bed capacity (B-10o) for in-patient health care buildings, licensed bed capacity (B-10p) for skilled nursing buildings, and number of guest rooms (B10q) for lodging buildings. These measures were each rounded in the following fashion:

    Fewer than 25 units
    25 to 49 units
    50 to 99 units
    100 to 249 units
    250 to 499 units
    500 to 999 units
    1,000 to 2,499 units
    2,500 to 4,999 units
    5,000 or more units

    no rounding performed
    rounded to nearest 5
    rounded to nearest 10
    rounded to nearest 25
    rounded to nearest 50
    rounded to nearest 100
    rounded to nearest 250
    rounded to nearest 500
    rounded to nearest 1,000

    Weather Variables: Heating and cooling degree-days which have a base 65 degrees Fahrenheit (F) are included on the data files. Also included on the data file are the annual mean and standard deviation of daily average temperatures. These can be used to compute approximate degree-days at any base temperature of interest, using a Gaussian (normal) approximation to the distribution of daily average temperatures.

    The heating degree-day variable has been inflated or deflated by a random percentage, normally distributed with mean zero and standard deviation 2.0. The mean and standard deviation of temperature and the base 65 degrees Fahrenheit (F) cooling degree-days have been modified to be consistent with the modified heating degree-days.

  • Survey Forms

1983

  • Building Characteristics Tables
  • Consumption and Expenditures Tables
  • Microdata
  • Released: July 2010

    The 1983 Commercial Buildings Energy Consumption Survey (CBECS) Public Use Files are comma separated value (.csv) files where each record corresponds to a single responding, in-scope sampled building. Each of the 7,140 records contains information for that building such as building size, year constructed, type of energy used, energy-using equipment and conservation features.

    The microdata files represent commercial buildings from the contiguous United States and the District of Columbia. The smallest level of geographic detail available is the Census region; no Census division or state level data are available.

    Both the 1979 and 1983 CBECS included buildings that were 1,000 square feet or less and included predominantly residential or industrial buildings if commercial activity was present. Beginning in 1986, CBECS only includes buildings larger than 1,000 square feet that are predominantly commercial.

    1983 CBECS Building Characteristics, Consumption and Expenditures
    layout files data files revised date
    File 1: Building Characteristics TXT CSV 6/10
    File 2: Imputation Flags for Building Characteristics TXT CSV 6/10
    File 3: Consumption and Expenditures TXT CSV 6/10
    File 4: Longitudinal Variables (transferred from 1979 survey) TXT CSV 6/10
    All Format Codes TXT 6/10
    All Layout Files and Format Codes TXT 6/10
    Questionnaire (questionnaire begins on page 219) PDF 6/10
    File Organization

    Because of the size of the CBECS questionnaire, the master file was divided into 4 files to make it easier to manipulate the data. The 1983 sample included all buildings that had been selected for the 1979 CBECS, as well as a sample of new buildings constructed since the 1979 CBECS. File 4 includes the 1979 values for buildings that participated in both the 1979 and 1983 CBECS. The 4 files are grouped by subject matter and each contains core variables that are particularly useful to analyze the data. The core variables are the following:

  • BLDGID2: building identifier, which is used to link files;
  • XSECWT2: cross-sectionally adjusted sampling weight (use LONGWT2 to estimate changes between 1979 and 1983);
  • STR402 and PAIR402: variance stratum and pair member which can be used for calculating variances;
  • REGION2: Census region;
  • SQFTC2: square footage category;
  • BCWM2C: principal building activity;
  • YRCONC2: year constructed category; and
  • ELSUPL2N, NGSUPL2N, FKSUPL2N, PRSUPL2N and STSUPL2N: a set of variables indicating whether electricity, natural gas, fuel oil, propane or district steam were used in the building.
  • For each group of variables, two items are provided: a layout file and a data file. The layout file is a text file that gives, for each variable on a file: the variable name, a description, and the corresponding format. The data file is a comma separated value file.

    To determine what the different values for each variable represent, use the text file "All Format Codes" (alternatively, use the text file "All Layout Files and Format Codes"). The formats are arranged in alphabetical order and are written so that they may be easily turned into a SAS format library.

    Each of the 4 data files can be used by itself or merged with other files for more complex analyses. By merging files, a new file can be created that contains, for each respondent, variables from two or more files. The variable BLDGID2 is used to link the files.

    The CBECS sample was designed so that survey responses could be used to estimate characteristics of the entire commercial buildings stock nationwide. All published CBECS tables report national estimates.

    In order to arrive at national estimates from the CBECS sample, base sampling weights were calculated for each building (these are the reciprocal of the probability of that building being selected into the sample). Therefore, a building with a base weight of 1,000 represents itself and 999 similar, but unsampled buildings in the total building stock. The base weight is further adjusted to account for nonresponse bias. The variables XSECWT2 and LONGWT2 in the data file are the final weight for cross-sectional and longitudinal estimates, respectively. In order to obtain a national estimate, each sample building's value must be multiplied by the building's weight.

    Examples

    To find the national estimate for... Do this... And you should get...
    Total number of buildings Sum XSECWT2 4,243,680.04 (or 4,244 thousand) buildings
    Total number of office buildings Sum XSECWT2 for cases where BCWM2C="11" 575,078.78 (or 575 thousand) office buildings
    Total floorspace Create a new variable (weighted square footage) by multiplying XSECWT2 by SQFT2 for each case, then sum this new variable 60,747,713,493.92 (or 60,784 million) square feet
    Total electricity consumption in thousand Btu Create a new variable (weighted electricity consumption) by multiplying XSECWT2 by ELBTU2 for each case, then sum this new variable 3,372,422,042,097.23 (or 3,372 trillion) Btu
    Total number of buildings discontinuing use of fuel oil or kerosene between 1979 and 1983 Create a new variable indicating fuel oil or kerosene use in 1979 (FKUSED1T=’1”) but not in 1983 (FKSUPL2N=’2’), multiplied by LONGWT2 for each case, then sum.

    320,326.63 (or 320 thousand) buildings

    Total number of commercial buildings (excluding agricultural, industrial, and residential buildings) Sum XSECWT2 for cases where BCWM2C does not equal 1, 7, or 13 3,711,810.10 (or 3,712 thousand) commercial buildings

    Imputation Flags

    File 2 contains "imputation flags," variables that begin with ”IMP.” Imputation is a statistical procedure used to supply values for missing items. Missing values for many, but not all, of the variables were imputed. The imputation flag indicates whether the corresponding data item was reported, imputed, or inapplicable. There are no corresponding imputation flags for variables from the CBECS questionnaire which were not imputed, variables where there was no missing data, and variables which were derived based on other variables.

    Confidentiality of Survey Respondents

    Several variables have been modified to protect the confidentiality of respondents. This section describes the procedures used.

    Square Footage: The reported value of square footage has been modified in two ways, depending on the size of the building. For buildings over one million square feet, the numeric square footage has been replaced with the weighted average square footage of all responding buildings over one million square feet. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South and West). For buildings one million square feet or less, the numeric square footage has been rounded to within ten percent of the upper limit of the buildings' square footage categories. However, if the rounded value fell below the lower limit of the category, the value was coded at this lower limit. For example, buildings in the range 5,001 to 10,000 square feet were rounded to the nearest 1,000 square feet (except that buildings rounding to 5,000 were coded as 5,001).

    Number of Workers: For buildings where the numeric number of workers was 5,000 or more, the reported numeric number of workers has been replaced with the weighted average number of workers of all responding buildings with 5,000 or more workers. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South and West).

    Number of Floors: The upper range of the number of floors has been replaced with two categories: 15 to 25 floors (coded as 15 on the file) and over 25 floors (coded as 26 on the file).

    Weather Variables: Heating and cooling degree-days which have a base 65 degrees Fahrenheit are included on the data files. The degree-day variables have been inflated or deflated by a random percentage, normally distributed with mean zero and standard deviation 2.0.

  • Survey Forms

1979

  • Building Characteristics Tables:
  • Consumption and Expenditures Tables:
  • Microdata
  • Released: July 2010

    The 1979 Commercial Buildings Energy Consumption Survey (CBECS) Public Use Files are comma separated value (.csv) files where each record corresponds to a single responding, in-scope sampled building. Each of the 6,221 records contains information for that building such as building size, year constructed, type of energy used, energy-using equipment and conservation features.

    The microdata files represent commercial buildings from the contiguous United States and the District of Columbia. The smallest level of geographic detail available is the Census region; no Census division or state level data are available.

    Both the 1979 and 1983 CBECS included buildings that were 1,000 square feet or less and included predominantly residential or industrial buildings if commercial activity was present. Beginning in 1986, CBECS only includes buildings larger than 1,000 square feet that are predominantly commercial.

    1979 CBECS Building Characteristics, Consumption and Expenditures
    layout files data files revised date
    File 1: Building Characteristics TXT CSV 6/10
    File 2: Imputation Flags for Building Characteristics TXT CSV 6/10
    File 3: Consumption and Expenditures TXT CSV 6/10
    All Format Codes (text file) TXT 6/10
    Questionnaire (questionnaire begins on page 289) PDF 6/10
    File Organization

    Because of the size of the CBECS questionnaire, the master file was divided into 3 files to make it easier to manipulate the data. The 3 files are grouped by subject matter and each contains core variables that are particularly useful to analyze the data. The core variables are the following:

  • BLDGID1: building identifier, which is used to link files;
  • XSECWT1: adjusted sampling weight;
  • STRATUM1 and PAIR1: variance stratum and pair member which can be used for calculating variances;
  • REGION1: Census region;
  • SQFTC1: square footage category;
  • BCWM1: principal building activity;
  • YRCONC1: year constructed category; and
  • ELSUPL1, NGSUPL1, FKSUPL1, PRSUPL1 and STSUPL1: a set of variables that indicate whether electricity, natural gas, fuel oil, propane or district steam were used in the building.
  • For each group of variables, two items are provided: a layout file and a data file. The layout file is a text file that lists, for each variable on the file: the variable name, a description, the position on the file and the file format. The data file is a comma separated value file that can be opened in Excel.

    To determine what the different values for each variable represent, use the text file "All Layout Files and Format Codes." The formats are arranged in alphabetical order and are written so that they may be easily turned into a SAS format library.

    Each of the 3 data files can be used by itself or merged with the other files for more complex analyses. The variable BLDGID1 is used to link the merged files.

    The CBECS sample was designed so that survey responses could be used to estimate characteristics of the entire commercial buildings stock nationwide. All published CBECS tables report national estimates. To provide national estimates from the CBECS sample, base sampling weights were calculated for each building (the reciprocal of the probability of that building being selected into the sample). Therefore, a building with a base weight of 1,000 represents itself and 999 similar, but unsampled, buildings in the total building stock. The base weight is further adjusted to account for nonresponse bias. The variable XSECWT1 in the data file is the final weight. In order to obtain a national estimate, each sample building's value must be multiplied by the building's weight.

    Examples

    To find the national estimate for... Do this... And you should get...
    Total number of buildings Sum XSECWT1 4,267,180.41 (or 4,267 thousand) buildings
    Total number of office buildings Sum XSECWT1 for cases where BCWM1="11" 544,954.01 (or 545 thousand) office buildings
    Total floorspace Create a new variable (weighted square footage) by multiplying XSECWT1 by SQFT1 for each case, then sum this new variable 54,636,363,945.58 (or 54,636 million) square feet
    Total floorspace in buildings with air conditioning Sum the new weighted square footage variable (see above) for cases where COOL1="1"

    43,215,194,087.24 (or 43,215 million) square feet

    Total electricity consumption in thousand Btu Create a new variable (weighted electricity consumption) by multiplying XSECWT1 by ELBTU1 for each case, then sum this new variable

    2,961,324,271,169.81 (or 2,961 trillion) Btu

    Imputation Flags

    File 2 contains "imputation flags," variables that begin with "IMP." Imputation is a statistical procedure used to supply values for missing items. Missing values for many, but not all, of the variables were imputed. The imputation flag indicates whether the corresponding data item was reported, imputed, or inapplicable. There are no corresponding imputation flags for variables from the CBECS questionnaire that were not imputed, variables where there was no missing data, and variables that were derived based on other variables.

    Confidentiality of Survey Respondents

    Several variables have been modified to protect the confidentiality of respondents. This section describes the procedures used.

    Square Footage: The reported value of square footage has been modified in two ways, depending on the size of the building. For buildings over one million square feet, the numeric square footage has been replaced with the weighted average square footage of all responding buildings over one million square feet. Separate weighted means were calculated for each of the four Census regions (Northeast, Midwest, South and West). For buildings one million square feet or less, the numeric square footage has been rounded to within ten percent of the upper limit of the buildings' square footage categories. However, if the rounded value fell below the lower limit of the category, the value was coded at this lower limit. For example, buildings in the range 5,001 to 10,000 square feet were rounded to the nearest 1,000 square feet (except that buildings rounding to 5,000 were coded as 5,001).

    Number of Workers: For buildings where the numeric number of workers was 5,000 or more, the reported numeric number of workers has been replaced with the weighted average number of workers of all responding buildings with 5,000 or more workers. Separate weighted means were calculated for each of the four Census regions.

    Number of Floors: The upper range of the number of floors has been replaced with two categories: 15 to 25 floors (coded as 15 on the file) and over 25 floors (coded as 26 on the file).

    Weather Variables: Heating and cooling degree-days which have a base 65 degrees Fahrenheit are included on the data files. The degree-day variables have been inflated or deflated by a random percentage, normally distributed with mean zero and standard deviation 2.0.

  • Survey Forms