Comparability — Argentina [top]

The 1991 sample identifies electricity as a source. The 2001 sample identifies compressed gas -- a category not distinguished in 1991. The literal translation is "gas from tube." It is coded with bottled gas but can be identified separately using the unharmonized source variable for 2001. The 2010 sample identifies Gas from a tank, a cylinder and a canister. It is coded with bottled gas but can be identified separately using the unharmonized source variable for 2010.

The 2010 sample does not include collective households. Its universe has been defined as "private dwellings" for consistency with previous samples, but the data does not include observations under the NIU category.

Comparability — Benin [top]

All censuses use the word "gas", which is assumed to be unspecified petroleum gas.

Comparability — Botswana [top]

In the 1981 and 1991 samples "gas" is unspecified. In 2001 and 2011 "gas" is specifically identified as "Liquefied Petroleum Gas (LPG)". The category "diesel" is included in 2011.

Comparability — Brazil [top]

For 1960-1970, households using improvised or portable stoves were classified as using no cooking fuel; for 1980-1991, such households were asked to indicate a fuel type. Additionally, in 1991 households were asked to indicate the type of cooking fuel available to the household, even if the residents currently occupying the housing unit did not cook.

The 1991 sample has a separate category for "bottled gas and wood" and two categories containing wood products, compared to one category in the earlier samples.

Comparability — Cambodia [top]

Samples are fully comparable. In the 2004 sample, the other or none fuel category is classified under "Other".

Comparability — Cameroon [top]

The 1987 sample identifies "wood, charcoal and sawdust" as a source. It is coded with "wood, coal, and other solid fuels" but it could not be identified separately. The 2005 sample identifies "firewood and charcoal" and "sawdust" as a separate categories. The first one is coded with "wood, coal, and other solid fuels" while the second one is coded with "others" given that it is not a solid fuel.

Comparability — Egypt [top]

The census had separate responses for each type of fuel, so there were possible response combinations in the data. The most frequent combinations have their own codes in FUELCOOK, while the less frequent are coded into a residual category. See the unharmonized source variables for the full information.

The question is concerned with energy source of the household. The universe includes all households in 1996, but is limited to private households in 2006. In the 2006 census the category "butane" was replaced by "propane". Both censuses allow responses for multiple fuel sources. Many are coded to the "other combinations" category, the majority of which represent electricity in combination with other fuels.

Comparability — Ethiopia [top]

In the 1994 sample, there are a different set of response categories for urban and rural households. See the unharmonized source variables for more detail.

In the 2007 census, respondents were able to indicate multiple cooking fuels. Households that indicated more than one cooking fuel are included in the multiple fuels category. Specific combinations of fuel can be identified separately using the unharmonized source variable for 2007.

Comparability — Ghana [top]

The data are mostly comparable across samples. In 2000, the non-wood plant material is coconut husks. In 2010, "Other" includes saw dust and crop residue.

Comparability — Greece [top]

The 2011 sample includes biomass as an option for fuel which is classified as "Other combined organic waste materials".

Comparability — Guatemala [top]

In 1964, unknown includes "does not cook".

Comparability — Haiti [top]

In the 2003 census, the question asked for the different fuel sources for cooking. It includes 7 different sources and all their possible combinations. The different combinations are coded under the multiple fuels category, but can be identified separately using the unharmonized source variable for 2003.

Comparability — Honduras [top]

The 1961 sample only has information for households consisting of a head.

Comparability — Indonesia [top]

The samples are mostly comparable with the following differences in response categories: the 1971 census combines wood and charcoal, the 2005 sample distinguishes coal from charcoal, and the 2010 census includes a response for no cooking fuel used. All censuses use the word "gas", which is assumed to be unspecified petroleum gas.

Comparability — Israel [top]

The sample has a large number of non-responses.

Comparability — Jamaica [top]

In the 1982 census, the question was asked to all households, but to only private households in the subsequent censuses. In the 1982 census "no fuel" and "other" responses were combined together and are coded into "other" in this variable. In the 2008 sample, wood and charcoal are treated as separate categories. The proportion of non-reponses is significantly higher in the 1982 sample than in the succeeding samples.

Comparability — Kenya [top]

The Kenya 2009 sample includes private households (and refugee camps), while the 1989 includes "not group quarters" and the 1999 sample includes all households.

Comparability — Lesotho [top]

The samples are fully comparable.

Comparability — Liberia [top]

The "electric" category in the 2008 sample refers to cooking using electricity from the household’s personal generator or cooking using electricity from a power company as fuel.

Comparability — Malawi [top]

This variable is only available in the 1998 and 2008 samples. In the 1998 sample, the question was asked of households with multiple dwellings, only information of the first dwelling is coded into this variable. The question did not identify households that had no source of cooking fuel in the 2008 census.

Comparability — Mali [top]

The 1998 and 2009 samples identify electricity and petroleum as sources of fuel not previously identified in the 1987 sample. The 1987 and 2009 samples distinguish firewood from coal while the 1998 sample combines wood and charcoal. The 2009 sample includes "solar oven" as a response category which implies cooking using solar energy as fuel.

Comparability — Mexico [top]

The 1960 Mexican sample is a sample of individuals, not households. Only heads of households have most household-level information.

The "gas" category in 1960-1970 includes a small percentage of households using electricity.

The 2010 sample distinguishes between gas from piping/distributed natural gas, and that from a stationary source (e.g. tank). The "other" category in this census also includes those who do not cook in the dwelling.

The 2015 and 2020 samples include an option for households not cooking in the dwelling. Samples before 2010 do not have specific instructions for these cases.

Comparability — Morocco [top]

The 2014 census asked responses for five separate fuel categories that indicate whether a specific fuel was used often, sometimes, or never. In FUELCOOK, if a single fuel is said to be used often (while others are used sometimes or never), then that is considered as the primary source. If multiple fuels are used with similar frequency, then those households are classified as using "other combinations." These criteria are consistent with the concept of "predominant" fuel used in FUELCOOK. The unharmonized source variables preserve all these details.

Comparability — Palestine [top]

The universe differs across samples. In both samples, the distinction is between gas and other sources of energy.

Comparability — Paraguay [top]

In 2002, Kerosene was not included as a response category.

Comparability — Peru [top]

The 2007 census question refers to the most common fuel used for cooking, while in 2017 multiple responses are allowed.

Comparability — Puerto Rico [top]

The 1970 and 1980 samples distinguish piped gas form other gas delivery means, but 1990 does not make this or many other distinctions.

Comparability — Saint Lucia [top]

The identified fuels differ between samples.

Comparability — South Africa [top]

The response categories are fairly consistent over time. The 1996 census distinguishes between electricity direct from municipality, and electricity from another source. The 2016 sample distinguishes between electricity from mains and other sources of electricity in the corresponding unharmonized source variable. The 2011 censes includes a category for households that do not cook.

The universe includes all households in 2001, but is limited to private households in 1996 and 2007. In 2011, the universe includes only private households residing in conventional housing units or converted hostels. Private households who spent census night in a hotel, or other collective dwelling or institution or that were in transit on census night are excluded from the universe. The 2016 sample includes all households.

Comparability — Tanzania [top]

In the 2012 sample, the options "Generators or private sources" and "electricity from wind" are classified as electricity.

Comparability — Thailand [top]

In all samples the question is asked only to private households. The universe is additionally restrictive in 1970 and 1990. In the 1970 sample, urban and sampled rural private households were asked the question, while in 1990 only private households with a kitchen were asked the question.

The 1980 census invited multiple responses while indicating which was the primary fuel. FUELCOOK indicates the primary source, but the various combinations are preserved in the unharmonized source variable.

Comparability — Trinidad and Tobago [top]

The 2000 and 2011 samples include the options for liquefied petroleum gas or cooking gas, as well as "none".

Comparability — Uganda [top]

In 2014, more detailed categories for households using electricity are preserved in the unharmonized source variables

Comparability — United States [top]

Those who had no cooking equipment or ate meals elsewhere were to be classified as using no cooking fuel.

Comparability — Uruguay [top]

The 1975 census does not distinguish between piped gas and liquefied petroleum gas as other samples do. The 2006 sample retains detail on whether the electricity was from the public utility or from a generator in the unharmonized source variable. The 2011 sample includes "super gas" as a response category, which is a form of biogas technology (fuel derived from plants or plant materials).

Comparability — Vietnam [top]

The question was only asked in the 2009 and 2019 census.

Comparability — Zambia [top]

The 2010 Zambia enumeration instructions indicate that candles should not be considered a source of energy for cooking. Nonetheless, they are indicated in the source and integrated variable.