Taking into account that an impact evaluation based on surveys is a quantitative statistical instrument, it is vital to establish the units for this quantification and for analysis of the results. This minimum quantitative unit is called an “analysis unit” or, in the context of sampling, it is also known as a “sampling unit” or “items” for sampling.

The analysis unit is not only essential for sample selection but also for the design of the questionnaire. The analysis unit for LAP impact evaluation could be the “beneficiaries” of the programme, however this option needs at least two analysis units to be defined: the household with a beneficiary unit and the owner or possessor.

Although the owner or possessor can be a member of the household, it is important to distinguish him or her because the characteristics of each are different. For example, certain characteristics of the owner, such as his or her age, gender, occupation, education, etc., can be identified which are very different from those of the household as this unit is a group of members with their own characteristics.

Furthermore, bearing in mind that one of the main objectives of the LAP is to provide legal certainty about land and that the evaluation aims to find out how this generates benefits in terms of value and access to infrastructure and services, we should consider a third analysis unit: the property.

As the initial activities of LAPs concern mapping the cadastre, the highest level of control for establishing the scope of the study (including the definition of the sampling frame), the property will be the most suitable analysis unit. The advantage of property is that the sum of the parts unambiguously gives us the whole without the risk of overlapping, whereas the sum of the households has the problem of overlapping in terms of land occupation, since there may be more than one household on the same property and one household may have more than one property. A further consideration is that not all properties are associated with a household, as in the case of unoccupied land or areas belonging to the state or other type of legal entity. Furthermore, as there may be more than one owner per property and a single individual can be the owner of several properties, the sum of owners can involve overlaps that prevent demarcation of the whole in its entirety. The recommended analysis unit is therefore the property, both for determining the sampling scheme and devising the questionnaire questions and the analysis of the results.

An evaluation aim considered in LAPs is to analyse the effect of RCT processes on productivity, the generation of additional income or food security, particularly in a rural setting. A further relevant analysis unit is consequently the production unit.

It is important to bear in mind primarily that not all properties have production units and that production or commercial units may or may not occupy the whole surface area of the property, as in the case of small businesses (sales of various products including foodstuffs) which may be associated with the home itself.

We therefore have four analysis units:

1. Property
   As the central unit.
2. Owner or possessor
   Individual considered the proprietor or occupant of a property.
3. Household
   It is important to stress that there may be more than one household on a property. A main household should be associated with a single home.
4. Production or commercial unit
   The production or commercial (monetary or non-monetary) unit on the property and which uses production factors within it.

It is important to consider that an analysis of the four types of unit defined can be done in the same survey, but the sampling frame should preferably be based on properties. For cases involving a baseline set up without preliminary information on properties, the home can be used as the analysis unit (named in population censuses); this option implies that unoccupied properties or those dedicated exclusively to production might not be included in the sample.

By combining analysis units but starting from the property, we therefore have the following groups:

Property with respect to the owner or possessor
Property with a single owner
Properties with two or more owners

Properties with respect to households
Properties without households
Properties with a single household
Properties with two or more households

Properties with respect to production units
Properties without any production activity
Property with one production unit
Property with two or more production units