From Opasnet
Jump to: navigation, search

<section begin=glossary />

Index is a set of locations that logically belong together based on their similar nature. For example, "time" is an index with locations like "2007" and "9.6.2008". Other typical indices include year, sex, diagnosis, and age. Indices are used to operationalise variables and assessments. In Opasnet Base, indices are used to explain results stored in the database. Indices and locations are also used to connect cells with the same locations from different variables.

<section end=glossary />

There are different kinds of indices. Their differences and use are described in the following example ovariable.

Age Sex Policy Income
34 Male BAU 2467
36 Male BAU 445
57 Male BAU 1305
25 Female BAU 2228
52 Female BAU 1280
75 Female BAU 1298
34 Male Tax deduction 2727
36 Male Tax deduction 3337
57 Male Tax deduction 812
25 Female Tax deduction 1488
52 Female Tax deduction 1393
75 Female Tax deduction 283
  • Non-marginal (joint) index: This is just a description of the observation row. However, an ovariable can be meaningfully merged by it with another ovariable. For example, Age describes the population age distribution within the population that is defined by the other indices (in this case, separately in males and females).
  • Marginal index: Defines the result (income in this case) distribution for the particular subpopulation. The number of male and female rows does NOT reflect the actual proportions of males and females in the whole population. Instead, the assumption is that there is an equal amount of rows for both.
  • Marginal index of scenarios: These are marginal indices as well, but non-scenario locations can exist simultaneously. In contrast, the Policy index here describes two alternative scenarios that are mutually exclusive. In this case, the tax deduction is either implemented or not (BAU). As you can see from other indices, they have the same rows twice, once for each scenario.

Allowed and non-allowed operations:

  • You can sum up across a joint index to get the total of the population.
  • You can sum up across a marginal index with caution: you have to understand that each location has equal weight, irrespective of its abundance in the population.
  • You cannot sum up across a marginal index of scenarios. (You would get a value that is N times too large, where is N is the number of locations.)
  • You can sample a marginal index with CollapseMarginal. This will result in non-marginal (joint) index. This can be done for both scenarios and non-scenarios.


Universal object, Opasnet Base, dimension, location

See also