"Options Analysis Tutorial and Help Guide" contains a section: "Using Options Analysis - Analysis Output" which includes the following:
"Alternative appeal data
An additional special set of columns will be found in the FundamentalsDat data sheet produced via a Fundamentals analysis. These columns are headed 'Alternative appeal data'. Each column contains the highest appeal amongst options other than option at the head of the column. Each column is a surrogate variable for all the other options included in the analysis.
This can be used to reduce the number of options in further analysis to two – the option of interest and the 'ALT' option. Or it can be used with a sub-set of options to be reduced (of, say, minor substantive interest) plus another option to create the 'ALT dummy', so that these “others” can be subsequently dealt with collectively, within a larger set of options. To generate these ALT options for these purposes you should use validated data without casewise adjustment. You can apply casewise adjustment when using the data in further analysis.
Using a reduced number of options will both simplify analysis and speed execution but ignore the options' individual option roles and influences. For some purposes the reduced analysis may be preferable, especially where the principle focus is a single option and 'the rest', and, extensive simulation is envisaged."
Generally, consolidation of options is not a preferred approach.
However, when the options in a selection set exceed the maximum number of options able to be accommodated in a single analysis, consolidation of sufficient options can allow the analysis to proceed.
Consolidation would commonly be best applied only to minor options and only to an extent sufficient to make analysis tractable. (An alternative or supplementary basis for consolidation may be similarity of options.)
Consolidation of options within a set is likely to have only minor direct effects on the shares, advantage or disadvantage of unconsolidated options. (Changes are generated when more than one consolidated option shares an equal highest appeal rating with a unconsolidated option.)
Obviously, there is no effect on the appeal parameters for unadjusted input appeal data of unconsolidated options. However, consolidation of options will effect adjusted input parameters for appeal of all options in a selection set.
Likewise, all analysis involving adjustment or changes to parameters, whether involving adjusted or unadjusted input data, will yield differing results from analyses of corresponding selections sets of unconsolidated options.
The substantial differences in results between sets with consolidated and unconsolidated options will depend mainly on:
Where consolidated options are used, it will be highly desirable to limit comparisons between analyses to the set involving the consolidation.
Consolidation involves the creation of a hypothetical option representing only the most attractive option for each individual amongst those in a consolidation set. Data for each individual on all the other options is discarded. If 10 options are consolidated, only one appeal value per respondent will be retained. Unsurprisingly, the parameters for the consolidated option can be very different from those of the options from which it is composed.
The truly hypothetical character of the consolidated option will be appreciated when it is realised that mutually contradictory appeals may be contained within it! So long as its role is residual this may not matter. But if its role is larger, attempts at interpretation of the consolidated option character may be overwhelmed by complexity and contradiction. (On the other hand, retention of unconsolidated options enables closer identification of sources of option advantage.)