Overview (ORTHOPLAN command)

ORTHOPLAN generates an orthogonal main-effects plan for a full-concept conjoint analysis. ORTHOPLAN can append or replace an existing active dataset, or it can build an active dataset (if one does not already exist). The generated plan can be listed in full-concept profile, or card, format using PLANCARDS. The file that is created by ORTHOPLAN can be used as the plan file for CONJOINT.

Basic specification

  • The basic specification is ORTHOPLAN followed by FACTORS, a variable list, and a value list in parentheses. ORTHOPLAN will generate cases in the active dataset, with each case representing a profile in the conjoint experimental plan and consisting of a new combination of the factor values. By default, the smallest possible orthogonal plan is generated.
  • If you are appending to an existing active dataset that has previously defined values, the FACTORS subcommand is optional.

Subcommand order

  • Subcommands can be named in any order.

Operations

  • ORTHOPLAN builds an active dataset (if one does not already exist) by using the variable and value information on the FACTORS subcommand.
  • When ORTHOPLAN appends to an active dataset and FACTORS is not used, the factor levels (values) must be defined on a previous ORTHOPLAN or VALUE LABELS command.
  • New variables STATUS_ and CARD_ are created and added to the active dataset by ORTHOPLAN if they do not already exist. STATUS_=0 for experimental cases, 1 for holdout cases, and 2 for simulation cases. Holdout cases are judged by the subjects but are not used when CONJOINT estimates utilities. Instead, the cases are used as a check on the validity of the estimated utilities. Simulation cases are entered by the user. They are factor-level combinations that are not rated by the subjects but are estimated by CONJOINT based on the ratings of the experimental cases. CARD_ contains the case identification numbers in the generated plan.
  • Duplication between experimental cases and simulation cases is reported.
  • If a user-entered experimental case (STATUS_=0) is duplicated by ORTHOPLAN, only one copy of the case is kept.
  • Occasionally, ORTHOPLAN may generate duplicate experimental cases. One way to handle these duplicates is to edit or delete them, in which case the plan is no longer orthogonal. Alternatively, you can try running ORTHOPLAN again. With a different seed, ORTHOPLAN might produce a plan without duplicates. See the SEED subcommand on SET for more information about the random seed generator.
  • If you do not specify OUTFILE or REPLACE and there are no open datasets, the procedure creates a new, unnamed dataset. If the active dataset and the ORTHOPLAN command contain the same factors, the factor values from the ORTHOPLAN command are appended to the factor values in the active dataset. If the active dataset is a named dataset, a new dataset is created. If the active dataset is an unnamed dataset, the active dataset is replaced. If the variables in the active dataset do not match factors in the ORTHOPLAN command, an error occurs.
  • The SPLIT FILE and WEIGHT commands are ignored by ORTHOPLAN.

Limitations

  • Missing data are not allowed.
  • A maximum of 9 levels can be specified for each factor.
  • The total number of levels for all factors cannot exceed 90.
  • A maximum of 81 cases can be generated by ORTHOPLAN.