setdefaults page_title: Demo Conjoint autosubmit: n appconfig singlepage: y progressbar: n poweredby: feedback: n showtext: CONJ_INTRO html: <The next few pages will demonstrate a basic choice-based conjoint (CBC) implementation.
Our conjoint-trained project managers regularly implement:
Respondents are tasked with a series of choice sets, a table with three products, in this case computers. The product attributes are produced according to a CBC design that we store in an external database.
This CBC demo is based on a design that would have 12 tasks (choice sets) per respondent, and 300 versions of 12 tasks. A good CBC design will demonstrate near-orthogonality (i.e., avoiding correlation of levels) and will thus collectively test as many attribute level-interactions (or pairings) as possible.
As this is a demo, only 3 of the 12 tasks will be shown. On each new screen, notice how the product descriptions change.
HTML set list: VERSLIST 1. 1 TEST_VERS. Please select a version for testing. type: pulldown required: n systemvar: y optsfrom: VERSLIST showif: $testmode VERS. Version type: radio selectby: weight,counts lock: y optsfrom: VERSLIST {weight: (anyChecked($QTEST_VERS,[id])*100)+1} setvalue question: QRND_TO_TM value: 2 condition: $testmode onload: y RND_TO_TM. Randomize task order in this test? type: radio ap: n systemvar: y showif: $testmode 1. Yes 2. No TASKORD. Task order type: text invisible: y cvalue: order( $QRND_TO_TM == 2 ? '(1..3)' : '[1..3]' ) systemvar: y setlist: TASKLIST series: 1..3 order: $QTASKORD block: <
type: radio
optsfrom: CONJOINT_OPTS
template: conjoint_stacked.tx
template_data: <