Why conjoint analysis (Session 1, pp. 15-24)
1. In direct surveys, respondents might say they consider all attributes important
a. Not informative
2. Conjoint enforces tradeoffs between attributes
a. All attributes evaluated at once
b. Respondents evaluate “complete” products with both strong and weak attributes
3. Conjoint reduces problem of socially desirable answers
4. Conjoint adds realism
a. In real-life consumers evaluate products, not isolated attributes (do they consciously
know which attributes matter?)
5. Conjoint analysis is straightforward
6. Suitable software is available (Sawtooth)
Types of conjoint studies (Session 1, pp. 41-52)
Ranking-based conjoint: Choose the most-preferred product, then the second most-preferred product,
until the least-preferred product
Rating-based conjoint: Give a score to each product in turn
- Problems of Rating-Based Conjoint:
o Not realistic: In real-life, we buy products rather than rating them
o Not clear whether spread in ratings is due to real preferences or due to response style
(E.g., small spread in example above, weak preferences or cautious answers?)
o Implications for sales levels and market shares are not clear. Sales and shares result from
consumer choices, not ratings. What would be the rating threshold?
Choice-based conjoint: Choice between different variants choose the most preferred product only
We record the choices made by every customer during the n tasks (i.e. choice sets). Because, in every
choice set, a different combination of attribute levels is used, we can derive the effect of different
combinations of attribute levels on choice.
Advantages of choice-based conjoint
Tradeoffs are enforced even more
Realistic: the choice-setting mimics real-life
Accommodates no-choice option
Avoids the need of ad-hoc rules to predict market shares
No subjective scaling
Choice is cognitively less demanding than ratings (Louviere 1994)
Random utility theory (Session 2, pp. 14-21)
Thurstone (1927) A consumer generally chooses the alternative that she likes the most, subject to
constraints such as income and time (the stimulus with the highest utility). Sometimes, they don’t
because of random factors.
Seven steps of conjoint (Session 2, p.26)
1
Eline van de Ven – Conjoint 2018-2019
, 1. Determine the type of study (e.g. rating / choice based)
2. Identify the relevant attributes (which and how many?) Conjoint Design
3. Specify the attributes’ levels (which and how many?)
4. Design questionnaire (which products to include?)
5. Collect data from respondents (which channel, format and layout?)
6. Estimate part worths (evaluate the attributes’ levels) Conjoint Analysis
7. Design market simulators (what if scenarios in hypothetical markets)
Choosing attributes & levels (Session 2, pp. 31-33 & 39-41)
Attributes in conjoint analysis should …
be relevant for the management (discuss with them!)
have varying levels in real-life (4 wheels for a car)
be expected to influence preferences (theory, qualitative research)
be clearly defined and communicable (respondent should understand correctly, e.g., verbal
descriptions, pictures, intro movie)
preferably not exhibit strong correlations (but price, brand name)
Number of attributes
Green & Srinivasan (1990): full-profile conjoint if # attributes ≤ 6
Levels of attributes should be
interesting for the management (discuss with them!)
unambiguous (“low” versus “high” is too imprecise)
separated enough (otherwise too little weight)
realistic (but allowed to be little bit outside current range)
such that no attribute can a priori be expected to be clear winner
Number of levels
Two levels is minimum
In case of linearity, two levels is both sufficient and efficient
In case of nonlinearity (e.g. quadratic), more than two levels are needed
More levels than necessary is inefficient:
o More parameters need to be estimated, and complexity for respondent increases
Equal number of levels:
o Attributes with more levels are found to be more important (Wittink, Krishnamurthi and
Reibstein, 1990)
Number of stimuli and alternatives (Session 3, pp. 10-17)
- How many stimuli (CONCEPTS) to include?
o Advantages of large questionnaire:
More observations per respondent
Increase in quality, as respondents learn how to answer
o Disadvantage of large questionnaire:
Decrease in quality, as respondents get fatigued or bored
o Respondents can complete up to 20 tasks without decrease in quality (Johnson and
Orme, 1996)
2
Eline van de Ven – Conjoint 2018-2019
1. In direct surveys, respondents might say they consider all attributes important
a. Not informative
2. Conjoint enforces tradeoffs between attributes
a. All attributes evaluated at once
b. Respondents evaluate “complete” products with both strong and weak attributes
3. Conjoint reduces problem of socially desirable answers
4. Conjoint adds realism
a. In real-life consumers evaluate products, not isolated attributes (do they consciously
know which attributes matter?)
5. Conjoint analysis is straightforward
6. Suitable software is available (Sawtooth)
Types of conjoint studies (Session 1, pp. 41-52)
Ranking-based conjoint: Choose the most-preferred product, then the second most-preferred product,
until the least-preferred product
Rating-based conjoint: Give a score to each product in turn
- Problems of Rating-Based Conjoint:
o Not realistic: In real-life, we buy products rather than rating them
o Not clear whether spread in ratings is due to real preferences or due to response style
(E.g., small spread in example above, weak preferences or cautious answers?)
o Implications for sales levels and market shares are not clear. Sales and shares result from
consumer choices, not ratings. What would be the rating threshold?
Choice-based conjoint: Choice between different variants choose the most preferred product only
We record the choices made by every customer during the n tasks (i.e. choice sets). Because, in every
choice set, a different combination of attribute levels is used, we can derive the effect of different
combinations of attribute levels on choice.
Advantages of choice-based conjoint
Tradeoffs are enforced even more
Realistic: the choice-setting mimics real-life
Accommodates no-choice option
Avoids the need of ad-hoc rules to predict market shares
No subjective scaling
Choice is cognitively less demanding than ratings (Louviere 1994)
Random utility theory (Session 2, pp. 14-21)
Thurstone (1927) A consumer generally chooses the alternative that she likes the most, subject to
constraints such as income and time (the stimulus with the highest utility). Sometimes, they don’t
because of random factors.
Seven steps of conjoint (Session 2, p.26)
1
Eline van de Ven – Conjoint 2018-2019
, 1. Determine the type of study (e.g. rating / choice based)
2. Identify the relevant attributes (which and how many?) Conjoint Design
3. Specify the attributes’ levels (which and how many?)
4. Design questionnaire (which products to include?)
5. Collect data from respondents (which channel, format and layout?)
6. Estimate part worths (evaluate the attributes’ levels) Conjoint Analysis
7. Design market simulators (what if scenarios in hypothetical markets)
Choosing attributes & levels (Session 2, pp. 31-33 & 39-41)
Attributes in conjoint analysis should …
be relevant for the management (discuss with them!)
have varying levels in real-life (4 wheels for a car)
be expected to influence preferences (theory, qualitative research)
be clearly defined and communicable (respondent should understand correctly, e.g., verbal
descriptions, pictures, intro movie)
preferably not exhibit strong correlations (but price, brand name)
Number of attributes
Green & Srinivasan (1990): full-profile conjoint if # attributes ≤ 6
Levels of attributes should be
interesting for the management (discuss with them!)
unambiguous (“low” versus “high” is too imprecise)
separated enough (otherwise too little weight)
realistic (but allowed to be little bit outside current range)
such that no attribute can a priori be expected to be clear winner
Number of levels
Two levels is minimum
In case of linearity, two levels is both sufficient and efficient
In case of nonlinearity (e.g. quadratic), more than two levels are needed
More levels than necessary is inefficient:
o More parameters need to be estimated, and complexity for respondent increases
Equal number of levels:
o Attributes with more levels are found to be more important (Wittink, Krishnamurthi and
Reibstein, 1990)
Number of stimuli and alternatives (Session 3, pp. 10-17)
- How many stimuli (CONCEPTS) to include?
o Advantages of large questionnaire:
More observations per respondent
Increase in quality, as respondents learn how to answer
o Disadvantage of large questionnaire:
Decrease in quality, as respondents get fatigued or bored
o Respondents can complete up to 20 tasks without decrease in quality (Johnson and
Orme, 1996)
2
Eline van de Ven – Conjoint 2018-2019
