The minimum sample size depends on your target market. The larger your target market, the larger your sample should be for statistically significant data. The general rule of thumb for Conjoint Analysis is usually a minimum of 200-300 completed surveys. This, however you can go down to 100 completed surveys if your target market is relatively small.
Please refer ti the following references for further info
According to Tang (2006) sample size recommendations are mostly based on two following approaches: relying on past experience with similar studies and general rules of the thumb or generating synthetic datasets and checking for sample errors of our part-worth estimates.
The probably most known rule of a thumb to estimate necessary sample size for a choice-based conjoint study (Orme, 1998) assumes that:
– having respondents complete more tasks is approximately as good as having more respondents,
– with increasing number of attributes number of parameters to be estimated grows but information that is gained in each task grows at the same rate.
Sorry guys, but the first question should have been, "What sort of conjoint analysis are you using"? because Choice-based conjoint only gets a partial answer from each respondent and thus requires a bigger sample. Second, required sample size does not increase linearly with the size of the target population to be sampled but with the square root. After a population of 2,000, the required sample size stabilizes below 400 for reasonable confidence.
Finally, the standard error, the desired confidence and confidence interval all enter the calculations for generalizability to a population.
John Ireland Thanks for the details. Please also suggest a book to study conjoint analysis & design types in depth.
Also, if you can briefly answer here: how to choose between full, factorial & orthogonal design? [I'm curious to understand how they affect the choice of products]
What should be the minimum number of cards to be generated?
Nisha Arora Thanks for the question. The easiest one to answer is full, factorial & orthogonal design. Full, I believe, is the same as factorial meaning that you try all alternatives. This is adequate for very simple products but produces too many alternatives for more complex products (many levels of many attributes). The disadvantage of orthogonal designs is that they do not test interactions. So, the tradeoff is between studying interactions and burdensome tasks.
Sawtooth developed ACA to test very complex products.
The number of cards to be generated depends on the design, the complexity of the product to be studied, the response elicited (choice, ordering, or rating), the population size, confidence and confidence levels required, etc.
I haven't read a book on Conjoint. Any text on multivariate analysis should give an introduction. I started with the very old SPSS explanation (still used), moved on to the Sawtooth systems and manuals (CVA, ACA and CBC) and Sawtooth's wonderful technical papers. See my article, "Just how loyal are Islamic Banking Customers" in the Intl. Journal of Bank Marketing for some references.
If you decide on choice-based conjoint, then Conjoint.ly offers great explanations and free access for academics. The contrast with Sawtooth is tremendous.