Your data analysis solution
Name is required.
Email address is required.
Invalid email address
Answer is required.
Exceeding max length of 5KB

MaxDiff standard deviation and tests of significance

By Aidan Mulloy | Jan 20, 2017 04:58PM CET

Hi,

I carried out MaxDiff using Hierarchical Bayes estimation with 730 respondents, following the steps of the tutorial on carrying out MaxDiff. The standard deviation for all the MaxDiff scores (descriptive statistics, showing the mean scores of all the attributes) are rather high and this concerns me. However in the example provided in the tutorial I noticed in the results section that the standard deviations were also pretty high and these weren't commented on. Is it usual for the standard deviation to be high for MaxDiff and should it be an area of big concern? I also noticed tests of significance weren't carried out for the mean scores of the attributes; is this also usual with Max Diff?

Thanking you,
Aidan

Best Answer
By Jean Paul | Jan 23, 2017 10:37PM CET | XLSTAT Agent

Hello Aidan,

Thank you for this very interesting question.
The Standard Deviation will change according to the number of MaxDiff respondents you have. In our online tutorial, we have very few respondents (5), which tends to widen the standard deviation. High standard deviations in your data may arise from the fact that there is no clear preference pattern among your attributes. In this case, the MaxDiff scores are not very reliable.
Can you please give us more details on the p-values you are mentioning? Do you mean comparing MaxDiff mean scores between attributes? Do you have a reference on that to share?

Best,

Jean Paul


Up

0

Down

By Jean Paul | Jan 23, 2017 10:37PM CET | XLSTAT Agent

Hello Aidan,

Thank you for this very interesting question.
The Standard Deviation will change according to the number of MaxDiff respondents you have. In our online tutorial, we have very few respondents (5), which tends to widen the standard deviation. High standard deviations in your data may arise from the fact that there is no clear preference pattern among your attributes. In this case, the MaxDiff scores are not very reliable.
Can you please give us more details on the p-values you are mentioning? Do you mean comparing MaxDiff mean scores between attributes? Do you have a reference on that to share?

Best,

Jean Paul

Post Your Public Answer

Contact Us

1c26995d494fb3061dd0ae8571ffc0a4@xlstat.desk-mail.com
https://cdn.desk.com/
false
desk
Loading
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
about
false
Invalid characters found
/customer/portal/articles/autocomplete
9283