Testing for difference between ranked and weighted data

[slogger]

Hello

I just joined this forum today and I'm very appreciative of those people who selflessly give their time and expertise.

I'm comparing two groups of visitors (Group A and B) to a national park, to see if there is a significant difference between their preferred method of visiting the park.

Respondents scored their relative preference on a 5-point Likert scale for 6 predetermined visit methods. Results were then scored and weighted by overall % popularity between the visit methods.

I now need to analyse the ranked percentage scores between the two groups to see if there is/is not a significant difference between them.

I am confused as to whether a Kruskal-Wallis test or Mann-Whitney test - other test - is appropriate.

I'd really appreciate it if someone can help, please. Thank you.

[slogger]

Dear posters

Re my previous post, I've now satisfied myself, via further reading and checking on various tests, that the Mann-Whitney U test is appropriate for analysing the ranking of data between my two sample groups, and have performed the test accordingly. However, Mann - Whitney does not seem able to test for significant difference based upon the weighting of the my data.

For example: an extract of my data is as follows:

Preferred visit method Group A Group B

With a club 44.4% 5.1%

As a commercial guide 22.2% 54.2%

With friends/family 33.3% 25.4%

As a commercial client 0.0% 15.3%

Alone 0.0% 0.0%

Other 0.0% 0.0%


Is there a test which actually compares relative weighting of responses and not merely response rank?

Many thanks