Can I use Ordinal Regression for Rank and Nominal data ?

Hi
I am need to design an experiment for my research work, which involves
ranking of a product finish(plastic panels).

Each of the Panels has 6 measurable variables associated (say
thickness, shininess, color, gloss, etc., which can be MEASURED by an
instrument) In a group of 100 panels I have selected only 18 panels
which have low, medium and high value of each Variable.
----------------------------------------------------
The above 18 panels are evaluated i.e, just RANKED by people based on
what they like (Ranking based on the overall Finish ONLY, so they are
not aware of the SIX variables, which could be measured its a
numerical data)

The people who are ranking the plastic panel are focusing on the
Product FINISH only. They Rank from 1 to 18 ( 1 = BEST and 18 =
WORST )

I have Y = Rank data , X1 = Thickness, X2 = Shininess, X3 = Color
etc.,
I am trying to correlate the Rank data with the a measurable data.

*************************************************************************************************
Median
(Rank) X 1 X 2 X 3
X4 X5 X 6

Panel 1. 4 22.5 38.5 41.8
27.5 22.7 11.8
Panel 2. 13 10 25.4 35
15.8 45 37.2
Panel 3. 9 15 17.5 22.4
54 23.8 22.4
Panel 4. 6.5
Panel 5. 3.7
..
.
..

Panel 20.
***************************************************************************************************
Note : ( X1 to X6 are Normally distributed and Median Rank - Ordinal
variable)

My aim is to determine which among the six variable is more
influential/ related to RANK and develop a STANDARD for this
instrument.

The Rank data does not follow a normal distribution even after
transformation.

I did calculate the Kendall Coefficient (W = 0.102).

Can I use the ORDINAL REGRESSION for this kind of problem ?

Can you please suggest me some solution for how approach this very
Tricky problem.
.

Relevant Pages