Data alignment

This forum is closed, and read-only.
Locked
Gnurkel
PLS Junior User
Posts: 3
Joined: Tue May 20, 2014 2:46 pm
Real name and title: Laura Verkerk

Data alignment

Post by Gnurkel »

Hi all,

I am a Methodology & Statistics in Psychology graduate student, and I am using this program for my internship and thesis.
Tonight I will purchase the new book to guide my thesis well :)

This is not about the .csv files, this all went well. I have also made some nice-looking models, which all worked.

At the moment I am struggling with how to format and use my data in SmartPLS:

I have 15 products, rated by 250 consumers.
These products are rated in 2 different stages (before preparation and after preparation) and an overall rating. So that's 3*15*250 data points.
I also have analytical data (ingredients, about 20) and sensory panel data (about 20). So, 3*15*20*20*20.

So, a lot of data.

I have made a simple model, for which I imported my data like this:
Rows = Products
Columns = Sensory attributes (about 20) for the "Before preparation", Colour (Lab scale) for the "Before preparation" and Consumers' Liking scores for the "Before Preparation".

This gives a dataset with 15 rows ("observations") and, well, 250 + 20 + 3 columns. You can see that it's much. My final goal is to create a holistic model for product experience, with all the data points i mentioned earlier, incorporated.

I am planning to reduce the data points by replacing the original variables by Factor Analysis means (anyone having a good guide for this?), but with the products as observations I can't imagine that this will work.


So, how should I fomat my data for the optimal use in SmartPLS?
I am uncertain if it is possible to use a longitudinal data file, because the program seems to work column-wise.

Suppose: if I use the Persons (250) as rows, I will need to repeat all the product sensory and analytical information x 250, and I am not sure if this will mess the estimation up (you essentially give an input of the data x 250, so maybe effects will be flattened out). Is this possible? Done more often? How do you import your data?

I hope to have informed you enough, and I hope someone can advise me on this.

Thanks in advance, Kind Regards, Laura Verkerk
Locked