Intensity scales

summary next

Foreword

In case of missing values in the extracted data, following calculation choices are suggested automatically:


Estimation option is favoured in most of cases.


1. Blocs Incomplets Equilibrés

These protocols, used to carried out product comparisons, (implemented in TASTEL, in product profile) are dedicated to optimise comparisons when assessors cannot evaluate in a same time the study product number.
Also, it is needed to know the total product number to compare, the product number per block, or the product number that will be to assess at the same time.

Then, the system:
  • On the first hand, identifies product comparisons to carry out to allow a balance, and a complete product comparison between them,
  • On the other hand, randomly codifies products so as to avoid recognition biases.

These designs come from works of W.G. COCHRAN and G.M. COX, under publication JOHN WILEY and SONS, INC. (1957) Part.1 - Plans for designs of experiments.

Remark: A data standardisation can be asked so as to eliminate the notation effect that can be important in case of incomplete blocks by the fact that all assessors have not assessed all products, to which case, the system proceeds to the following transformation:

Xn = (X - moyij)/ectij, with:
Moyij: mean of the concernec assessor by attribute
Ectij: standard deviation of the concerned assessor by attribute.


Top