The score per latent variable
The score of each latent variable is a weighted index score computed from the customer responses to a small set of required questions per variable with an answering scale 1 to 10 value. The weighting factors are equally distributed. Meaning each question weights the same as we have set the importance and relevance of each question at the same level.
The total score of a variable represent the weighted sum of the set of questions which are transformed into 0 to 100 scale value.
Computing the score is done in three steps:
- Compute the mean of each question
- Compute the 0 to 100 scale value for each question
- Compute the latent variable score
Example: perceived product quality
The survey has four question and you have determined that each question is equaly important > each question weight (100% : 4 =) 25%.
Step one, computing the mean....let us assume:
- The mean of question 1 = 8.31
- The mean of question 2 = 8.23
- The mean of question 3 = 7.88
- The mean of question 4 = 6.98
Step two, the mean values from the data set must first be transformed to the value on a 0 to 100 scale. This is done by:
- subtracting 1 from the mean values (sample approach N-1),
- dividing the result by the scale minus 1 and
- multiplying the whole by 100.
Compute 0 to 100 value score for each question:
- Score question 1 = ((8.31-1) : (10-1)) * 100 = 81.25
- Score question 2 = ((8.23-1) : (10-1)) * 100 = 80.32
- Score question 3 = ((7.88-1) : (10-1)) * 100 = 76.46
- Score question 4 = ((6.98-1) : (10-1)) * 100 = 66.50
Step three, compute the weighted variable score:
- Score value question 1 = 81.25 * 25% = 20.31
- Score value question 2 = 80.32* 25% = 20.08
- Score value question 3 = 76.46 * 25% = 19.12
- Score value question 4 = 66.50 * 25% = 16.62
The score of the perceived product quality is the sum of the above = 76.14