Sunita Rai You can consult to Narendra Raj Poudel associated to Public Administration Campus, Balkhu, Kathmandu. He has a good idea regarding this.

Hello Sunita,

Ordinarily, for a 7-point Likert-type scale, you'd expect the response values for a single item to be a set of successive integers (e.g., 1,2,..7). Hence, for a 12-item scale set, the minimum and maximum scores would be expected to be 12x1 = 12 (which matches the value you cite) and 12x7 = 84 (which does not match the value you cite).

Perhaps the authors did something unusual, such as the squared integer value for each item's response scale (1, 4, 9, ... 49). That would be consistent with the 12x49 value. However, there really should be some justification given in the article for why such a scheme was imposed.

Good luck with your work!

When your Likert scale has seven steps you can use it as a normal continuous variable and calculate the mean as you always do. It says " the sum of ratings of all the items was taken" and then divided with n = sample size.

The number of steps in a Likert-scored item does not affect whether it is linear (and thus can use a mean) or merely continuous (and thus allows only for a median).

In particular, a linear variable requires that every step in the continuum of scores is equal, so that the "distance" between 1 and 2 is the same as between 2 and 3, etc. But there is no way to guarantee of this with an ordinal scale, such as that used in Likert-scored items.

This difference is the main reason why multiple-item scales (which can be treated as linear) are created from sets of single, Likert-scored items (which are each ordinal).