oky i tested 270 students a psychological test Likert type (5) the total score of this test is 320 which is the sum of 8 sub scales, how I can deiced the level of the psychological adjustment for all students whether very high, high, moderate and lowest.

So, if I understand this correctly, each student surveyed is given a composite score from 0 to 320, and you want to subdivide that range into four different classes.

i did compose the total score for all student, but it is matter of how I decide who how my students preform on the test , I mean I want to categorize the total score into 4-5 categories and encoded.

Customer:replied 5 months ago.

like very high adjusted students High adjusted moderated adjusted low adjusted

This sounds like something that would be a judgement call on your part, as the researcher conducting the survey.

Depending on what was measured, and the various subscores, this may be more complicated than just dividing the total score range into equal-width subranges.

I would suspect that a reasonable approach would be to find out how other similar studies may have been done, so that your approach is consistent with comparable studies.

With no other research to use as a guide, you're really kind of on your own as far as designing your experiment. Whichever way you decide to go, though, you want to be very clear about how you are classifying various scores, and why you have chosen that particular breakdown.

As for the subscores, you can use a weighted mean to calculate the composite score for each survey participant. However, you will have to determine what weighting factors to use for each of the eight subscores, based on the data that is used to make up each subscore, and how that relates to the partipant's "adjustment".

You could go a step farther and calculate the standard deviation of the scores.

You could then establish a scale along the following lines:

More than 2 standard deviations above the mean -> very high adjusted students Between the mean and 2 standard deviations above the mean -> High adjusted Between the mean and 2 standard deviations below the mean -> moderated adjusted More than 2 standard deviations below the mean -> low adjusted

However, this is somewhat arbitrary, and a peer reviewer might have criticisms about your methodology.

this the Likert scales (5) the idea is to use the interval and then along 4, or 5 or even 6 I repeat to the next like 4.2-0.8= 3.4 3.4/5*100= 68 68*320/ 100= 217.6 all students between 217.6 and 268.8 are in the high level which is the second

If I have two measurements and i want to find how the third measure mediated the relationships between the two measure what is the test to do that and any condition require?