Developing National, Local and/or Specialized Norms for the MSCEIT
Developing Norms for Special Groups
If you want to develop norms for specialized groups, the easiest way to do this is to test within that specialized group, have MHS score the scales, and then report the norms for the group. If you publish norms for a special group, please let us (the authors and/or MHS) know about it so that we can cite your work.
Developing Regional/National Norms
Let's say you want to develop national norms for the MSCEIT. This is of specific value if you are looking at, say, cross-national comparisons.
The first decision you will want to consider is whether to "go it alone" or to develop specific norms in cooperation with MHS (the test publisher). MHS has some experience in, and is interested in supporting national norms for countries around the world. Contact the test authors or the research department at MHS to find out more. If they agree to help you develop local norms, they will be able to contribute considerable expertise to assist you.
The Do-It-Yourself Approach
Let's say you want to develop your own consensus-scoring norms independently. In that case, you would need some of the published booklets from MHS to test people with, and some answer sheets. (The more booklets, the more people you can test in a single setting). You could then hand-enter data from the answer sheets for the 141 items of the MSCEIT into the data file of a program like SPSS. After you obtained about 150 responses, you would run the Frequencies program. That would tell you how many people responded to each of the alternatives for a given item. Let's consider the first item on the MSCEIT. Say that:
10% of the people chose 1
15% of the peopel chose 2
20% of the people chose 3
50% of the people chose 4, and
5% of the people chose 5.
Lets say you named the raw response for item 1, "RR01", and 'CNS01' will be the consensus-scored version. You would write a compute statement that looked something like this:
COMPUTE CNS01 = 0.
IF (RR01 EQ 1) CNS01 = .1
IF (RR01 EQ 2) CNS02 = .15
IF (RR01 EQ 3) CNS03 = .20
IF (RR01 EQ 4) CNS04 = .5
IF (RR01 EQ 5) CNS05 = .05
Note the proportions reflect the above percentages of those who endorsed each alternative. Next, you would repeat this process for every item on the test (incrementing the '01' to '01', etc., item-by-item).
Once you had the item-by-item consensus scores, you would add up (via more compute statements) items on each scale to get the scale scores. That would be your consensus scoring. The more participants, the more accurate the scoring would become. That is, we recommend updating your consensus scoring algorithm every few hundred participants (and less frequently, after a thousand individuals).
To get EIQ scores, you would convert the obtained scores to z-scores, and then multiply them by 15 and add 100.
The process used at MHS is more complex than this and ensures normal distributions and other positive qualities.
If you choose to develop "in-house" norms, the above procedure should provide you with a good proxy for normed scores. That is to say, the scale you have obtained will be good for correlating with criteria.
Note that special-group norms will be most useful if they are compiled using the MHS scoring facilities, so that a direct comparison can be made. I would also recommend that in the case of comparing national or special-group norms, some coordination with the MHS research department be considered, as they have considerable expertise concerning the standardization/ normalization process.