The repertory grid: name the people who matter, take them three at a time, say how two are alike and the third differs. The constructs are yours, in your words. The page does the arithmetic and prints no type, no score and no label.
1. The people
Six to ten is plenty. Roles, not just names: a teacher you liked, a teacher you disliked, your mother, someone you find hard, the most interesting person you know.
2. Three at a time
Add at least three people above.
The words are yours. The page never supplies one, which is the whole point of the instrument and also its flaw.
3. The grid
Rate each person on each construct, 1 (the left pole) to 5 (the right pole).
4. What travels together
Grids kept
The honest limit, stated first. The grid uses your words, so it cannot put anything in your mouth, and that is also what is wrong with it: two people's grids are not comparable, because they are not made of the same material. The statistics assume the ratings are evenly spaced, and one worker in the field said plainly that testing that assumption would need the same grid filled in many times over, which is a condition a single elicited grid cannot satisfy. Which three people you happen to pick changes which constructs come out. Nobody has shown that a shrinking distance between how you rate yourself and how you rate your ideal corresponds to anything outside the grid. It may be the most elegant mirror anybody ever built, and a mirror is not a measurement. This page computes which of your own constructs move together and stops there. It prints no type, no score and no diagnosis, because the instruments in this volume that tried to were measuring something other than what they said.