Editor's note: E.B. Feltser is a freelance writer based in San Diego. She has worked as a marketing research interviewer and survey writer.

A multi-point scale is a wonderful thing. Subtle, nicely objective, neatly quantifiable. Interviewers appreciate them because they're fast and don't entail all the typing or handwriting work associated with open-ends. It's easy to understand why they are so common in surveys.

Consider the basic, garden-variety five-point scale:

5. Extremely satisfied

4. Satisfied

3. Neither satisfied nor dissatisfied

2. Dissatisfied

1. Extremely dissatisfied

6. Don't know (DO NOT READ)

7. Refused (DO NOT READ)

Read it aloud, at speed. You're in an echo chamber, right? That's thanks to numbers 2 and 4. The easy fix is to add an adjective to 2 and 4: somewhat satisfied, maybe, or mildly satisfied or routinely satisfied or whatever. That kind of parallel construction gets a rhythm going that respondents seem to remember more easily. It also cuts down on the "I'm quite satisfied." "Would you say you're extremely satisfied or satisfied?" "Quite satisfied." round-and-round that drives interviewers up the wall and inspires them to key in 5 or 4 as the mood strikes, just to get on with the survey. And in fact, "satisfied" and "dissatisfied" are open-ended emotional states that logically include all other levels of satisfaction or dissatisfaction, including the "extremely" level. It's no wonder some respondents have a tough time nailing that one. Incidentally, why does the mid-point have to be so user-unfriendly? Why not a simple "I'm neutral about it"? It takes less time to say, and surveys, especially long ones, are like Olympic track events: seconds count, and they add up fast.

Don't know (DO NOT READ) and the mid-point paradox

"Gee. I've never tried it. I guess I'd give it a 3. That's neutral, right?" Thus does the respondent valiantly try to fit the square peg into the round hole. But what's a poor interviewer to do? The classic instruction is to go with the respondent's numerical rating. And yet, it must be important from a marketing standpoint to distinguish between potential devoted consumers (I've never tried it) and experienced but disenchanted ones (It's OK, nothing special. Give it a 3.).

Perhaps it's time to replace the classic instruction with something more useful to an interviewer. Treating "don't know" as a naughty little secret instead of a viable response forces interviewers to make subjective (often highly subjective) choices, and that turns mid-point quantification into mush. Besides, the endearing thing about most respondents is that they truly do try to give honest opinions within the response framework. Even those who are aware that "don't know" is an option tend to ignore it whenever they can. Tipping the respondent to the "don't know" option cuts down on such pleasant responses as "I've already told you three times I never heard of it. I don't want to answer any more of these stupid questions. Good-bye!" Anything that cuts down on mid-terms is a thing of joy and delight to an interviewer.

Multi multi-point scales

(Q. 22) Now, using a seven-point scale where 7 means very satisfied. . .

(Q. 24) Now, using a five-point scale, where 5 means liked very much. . .

(Q. 26) Now, using a seven-point scale where 7 means disliked very much. . .

Well, OK, in truth it's been a while since I've done a survey where the multi-pointers reversed polarity in mid-stream (thank Heaven for small favors!). But why do some topics warrant a cruder response than others - a five-pointer set like a grubby medieval serf amongst the aristocratic, fine-tuned seven-pointers? From the interviewer's point of view, it means using time to explain the new scale, and then using more time to reintroduce the returning seven-pointer. Most respondents grasp the concept of a multi-point scale quite easily, but too much shifting around gives even the sharpest of them whiplash. Worse, they get confused, and even more time is used up explaining the scale yet again to bewildered respondents ("Five. No, seven. Seven? Is seven the top this time?").

Words, words, words

Time the following while you read it aloud:

"You mentioned that you have purchased Dy-No-Mite Drain Cleaner. Now I'd like to get your opinion of that product. Would you say you are:

1. Completely satisfied with the product;

2. Somewhat satisfied with the product;

3. Neither satisfied nor dissatisfied with the product;

4. Somewhat dissatisfied with the product;

5. Completely dissatisfied with the product."

Now, time this version:

What's your opinion of Dy-No-Mite Drain Cleaner? Are you:

1. Completely satisfied with it

2. Somewhat satisfied

3. Neutral about it

4. Somewhat dissatisfied

5. Completely dissatisfied"

The second version takes about half as long to read, and holds the respondent's interest better because it comes on faster. Like reading an exciting novel, respondents should get caught up in the challenge of the next question before they have time to reflect on how long they've been at it. Time is of the essence to interviewers as well: the more dialings, the more completes. Anything that needlessly uses up time in a survey jeopardizes the chances of the complete and reduces the interviewer's total dialings for the shift.

The inappropriate multi

I recently worked on a survey that used a multi-point scale to determine whether a business had certain amenities. Let's say it involved the respondent's favorite restaurant (it didn't, actually). How do you interpret a mid-point scale response to the question of, say, whether the restaurant had full menus? It had menus but not full ones? It had full menus but only on alternate days? It had full menus but only gave them to special customers? That's an example of a survey writer getting carried away with the wonders of multi-point scales (or, more likely, just getting sloppy). Whatever the reason, it wasted a lot of interviewing time as many respondents diligently worked out mid-point choices, although it did provide some giggles for respondents sharp enough to notice.

Well-structured, carefully worded (and appropriate) multi-point scales are a great tool. Streamline them and that purring sound will be interviewers throughout the land murmuring their thanks.