Method #14 (Semantic Differential)
This scale has been prepared so that you can indicate how you feel. There are
pairs of words with opposite meanings. Place a check in the blank closet to the
word that indicates your feelings. If you are neutral, place a check in the
middle blank.
1. Anxious ____ ____ ____ ____ ____ ____
____ ____ ____Calm
2. Excited ____ ____ ____ ____ ____ ____
____ ____ ____ Indifferent
3. Elated ____ ____ ____ ____ ____
____ ____ ____ ____ Depressed
Method I (Likert) could also be changed to true or false. That is, either it did
occur or it did not.
The Likert method is recommended as a standard; the other methods are for
special cases. The problem with true-false formats (and consequently the
check-list) is that it produces dichotomous data and in a sense is a
forced-choice item. Statistical power is lost in dichotomous data, and some
participants indicate that they cannot respond with their true feelings on
forced-choice items. The semantic differential is essentially like the Likert
method but takes up more room on a page and takes a lot more time to write
items. The problem with ranking is that many respondents answer such items
incorrectly by responding to them like a checklist or leaving out some of the
response set. (Don't ever use the ranking method.)
The following discussion is written for the beginner in writing items. It is
written to give hints about making goals into criterion variables.
State the goals so that it appears that two or more people could agree when the
target person had reached the goal. This can later be tested as reliability but
at present the interest is in writing items so don't be too much of a stickler.
For example, the goal may be for a psychiatric patient to stop talking to
himself. Suppose that the client finds out that this is the goal and talks to
himself only when he is not being observed. It should be noted that the behavior
has not stopped, but it will not be measured accurately. On the other hand, if
the goal is for the person not to talk to himself when someone else is present,
the measure is accurate. Such issues should be clarified.
In another example, the goal may for a client to be "free from anxiety." The
problem is to make such a goal into a statement in which it could be determined
by two or more people that the person was "free of anxiety." The best way is to
make as many statements as possible about anxiety and then test the reliability
of such statements. Some such statements might be: (1) Do his/her hands shake?
(2) Will s/he go out on his/her own? (3) Is s/he fearful? (4) Is s/he afraid of
people? etc.
The main point to be made here is that the statement must describe a situation
in which some quantitative judgement can be made as to whether the client has
reached the goal or not.
The next step is to make as many statements as are feasible to include all the
goals which you have stated. The following are some informal rules for writing
items for measurement. They are not absolute rules, but the more of them that
are broken, the less likely is the reliability of the item. Apply these rules to
the statements which you make about the goals.
1. Avoid statements that may be interpreted in more than one way.
2. Avoid statements that are irrelevant to the psychological object under
consideration.
3. Avoid statements that refer to the past rather than the present (unless it is
specifically the past you are concerned about).
4. Avoid statements that are likely to be endorsed by almost everyone or almost
nobody.
5. Select statements that are believed to cover the range of the goal or goals
you have outlined.
6. Keep the language of the statements simple, clear, and direct.
7. Statements should be short, rarely exceeding 12 words.
8. Each statement should contain only one complete thought.
9. Statements containing universals such as all, always, none or never, often
introduce ambiguity and should be avoided.
10. Such words as only, just, merely, and others of a similar nature should be
used with care and moderation in writing statements.
11. Whenever possible statements should be in the form of simple sentences
rather than in the form of compound and complex sentences.
12. Avoid the use of words that may not be understood by those who will be using
the scale.
13. Avoid the use of double negatives.