LATITUDE OF ACCEPTANCE AND THE SEMANTIC DIFFERENTIAL*
Department of Speech, Michigan State University
James C. McCroskey
Over the past decade there has been a steady increase in interest in latitude of acceptance and latitude of rejection in the assessment of attitude and attitude change. It has been suggested that the determination of the positions on an issue that are acceptable to an individual (latitude of acceptance), in addition to his most acceptable position, and the positions that are objectionable to an individual (latitude of rejection), in addition to the position he finds most objectionable, yields a more meaningful indication of an individual's attitude than the usual single score determined by conventional attitude scales (4, 5, 6, 7, 8). A growing list of studies utilizing the "Sherif-Hovland technique" tends to support this belief (1, 4, 5, 6, 7, 8).
An important conclusion drawn from these studies is that individuals upholding strong stands on an issue tend to have a smaller latitude of acceptance and a greater latitude of rejection than individuals upholding more moderate positions (5, 7). However, an examination of the type of instrument used in the Sherif-Hovland technique suggests the possibility that ceiling effects of measurement may partially explain these findings. The usual nine-statement measure employed in these studies provides room for the moderate individual to indicate an acceptable position on either side of his most acceptable position. However, for the person upholding a strong position (1 to 9), the only possible acceptable positions, other than his most favored position, are ones of less strength. In short, the latitude of acceptance of the person with a strong stand is artificially restricted by the measuring instrument.
Recent research reported by Diab (2) raises another question concerning the Sherif-Hovland technique. Diab found that his moderate subjects rejected significantly more items than they accepted, a finding expected only for extreme subjects (2). This caused Diab to question the neutrality of the so-called moderate subjects as measured by the Sherif-Hovland technique. Diab found that by employing the semantic differential technique (3) in conjunction with the Sherif-Hovland technique, the behavior of his so-called moderate subjects could be more fully explained (2).
As a result of his findings, Diab (2) suggested a modified measurement technique in which latitudes of acceptance are obtained on semantic-differential scales and a subject's score is based on the mean of his latitude of acceptance. Diab called for investigation of this new procedure in future research (2).
The present investigation was concerned with three questions. Do latitudes
of acceptance decrease with concurrent increases in attitude strength?
Does the Diab semantic-differential scoring procedure yield scores substantially
different from those yielded by the usual semantic-differential scoring
procedure? Are results obtained by Diab's modified measurement technique
employing the semantic differential comparable to those obtained by the
Sherif-Hovland type nine-statement questionnaire?
The sample used in this study consisted of 84 undergraduate students,
54 from Pennsylvania State University and 30 from Michigan State University.
All of the subjects were enrolled in basic oral communication courses.
The median age of both groups was 20.
A series of six bipolar semantic-differential scales for each of 12 concepts was administered to the subjects. Instruction followed the procedure suggested by Diab (2). The subjects first completed the semantic differentials following the usual Osgood (3) instruction. When all the subjects had completed the 12 semantic differentials, they were asked to indicate all other acceptable positions by marking "A" in the appropriate spaces.
Scales employed were good-bad, wise-foolish, fair-unfair, beneficial-harmful, right-wrong, and positive-negative. Pretesting indicated that the six scales all had factor analytic loadings of .80 or higher on the evaluative dimension for each of the 12 concepts.
After the subjects had completed the semantic differentials, a nine-statement
(A to I) questionnaire on "Legalized Abortion" was administered
following the Sherif-Hovland technique. Responses to all measures were
The responses to the semantic differentials were scored in the usual manner (3). Additionally, they were scored in the manner suggested by Diab (2), the individual's score on each scale being represented by the mean value of his latitude of acceptance. Obtained correlations between the scores on each of the 12 concepts obtained by the two procedures were as follows: Socialized Medicine, .96; Censorship, .98; Legalized Abortion, .95; Government Price Controls, .93; School Segregation, .98; Right-to-Work Laws, .90; Premarital Sexual Relations, .91; The Supreme Court's Decision on Prayer in Public Schools, .96; Playboy Philosophy, .94; Federal Aid to Education, .97; Organized Religion, .93; and This Course, .90. The obtained correlation between the scores obtained by the two procedures for all 12 concepts combined was .96. These consistently high obtained correlations indicate that the scoring procedure suggested by Diab procedures scores not unlike those produced by the normal Osgood procedure.
The subjects' scores on the 12 concepts were sorted into three groups, those scoring above the hypothetical neutral point (25-42), those scoring below the hypothetical neutral point (6-23), and those scoring at the hypothetical neutral point. Correlations between semantic-differential scores (Osgood procedure) and latitude of acceptance for the non-neutral groups of subjects were computed. The obtained correlation for the above neutral group was--.30; for the below neutral group, .28. These correlations, though low, were significant and in the expected direction (p < .05). However, as was noted above, it was suspected that ceiling effects might artificially restrict latitudes of acceptance of subjects holding extreme positions. Therefore, subjects with scores representing the five points of most extreme attitude were removed from each of the non-neutral groups. Thus, the range in the high group became 25-37 and the range in the low group became 11-23. The obtained correlations between the semantic-differential scores of these groups and latitude of acceptance were very low and nonsignificant. The correlation for the above neutral group was--.08; for the below neutral group, .11.
These findings suggest that there is no linear relationship between latitude of acceptance and strength of attitudes and that the significant correlations found when the extreme subjects were included most likely were produced by ceiling effects of measurement.
Table 1 reports the results of the Sherif-Hovland type legalized abortion
questionnaire. Very few subjects selected the most extreme positions, too
few for meaningful interpretation. However, the remaining responses indicate
no apparent trend in either latitude of acceptance or latitude of rejection
for the other seven attitude positions. The subjects' scores and latitudes
of acceptance as measured by the semantic differential and the Sherif-Hovland
type questionnaire (scored 1-9) were highly correlated. The correlation
between the scores on the two measures was .87 (p < .001). The
correlation between the latitudes of acceptance on the two measures was
.78 (p < .001).
Latitudes of Acceptance and Rejection Obtained from Sherif-Hovland
Type Legalized Abortion Questionnaire
|Mean Number of Items Acceptable||Mean Number of
|A (Extreme Pro)||5||2.0||3.2|
|I (Extreme Anti)||3||2.0||4.0|
The results of this study call into question the presumption that latitude of acceptance decreases with a corresponding increase in attitude strength. When the most extreme attitude positions were omitted, there was no significant correlation between attitude strength and latitude of acceptance. The observed relationship between attitude strength and latitude of acceptance in previous relationship between attitude strength the latitude of acceptance in previous studies may have been an artifact of the measuring instrument employed. Some may interpret this finding as a challenge to the usefulness of the Sherif-Hovland technique. The writer believes that such an interpretation should not be made at this point. There are major differences in latitude of acceptance between subjects. The results of this study suggest that these differences are not caused by strength of attitude, but no alternative explanation is provided other than the possibility that these are merely random fluctuations. Further research is needed to determine the cause of these differences.
The use of a semantic differential in conjunction with the Sherif-Hovland technique as suggested by Diab (2) was demonstrated to provide results comparable to those obtained by a Sherif-Hovland type questionnaire. Only the evaluative dimension of the semantic differential was employed in this study and, as suggested by Diab (2), increased precision may result if other dimensions are added. However, if only latitude of acceptance is desired by the researcher, the use of the Sherif-Hovland type of questionnaire alone is advisable. The correlation between latitude of acceptance as measured by the semantic differential and the Sherif-Hovland type questionnaire in the above reported study was quite high. The Sherif-Hovland type questionnaire is somewhat easier to administer.
One aspect of the Diab technique, however, is questionable in light of the results of this study. The usual method of scoring the semantic differentials was found to yield scores highly correlated with those yielded by the Diab mean latitude of acceptance technique. Since the usual method of scoring is much simpler, it is to be preferred over the Diab method.
A careful inspection of the data provided by the subjects who scored at the hypothetical neutral point on the semantic differentials produced a noteworthy finding. Ninety-three of the 1008 scores were at the hypothetical neutral point. Of these, 26 indicated a latitude of acceptance of 1. The subjects marked no other position acceptable. The other 67 neutral responses were accompanied by latitudes of acceptance ranging all the way to 37--every position being acceptable. Marginal comments of one subject who scored at the neutral point on two concepts suggested an interpretation of this extreme variance. On the concept of "Playboy Philosophy" the subject indicated no acceptable position other than neutrality and wrote in the margin, "I don't know anything about it." On the concept "Right-to-Work Laws" the subject marked all neutral responses but indicated all other positions were acceptable. The comment in the margin was "I don't care one way or the other."
This finding suggests that Diab's (2) concern with so-called "moderate"
or "neutral" responses was well founded. The "neutral"
position may represent very different things to different people. The latitude
of acceptance obtained by the Sherif-Hovland technique may provide a means
of identifying the individuals who "don't care one way or another."
These individuals will presumably have a large latitude of acceptance.
It will not, however, be possible to distinguish between the "ignorant"
neutral and the "intense" neutral on the basis of latitude of
acceptance. Both are likely to have a very small latitude of acceptance.
However, it may be possible to distinguish between them on the basis of
latitude of rejection. The "intense" neutral will presumably
have a relatively high latitude of rejection, while the "ignorant"
neutral will presumably have a relatively low latitude of rejection. The
results of this study provide no relevant data to support this latter speculation.
The purpose of the present study was to investigate the relationship between latitude of acceptance and strength of attitude, to learn whether the Sherif-Hovland technique employed in conjunction with the evaluative dimension of the semantic differential as suggested by Diab (2) would yield results comparable to results obtained by the Sherif-Hovland type nine-statement questionnaire, and to determine whether the mean latitude of acceptance scoring procedure suggested by Diab (2) yields scores substantially different from those yielded by the usual semantic-differential scoring procedure.
Six-scale semantic differentials on 12 concepts and a Sherif-Hovland type questionnaire were administered to 84 college student subjects. The results indicated that latitude of acceptance was not significantly correlated with attitude strength when extreme respondents were omitted and only slightly correlated when extreme respondents were retained. The Diab procedure yielded results very comparable to those yielded by a Sherif-Hovland type nine-statement questionnaire on legalized abortion. The Diab procedure for scoring semantic differentials was found to yield scores highly correlated with those yielded by the usual Osgood scoring procedure.
A wide range in latitude of acceptance for "neutral" subjects
was observed. Interpretation of these responses was not possible with the
Sherif-Hovland technique as modified by Diab. It was suggested that the
full Sherif-Hovland technique should be employed with the semantic differential
to provide data to facilitate interpretation of "neutral" responses.
1. Diab, L. N. Studies in social attitudes: I. Variations in latitudes of acceptance and rejection as a function of varying positions on a controversial social issue. J. Soc. Psychol., 1965, 67, 283-295.
2. -----. Studies in social attitudes: III. Attitude assessment through the semantic-differential technique. J. Soc. Psychol., 1965, 67, 303-314.
3. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. The Measurement of Meaning. Urbana, Ill.: Univ. Illinois Press, 1957.
4. Sherif, C. W., Sherif, M., & Nebergall, R. E. Attitude and Attitude Change. Philadelphia, Pa.: Sanders, 1965.
5. Sherif, M. Some needed concepts in the study of social attitudes. In J. G. Peatman & E. L. Hartley (Eds.), Festschrift for Gardner Murphy. New York: Harper, 1960.
6. -----. Conformity-deviation, norms, and group relations. In I. A., Berg & B. M. Bass (Eds.) Conformity and Deviation. New York: Harper, 1961.
7. Sherif, M., & Hovland, C. I. Social Judgement. New Haven: Yale Univ. Press, 1961.
8. Sherif, M., & Sherif, C. W. An Outline of Social Psychology.
New York: Harper, 1956.
Department of Speech
College of Communication Arts
Michigan State University
East Lansing, Michigan 48823
Click Here To Go Back To PERIODICALS