APPENDIX E
Simple Instructions for Use of Reiss'
Premarital Sexual Permissiveness Scales
These instructions are not intended to replace or substitute for the full Guttman-scaling procedures that can be used with these scales. Rather, they are intended as a guide for those who want an efficient way of handling the scales that does not involve much technical background or reading.
First, always use all twelve items in each of the two scales (male scale and female scale) when administering them to respondents. The reasons for this are many. With all twelve items available, the researcher can select the four dealing with petting and use them alone for a petting scale. He can do the same with the four items concerning coitus or kissing. Also, he can select the three items dealing with situations where affection is not present (4, 8 and 12) and use them as a nonaffection scale.
Secondly, also use the six-way choice after each item, for that makes the respondent feel he can present his feelings more accurately, and you may at some future time want to use this full information. Most respondents can make this six-way choice for all twelve questions in a scale in a matter of a few minutes, so not much time is taken up by asking for a full answer to each of the twelve items.
After administering both the male and female scales in full form, the researcher may assume that he has a Guttman scale, since these scales have been tested on thousands of individuals in various groupings and always have worked properly. However, as noted previously, ideally the researcher should test for a Guttman scale using one of the Guttman-scaling procedures available on computers (there is one at the University of Iowa Computer Center that you may write for) or using a hand method such as that described by Louis Guttman in "The Cornell Technique for Scale and Intensity Analysis," Educational and Psychological Measurement, 7 (Summer 1947), pp. 247-280. Treat all responses as dichotomies so that any degree of agreement is counted as agree and any degree of disagreement is counted as disagree. This greatly simplifies the procedures involved. For other purposes, such as intensity analysis, use the open responses; but for simply testing for a Guttman scale, use the dichotomous form.
These instructions have been prepared under the assumption that the researcher will sometime make the above official Guttman-scale check, but that he wants a quick method for the moment. A scale that uses items 5, 6, 7, 9, 10, 11, and 12 - seven of the twelve items - is recommended. Scores would be assigned to individual respondents on the basis of the following diagram.
SCALE TYPE[a]
ITEM NUMBER 0 1 2 3 4 5
5 or 6 or both - + + + + +
7 - - + + + +
9 or 10 or both - - - + + +
12 - - - - - +
a. A plus sign indicates agreement with the item.
The diagram indicates that respondents would be assigned to one of the six scale types on the basis of how he or she answered the seven items. For example, a person would be given a scale type of zero if he disagreed with all items. He would get a scale type of five if he agreed with all items. The other scale types fall in-between these two extremes. Items 5 and 6 have been combined, so that if a person answered Yes to one or both of these, he would be counted as agreeing; in short, only those who answered No to both of these would be counted as disagreeing. Items 9 and 10 have also been combined into one item in exactly the same way. This was done because the responses to these four items were so close that it was not important to separate them from each other.
You can simply look at how each respondent answered these questions and place him in one scale type or another accordingly. There are a few minor problems that may arise. First, the researcher should drop all respondents who have not answered all twelve questions; even though he uses only seven of them here, it is best not to include those who have not fully replied, for he cannot be sure he possesses their complete attitude. Second,-and this is more complicated -is the question of what to do with those respondents who are not "pure" scale types, that is, who do not fit perfectly into any of these scale types. In reality the researcher should not have to handle many error types, for more than ninety percent should be pure scale types. But when error types do appear, he should try to place the person in the scale type closest to his response, and in so doing try to place him nearer the center of the scale than toward either extreme (scaletypes zero or five). For example, if a person agreed with items 5 and 6, disagreed with item 7 and agreed with item 9, he would have a scale-type three response with one error (item 7), and he could simply be called a scale-type three, which is the middle of the scale. However, if in addition to the above response the person had disagreed with item 11 and agreed with item 12, he would not be called scale-type five, for that is an extreme type, and it is best not to put people there unless they clearly fit. Instead he would be called scale-type three and placed in the middle of the scale. One other example: if a person responds Yes to items 5 and 6 and also agrees with items 7 and 12, he would be called scaletype two, which would keep him near the middle of the scale and would also produce only one "error" in his pattern. (The error would be his agreeing to item 12.) If he were called scale-type five, he would have two errors, for he would be missing on agreement to item 11 and to the combined item (9, 10). Thus, he fits better (has less error) in scale-type two. Making better fits and keeping toward the center of the scale are the two guides to use in assigning cases that are not pure scale types. A little practice and it should be quite easy. This sort of thing is required in the "Cornell" method of Guttman-scaling described in the article by Guttman noted above. Thus, this practice is useful for learning Guttman-scaling techniques.
There is another very important point. When the researcher obtains answers for both the male scale and the female scale for each respondent, he should first give scale types for each respondent on the scale that is the same sex as the respondent (women on the female scale and men on the male scale) and call this their same-sex permissiveness scale type. Then he should scale the respondent on the scale type of the sex opposite to himself and call this the opposite-sex permissiveness scale type. To do this most easily, it would be best to group separately all questionnaires answered by men and all answered by women respondents. If the researcher wants to measure the permissiveness of his respondents, he should use the same-sex scale types. If he wants to measure equalitarianism, he should compare the same-sex scale types with the opposite-sex scale types and classify the people who differ as nonequalitarian. He could distinguish the nonequalitarians by direction and by how many scale types they differed on the two scales. If he wants to measure the permissiveness allowed the opposite sex, he would use the opposite-sex scale types.
For the researcher's own purposes, he may want to reduce the six scale types to two or three groups of scale types. He should then combine all those who are scale-types zero, one, and two and call them low permissives and all those who are scale types three, four, and five and call them high permissives. That would afford a dichotomy of the scale. For a trichotomy, call scale-type zero low, scale-types one and two medium, and scale-types three, four, and five high. This is useful for dealing with a sexually conservative group. With a high-permissive group, trichotomize by combining scale-types zero, one, and two as low, three and four as medium, and five as high. Note that scale-type zero will be mostly those who only accept kissing; scale-types one and two also accept petting, and scale types three, four, and five also accept coitus.
By following these instructions the researcher should be able to handle responses to both the male scale and the female scale quite easily and to assign each respondent a scale type for each of these two scales. 1
----------------------
NOTES:
1 If one is interested in using sexual standards, see fn. 13 in Chapter 2 for instructions.