August 4, 2004
Berscheid, E., Snyder, M., & Omoto, A.M. (1989). The Relationship Closeness Inventory: Assessing the Closeness of Interpersonal Relationships. Journal of Personality and Social Psychology. 57 (5), 792-807.
Relationship closeness has always been a subject of much interest and will most likely continue to be. The authors of Relationship Closeness Inventory: Assessing the Closeness of Interpersonal Relationships found it to be especially interesting that there had been no instrument created that would be capable of measuring relationship closeness in a variety of relationship types while still retaining reliability and validity. The authors of the Relationship Closeness Inventory (RCI) set out with a goal of creating an index that would indeed be capable of measuring relationship closeness in any type of relationship. Until this study was conducted most tests of relationship closeness were only focusing on one specific type of relationship at a time (usually romantic relationships). Berscheid, Snyder, and Omoto (1989) though it was time to create something that could be used in multiple situations and that would perhaps tap into the realm of other types of relationships (such as familial relations or friendships). Their preliminary tests, however, proved to give them more trouble than they had anticipated and ended up yielding similar results to tests that were already in existence that exclusively measure romantic relationships.
The need for the Relationship Closeness Index stems from the inability of researchers to observe subjects on an ongoing basis in their natural environments. While having that type of observation would be ideal in assessing relationship closeness, it is quite literally impossible. The RCI’s intended purpose is to provide a window into the relationship or relationships of an individual by assessing the degree of closeness of a given relationship. Further implications include the ability of the RCI to predict the stability and even longevity of certain relationships.
The RCI was designed to measure relationship closeness based on measures of three separate but equally important constructs. The first item measured is the time that two individuals spend alone together during their waking hours. This was originally measured in number of hours spent together at different times of the day on an average day but the authors later decided to use a more general measure of hours spent together in the week leading up to the day the participant responded to the inventory. The second construct under investigation is the degree of influence the respondent believes his or her partner has on him or her when making decisions or in general. Finally, the diversity of activities partners participate in together in the week before responding to the inventory was examined as well to help determine relationship closeness. Two of these areas were measured by presenting the test taker with a set of options from which they could choose as many as applied to them. These options included items such as “did laundry” and “discussed things of a personal nature” when reporting on diversity of activities performed together (of which there are 38 choices available) and items such as “does not influence everyday things in my life” and “influences the way I feel about the future” when reporting on the strength of the influence the partner has on the subject (there are 34 items to choose from in this category). For “time spent together” on the Frequency scale the respondent was simply asked to record how much time was spent together in hours and minutes over the past week.
When the scales of measurement were created the authors realized that the scores should all
be displayed on a 10-point equal interval scale, with scores near 10 conveying a very close relationship and scores nearer to 1 being far less close. This desire for a uniform rating scale to be used in comparisons and correlations posed an additional need, that of statistical transformations of the raw scores of both the Frequency scale and the Diversity scale. To accomplish this the authors used a square root transformation on the raw scores of both the Diversity scale and the Frequency scale. The strength scale was straightforward and needed no transformations. The authors felt that this change would allow for the scale to be the same for any sample population and would also prove to yield more meaning when making comparisons between the three sub-groups.
The reliability and validity of this inventory are good in some respects and not so good in others. What the authors truly set out to do was to create a measure that would report accurately on the closeness of any type of relationship. What the authors ended up with was basically one more inventory measuring romantic relationship closeness. This, however, was not entirely their fault since the population used to construct the test consisted of college students who chose to report mostly on romantic relationships. That said, let us discuss what reliability and validity there was both over all and relating to specific relationships.
The test-retest reliability posed a special problem because even if the test itself proved to be stable and valid, relationships can be unstable and may not lend themselves as good subject matter for this type of reliability. Nonetheless the authors felt it was necessary to find this type of reliability if at all possible and they decided on a time period of 3 to 5 weeks after which the original test subjects would come and respond again to the inventory. The correlation between the first total scores of the RCI and the second total scores ended up being r (75) = .82, p<. 001 with the sub scores being similar. The t scores all fell < 1.7 ns so the test retest reliability was found to be acceptable. As far as validity is concerned, it was determined that the test has only moderate validity when it came to relationship closeness in all types of relationships and being able to distinguish close from not close. The discriminant validity was tested by distinguishing
the RCI from other instruments in which the respondent simply reports on how close he or she believes their closest relationship is and how they believe their close relationship compares to other close relationships they have observed. Through the testing of discriminant validity the authors found that the correlation between the total score of the RCI and the total score of the Subjective Closeness Index were statistically significant but not extremely.
The RCI and the Subjective Closeness Index scores differed because many people assume that close relationship is a positive one while the RCI reveals not just positive close relationship but also negative ones. Several other tests were used to find correlations and meaning in order to prove the validity and reliability of the RCI but the extent of these tests and their results are too numerous and extensive to cover in detail here. Suffice it to say that this test can claim validity and reliability of a reasonable measure because of the comparisons and correlations (or lack of correlations) that they have come up with through extensive and arduous testing of subjects.
While the test did prove to be reliable and valid, both of these measures tend to go only so far as to apply to romantic relationships. The predictions that can be made from this test are only in relation to romantic partners and exclude other types of relationships such as those involving family members or friends. The authors can be proud that they created something valid and useful but they have also come up somewhat short in the area in which they set out to improve. However, the authors are fully aware that they have made only a small dent in the massive field of research regarding relationship closeness in multiple types of relationships and they have encouraged others to work beyond what they have presented so far. Personal suggestions include perhaps constructing an inventory specifically for familial relationship closeness, another for friendship closeness, and still another for more for measuring closeness in casual relationships. In this way the tests could be combined to produce more accurate and predictive results. In addition, the way in which the RCI was constructed could easily have been changed to the benefit of the test creators. They started out by asking the subjects to report only on their closest relationships and later realized they should ask about people’s relationships that were not as close. What might have been done is each respondent could have reported on a set number of relationships, perhaps 5, with one being the closest and the next being not quite as close and so on until the respondent was reporting on a “not close” relationship. In this way the authors might have found more helpful and valid information for their test. Also, the test makers could have required the respondents to report on each type of relationship so that each respondent would have done at least 3 reports, one for a familial relationship, one for a friendship, and one for a romantic relationship (or if none exists, another friendship or family relationship). In any case, tests are most likely being constructed today that are contributing to this pool of new information and it will be most interesting to see if a test ever comes into existence that can truly assess the closeness of various relationship types with one quick administration.
Outline August 4, 2004
1. Presenting an instrument to assess relationship closeness
2. Close Relationships can be characterized by:
a. The time two individuals spend alone together
b. The strength of the influence two people have on one another when they spend time alone
c. The diversity of the types of activities two individuals participate in together
3. The indicators of closeness chosen for this test should be broad and include many relationship types and diverse populations.
4. The test instrument should be easy to administer and should not take too much time.
5. Designed as a self-report instrument
6. The test creators began preliminary development by collecting information on a population of college students’ closest relationships as identified by the subjects themselves.
7. Problems with the study include:
a. Little variability because the respondents chose only their closest relationship to report on.
b. Limited population (Minnesota college students)
8. Identifying the “closest relationship”:
b. Choose from 15 possibilities
c. Each of those 15 options to be fit into a broader category
9. The majority of subjects named a romantic partner as the person with whom they shared their closest relationship
1. The RCI (Relationship Closeness Inventory)
a. Frequency of impact indicated by amount of time the subject spent alone with their partner while awake in the past week (current edition) or on an average day (preliminary edition).
b. Diversity of activities performed together measured by subject’s indication of number of different activities he or she shared with their partner in the last week. The subject can choose from a list of 38 items.
c. The strength of impact (or influence two people have on one another) assessed by having subject choose from scale of 1-7 how much influence they believe their partner has on them in making certain decisions
2. The duration of the relationship is not seen as the best measure of closeness.
3. A high level of interdependece in a relationship would be a great measure of relationship closeness but the authors could not come up with a scale that could accurately measure such a construct.
4. Each of the scales in the RCI is measured independently and has its own method by which its raw score is converted to fit the uniform 10-point scale used to score the test. The higher the score, the closer the relationship.
5. The authors expect that the three scales will all be intercorrelated in regards to people’s closest relationships.
1. Test-retest reliability:
a. The authors chose a 3 to 5 week retest period
b. Even if a test is actually stable and reliable, relationships that people have are not necessarily stable so that could skew the results
c. The RCI ended up having what the authors call “acceptable” test-retest reliability
2. Closest Relationships
a. Romantic relationships were reported to be the closest type of relationship
b. Family and Friend relationships did not differ from one another in closeness but were less close than romantic relationships
c. The authors found no differences between the sexes as far as how they rated their closest relationships
3. Longevity and Closeness
a. As the authors suspected, the longevity of a relationship did not necessarily correlate with the closeness of a relationship
b. Family relationships tend to be longer standing than romantic relationships yet the romantic relationships were generally rated as being closest.
c. Short-term friendships were actually found to be closer than long term ones.
4. Relationships that are seen as “not close”
a. The RCI’s validity is demonstrated through the test’s closeness scores being nearly twice as high for people who were rating a close relationship than or those who were rating a “not close” relationship in preliminary administrations of the RCI.
5. Reliability and Validity
a. The reliability and validity are satisfactory
b. The discriminant validity was tested by distinguishing the RCI from other instruments in which the respondent simply reports on how close he or she believes their closest relationship is and how they believe their close relationship compares to other close relationships they have observed.
c. Through the testing of discriminant validity the authors found that the correlation between the total score of the RCI and the total score of the Subjective Closeness Index were statistically significant but not extremely.
d. The RCI and the Subjective Closeness Index scores differed because many people assume that a close relationship is a positive one while the RCI reveals not just positive close relationships but also negative ones.
6. Additional Considerations
a. The authors also tested preliminary participants for the affective quality of their relationship with their partner using the Emotional Tone Index.
b. There was no reported correlation between the scores on the RCI and the Emotional Tone Index.
c. There was a significant correlation between the Emotional Tone Index and the Subjective Closeness Index (because people tend to choose positive relationships for the Subjective Closeness Index).
d. The authors used a sub-sample of 83 romantically involved heterosexual couples to see if there is trend of a significant correlation between two partner’s RCI scores. There was, but it was nothing to brag about.
e. The RCI also proved to be able to predict (to a certain extent) whether or not romantically involved couples will remain together by assessing how close they rated their relationship to be.
1. Many articles have been written specifically about relationships involving romantic partners but few have attempted to discuss relationship closeness as it relates to multiple types of relationships.
2. The authors stress that the RCI has made only a tiny dent in the vast universe of attempting to separate “closeness” from a specific type of relationship (usually romantic).
3. Since this report included mostly romantic relationships, however, that ended up being a major focus of the study and of the creation of the RCI.
4. The RCI has proved to be useful when studying romantic relationships and their degree of closeness but it needs to be used more in other relationship populations in order for its usefulness there to be assessed.
5. The authors feel that even though they did not accomplish what they initially set out to do they have made important advancements in the area of assessing relationship closeness in relationships that are not romantic because they have at the very least brought up new questions which others may be able to answer over time.