19
Measuring distance in personal relationships: The Relational Distance Index JON A. HESS University of Missouri–Columbia Abstract This article describes the development of an instrument that measures people’s reported use of distancing tactics in personal relationships. Two versions of the Relational Distance Index (RDI) were developed, an 8-item version and a 17-item version. Four studies were conducted to examine the instrument’s psychometric properties. The index exhibited a stable two-factor structure (unfriendly and withdrawal), had acceptable internal and temporal reliability, and performed well on the tests of convergent and discriminant validity. The data suggested the RDI is suitable for use with both student and nonstudent samples. The degree of closeness or distance is one of the most fundamental dimensions of every relationship. It is often the first quality that people mention when describing a relation- ship, and it seems to be one of the most important aspects of relational communica- tion. Jacobson (1989) noted that, ‘‘As we became increasingly involved in the task of identifying interactional themes in recurring marital conflicts, we began to notice that the struggles were almost invariably around how much closeness or distance there was going to be in the relationship’’ (p. 30). Kantor and Lehr (1975) argued that dis- tance (both psychological and spatial) is the single most central element in family interaction. Given the centrality of closeness and dis- tance in personal relationships, it is not sur- prising that researchers have devoted considerable attention to defining what is meant by a ‘‘close’’ relationship (e.g., Kelley et al., 1983; Parks & Floyd, 1996) and to assessing the degree of closeness in subject- ively close relationships (e.g., Berscheid, Snyder, & Omoto, 1989; Ginsberg & Gottman, 1986; Kelley et al., 1983; Shortt & Gottman, 1997). What is surprising is how little attention researchers have paid to the other end of the spectrum. Sparse attention has been paid to defining what is meant by a ‘‘distant’’ ongoing relationship (Helgeson, Shaver, & Dyer, 1987), and little effort has been made to assess the degree of distance in personal relationships that are noticeably detached. At present, there is no validated instrument specifically designed for measur- ing distance in personal relationships. The lack of an instrument that measures distance is detrimental to the study of per- sonal relationships. A valid and reliable measure of distance could be a useful tool for scholars who study personal relation- ships. Although distancing by one or both relational partners is typically a sign of strained relations, such as negative affect or ongoing conflict (e.g., Jacobson, 1989; Napier, 1988; Sullaway & Christensen, 1983), appropriate distancing can facilitate An earlier version of this article was presented at the 2002 meeting of the International Association for Relationship Research in Halifax, Nova Scotia. The author thanks Lance Holbert for contributions to the preparation of this manuscript. Correspondence should be addressed to the author at the University of Missouri–Columbia, Department of Communication, 115 Switzler Hall, Columbia, MO 65211; e-mail: [email protected]. Personal Relationships, 10 (2003), 197–215. Printed in the United States of America. Copyright # 2003 ISSPR. 1350-4126/02 197

Measuring distance in personal relationships: The Relational Distance Index

Embed Size (px)

Citation preview

Page 1: Measuring distance in personal relationships: The Relational Distance Index

Measuring distance in personal relationships:

The Relational Distance Index

JON A. HESSUniversity of Missouri–Columbia

AbstractThis article describes the development of an instrument that measures people’s reported use of distancing tactics

in personal relationships. Two versions of the Relational Distance Index (RDI) were developed, an 8-item

version and a 17-item version. Four studies were conducted to examine the instrument’s psychometric properties.

The index exhibited a stable two-factor structure (unfriendly and withdrawal), had acceptable internal

and temporal reliability, and performed well on the tests of convergent and discriminant validity. The data

suggested the RDI is suitable for use with both student and nonstudent samples.

The degree of closeness or distance is one ofthe most fundamental dimensions of everyrelationship. It is often the first quality thatpeople mention when describing a relation-ship, and it seems to be one of the mostimportant aspects of relational communica-tion. Jacobson (1989) noted that, ‘‘As webecame increasingly involved in the task ofidentifying interactional themes in recurringmarital conflicts, we began to notice thatthe struggles were almost invariably aroundhow much closeness or distance there wasgoing to be in the relationship’’ (p. 30).Kantor and Lehr (1975) argued that dis-tance (both psychological and spatial) isthe single most central element in familyinteraction.

Given the centrality of closeness and dis-tance in personal relationships, it is not sur-prising that researchers have devoted

considerable attention to defining what ismeant by a ‘‘close’’ relationship (e.g., Kelleyet al., 1983; Parks & Floyd, 1996) and toassessing the degree of closeness in subject-ively close relationships (e.g., Berscheid,Snyder, & Omoto, 1989; Ginsberg &Gottman, 1986; Kelley et al., 1983; Shortt &Gottman, 1997). What is surprising is howlittle attention researchers have paid to theother end of the spectrum. Sparse attentionhas been paid to defining what is meant by a‘‘distant’’ ongoing relationship (Helgeson,Shaver, & Dyer, 1987), and little effort hasbeen made to assess the degree of distancein personal relationships that are noticeablydetached. At present, there is no validatedinstrument specifically designed for measur-ing distance in personal relationships.

The lack of an instrument that measuresdistance is detrimental to the study of per-sonal relationships. A valid and reliablemeasure of distance could be a useful toolfor scholars who study personal relation-ships. Although distancing by one or bothrelational partners is typically a sign ofstrained relations, such as negative affector ongoing conflict (e.g., Jacobson, 1989;Napier, 1988; Sullaway & Christensen,1983), appropriate distancing can facilitate

An earlier version of this article was presented at the2002 meeting of the International Association forRelationship Research in Halifax, Nova Scotia. Theauthor thanks Lance Holbert for contributions to thepreparation of this manuscript.

Correspondence should be addressed to the authorat the University of Missouri–Columbia, Departmentof Communication, 115 Switzler Hall, Columbia, MO65211; e-mail: [email protected].

Personal Relationships, 10 (2003), 197–215. Printed in the United States of America.Copyright # 2003 ISSPR. 1350-4126/02

197

Page 2: Measuring distance in personal relationships: The Relational Distance Index

relational growth (Ryder & Bartle, 1991),might help prevent dysfunctions such ascodependency (e.g., Wright & Wright,1995), and can be an adaptive way tomaintain nonvoluntary relationships withdisliked partners (Hess, 2000). Peopleregulate closeness and distance in everyinteraction, so the ability to study distancemore systematically offers the potential toreveal new insights about how relationshipsare maintained. A good measure of distancealso offers the prospect of facilitating theor-etical advances. Distance has receivedincreased attention in recent theoreticalperspectives on relational dynamics (e.g.,Baxter & Montgomery, 1996; Hess, 2000;Vangelisti & Maguire, 2002), and with therecent surge of research examining the darkside of close relationships (e.g., Spitzberg &Cupach, 1998) distance may become evenmore important as scholars attempt todevelop a more balanced and comprehen-sive understanding of relational processes.The ability to empirically examine distancewould assist scholars in formulating andtesting theories that specify the role of dis-tance in personal relationships.

Conceptualizing and Measuring Distance

Conceptualizing distance and closeness

The concept of distance is based on the spatialmetaphor, and it is used to refer to a lack ofrelational closeness (Kreilkamp, 1981). It hasbeen defined as ‘‘a feeling of separation fromanother’’ (Hess, 2002, p. 664). Because dis-tance is the opposite of closeness, the reverseof the conditions that Berscheid et al. (1989)supplied for closeness—frequent, strong, andvaried causal connections—will lead to dis-tance. Thus, people can create distancethrough infrequent, weak, and unvariedcausal interconnections.

Closeness and distance are the two poleson a single spectrum, and are thus two facesof the same construct. However, it may beinappropriate to conceptualize or study dis-tance simply as a lack of closeness. Theliterature on distance treats it as somethingmore than the absence of closeness; instead,

distance is conceptualized as the presence offactors that create a noticeable rift in anongoing relationship (e.g., Helgeson et al.,1987; Kreilkamp, 1981).

At first glance, the assertion that distanceis something other than a lack of closenessmay sound contradictory, but it is actuallya common type of conceptualization. Forexample, trust and distrust are two ends ofa single continuum, yet for a person withoutpredispositions, the absence of trust ismerely vigilance or wariness. It takes anact of treachery or deceit to create convic-tions of distrust in the other. Consider acontinuum on which closeness is repre-sented by a high positive score and distanceis represented by a high negative score.A relationship characterized by significantamounts of closeness-producing behaviorswould be placed on the positive end of thescale. The absence of such behaviors wouldnot move the relationship to the far negativeend of the scale, but, rather, back towardthe midpoint. It is the presence of distancingbehaviors that propels a relationship to thefar extreme.

An examination of specific behaviorsshows how this effect can happen. Makingreference to shared interests, tastes, orbeliefs is one way people create closeness(Parks & Floyd, 1996). Although the failureto make such a comment may prevent arelationship from getting closer, this missedopportunity may not by itself create a strainon the relationship. In contrast, excluding aperson from a gathering is one way peoplecreate distance (Kreilkamp, 1981). If a per-son seizes the opportunity to exclude some-one from a social gathering, the resultingostracism can certainly create a rift in theirrelationship. But the mere lack of exclusionfrom an event is unlikely to foster signifi-cant closeness in the relationship. Althoughmultiple forces are in action, it is appropri-ate to conceptualize the degree of relationalassociation as a single continuum becausethe forces combine to shape a single judg-ment of how close or distant the relation-ship is at that point in time.

Motion on the closeness-distance con-tinuum is governed by two forces—one that

198 J. A. Hess

Page 3: Measuring distance in personal relationships: The Relational Distance Index

pulls people toward closer relations and onethat pulls them toward more distant rela-tions. Existing psychometric research hasfocused almost exclusively on behaviorsthat foster closeness, but recent researchhas identified ways people may create dis-tance (Hess, 2000, 2002). These recent find-ings may offer the opportunity to create aninstrument to measure behaviors that leadto distance.

Measuring distance in relationships

Although observable behaviors form thestimulus for assessing relational distance,actors’ appraisals are an essential componentof distance judgments because psycho-logical distance is a perceptual phenom-enon. How close or distant a relationshipis perceived to be depends on the meaningan individual assigns to self and partnerinteraction patterns. For instance, percep-tions of distance can result not just fromwhat behavior was done or left undone,but from discrepancies between people’sexpectations or typical behaviors and thelevel of closeness actually enacted by theself or other (Helgeson et al., 1987). So,although trained observers could accuratelyrate the closeness or distance in people’sovert behaviors according to an externalstandard, they could not accurately capturethe subjectively felt distance between therelational partners, regarding either specificacts or the relationship as a whole. A lack oftouch between two people might be a sig-nificant distancing behavior in an instancewhere their relational history would maketouching the expected behavior, but mightnot create distance in the same context ifsuch behavior were atypical in their rela-tional history. For researchers wishing tounderstand the phenomenological experi-ence of relationships and relational events,and how perceived distance may be relatedto other relational characteristics—whetherthose other elements are perceptual (e.g.,relational satisfaction) or behavioral (e.g.,negativity in communication or psychoso-matic well-being) —a self-report measure ofdistance is necessary.

Researchers have often examined dis-tance through the use of a single holisticjudgment relying on the spatial nature ofthe metaphor. For example, researchershave measured perceived relational distancebymeasuring the distance between the place-ment of dolls (e.g., Gerber, 1973; Weinstein,1965) or blocks (e.g., Berry, Hurley, &Worthington, 1990) representing self andother(s). Other researchers (e.g., Nelson &Frantz, 1996) have simply used a singleLikert-type question asking participantshow close they felt to the relational partner.Although the use of a single holistic item is auseful way to measure distance, researchersfrequently favor the use of multiple-itemscales for psychometric purposes (Kline,1998). Thus, a multiple-item scale for meas-uring distance should be a useful tool forresearchers.

Scholars have created scales to examinedistance among social groups by askingquestions such as whether one wouldmarry a member of the other group orallow such a person to live in her or hishouse (e.g., Bethlehem, 1977; Singh, 1977).However, instruments tailored to personalrelationships are less readily available.Anderson and Sabatelli (1992) created adifferentiation in the family system (DIFS)scale. Although the DIFS scale does incorp-orate elements of distance, it is designed toassess whether family members develop anappropriate degree of differentiation thatwill allow them to balance both closenessand distance. Christensen and Sullaway(1984) created a communication patternsquestionnaire that contains a 3-item mutualavoidance subscale. Although mutualavoidance is one way that people can dis-tance themselves, this subscale is onlyintended to examine the ways partners dealwith problems. It neither addresses distan-cing tactics in other aspects of relationallife, including mundane interactions, norany of the many other ways people cancreate distance. So neither of these scalesprovides a satisfactory measure of distancein personal relationships.

A recent study (Hess, 2002) provides apotential foundation for a distance scale.

Relational Distance Index 199

Page 4: Measuring distance in personal relationships: The Relational Distance Index

That study was conducted to delineate theways people create distance and to examinethe conceptual structure of these distancingtactics. In it, an extensive list of distancingtactics (from Hess, 2000) was reduced forgreater parsimony, and the new list wasstructured based on functional similarityinto three overarching strategies for distan-cing: avoidance, disengagement, and cogni-tive dissociation. In both that study and thepresent one, strategies refer to general planspeople have for accomplishing certaingoals, and tactics refer to the specific behav-iors used to implement the strategies. Eachof the three distancing strategies has mul-tiple tactics:

1. Avoidance. Five tactics are avoidantin nature, meaning that they aredesigned to reduce the amount ofinteraction with the relational part-ner. These tactics are avoiding con-tact, ignoring the other person whenin her or his presence, reducing thelength of an interaction, interactingonly in the company of others, andspeaking less when conversing withthe other person.

2. Disengagement. There are seven wayspeople disengage themselves fromothers. When doing so, people com-municate with less personal connec-tion than they would otherwise. Thesebehaviors include avoiding intim-ate topics of conversation, deceivingthe other, being inattentive, reducingimmediacy, not joking with the other,treating the other impersonally, andmerely humoring the other.

3. Cognitive dissociation. Three tacticsfall into this category, which refersto changing perceptions about therelationships and the meaning ofpeople’s actions in effort to perceiveless association in the relationship.These tactics include disregarding theothers’ messages, mentally derogatingthe other, and feeling detachment.

It is plausible that this set of 15 tactics wouldmake good items for a scale that could be

used to assess distance in personal relation-ships. However, the previous study providedno data on the psychometric properties of ascale constructed from these items. Thisproject was undertaken to examine the reli-ability, factor structure, and validity of adistance scale based on these tactics.

Study 1: Reliability and Factor Structure

Research questions

The first study examined the reliability anddimensionality of a scale based on the15-item distance typology. Two researchquestions were asked:

RQ1: What is the factor structure of thedistancing items?

RQ2: What is the overall reliability of theinstrument, and the reliability of anysubscales that might be derived?

Method

Participants

The participants were 157 studentsrecruited from introductory communica-tion classes at the University of Missouri–Columbia. They ranged in age from 18 to24, with an average of 19 years. Forty-twopercent reported their sex as male, and 58%as female. Participants reported their raceas Caucasian (87%), African-American(6%), Asian (4%), Hispanic (2%), andmixed race (1%), and their year in schoolas freshman (37%), sophomore (32%),junior (17%), and senior (14%).

In the study, participants were asked toreport their behaviors in a specific affect-ively negative relationship. Fifty of the par-ticipants reported ongoing relationships.These relationships had an average dur-ation of 2.8 years and a median durationof 1.3 years. One hundred students reportedrelationships that had previously ended.These relationships had an average dur-ation of 2.6 years and a median duration

200 J. A. Hess

Page 5: Measuring distance in personal relationships: The Relational Distance Index

of 1.1 years. Seven students did not reportwhether they were still in the relationship.

Instrument construction

The items on the survey were derived fromthe Hess (2002) typology of distancing tac-tics. Because there was an uneven distribu-tion of items within the three categories(five avoidant, seven disengaging, andthree cognitive distancing tactics), two ofthe cognitive tactics were divided. This wasdone to better balance the categories and sothat if any of the cognitive items had to bedropped from the final scale, that categorycould still have enough items to be mean-ingfully represented. The two items dividedwere degradation and detachment, becausethese items encompassed the broadest con-ceptual diversity. Degradation was dividedinto one item reflecting degradation of char-acter and one item assessing the tendency toignore the other’s feelings. Detachment wasdivided into an item that reflected emo-tional detachment and an item focusing oncognitive detachment. In both of thesecases, the separation was based on the cen-tral elements that were described as com-prising the distancing tactic. The itemsfrom each category were systematicallyinterspersed to prevent the three types ofdistancing tactics from artificially factoringtogether due to placement order on thequestionnaire (see Table 1).

Procedure

Participants were asked to fill out the ques-tionnaire during class time, and were notoffered extra credit because the task tookjust a few minutes. Participants wereallowed to decline participation, althoughnone chose to opt out of the study. Therespondents were instructed to think of arelationship with someone they stronglydisliked, because such relationships aretypically characterized by a high degree ofdistance (Hess, 2000). Previous researchsuggests that these nonvoluntary relation-ships with disliked partners exist in a widevariety of social contexts, such as relation-ships with coworkers, relatives, teammates,

and mutual acquaintances of close friends.For the relationship in question, studentswere asked to rate how frequently they dideach of the distancing tactics in that rela-tionship on a 7-point scale (anchored by1 ¼ I never did this to 7 ¼ I did thisevery time possible).

Results

The two research questions asked about thereliability and factor structure of the dis-tance items. Factor analysis was done onSPSS using principal axis factoring andpromax oblique rotation. Although threefactors had eigenvalues greater than one,the third factor had only a single item andthe scree plot suggested that a two-factorsolution was more appropriate. Thus, thetwo-factor solution was selected (see Table 2).The first factor was labeled ‘‘unfriendly’’because the items on this factor largelysuggested negative or antisocial intentionstoward the other. Items on this factorincluded mentally degrading the other,ignoring the other person when in thesame location, lying to her or him, andtreating that person as a stranger. In con-trast, the second factor contained beha-viors more oriented toward escapinginteraction with the other person, so itwas labeled ‘‘withdrawal.’’ The items load-ing onto this factor included beingreserved, avoiding intimate topics of con-versation, avoidance, and nonimmediacy.The scale’s overall reliability was .88. Theitems on the unfriendly factor had a coef-ficient alpha of .86 and the items on thewithdrawal factor had a coefficient alphaof .74. The factor intercorrelation was .67,and the two factors accounted for 46% ofthe variance.

Although the factors that emerged fromthis analysis were interpretable, there wasconsiderable cross-loading. To create moredistinct factors, items that did not load wellwere eliminated, and then a factor analysiswas conducted again on the remainingitems. Although it was impossible to com-pletely eliminate cross-loading, an 8-itemversion of the scale displayed significantly

Relational Distance Index 201

Page 6: Measuring distance in personal relationships: The Relational Distance Index

improved factor clarity while retainingacceptable reliability (see Table 3). Theeight items provided an even representationof the two factors (four unfriendly and fourwithdrawal) and drew from all three of theoriginal conceptual strategies (four avoid-ance, two disengagement, and two cognitivedissociation). The shortened scale’s overallreliability was .80. The items on theunfriendly factor had a coefficient alpha of.78 and the items on the withdrawal factorhad a coefficient alpha of .73. The factorintercorrelation was .54, and these two fac-tors accounted for 59% of the variance.

Discussion

The data from this study show that theseitems had sufficient reliability and struc-tural clarity to be used as a scalar measureof distance in personal relationships.Because the items in the instrument measurethe behaviors that comprise distance(‘‘cause indicators’’) rather than behaviorsthat result from perceived distance (‘‘effectindicators’’) the instrument is more accu-rately termed an index than a scale(DeVellis, 1991). Thus, this measure can bereferred to as the Relational DistanceIndex (RDI).

Table 1. The distancing questionnaire

*1. When talking to ___, I kept the conversation away from topics thatwere intimate or personal.

*2. I changed my behavior to avoid encountering ___ whenever possible.3. When listening to ___, I disregarded her/his message or interpreted it in a

way to minimize its importance.4. I intentionally misled or lied to ___ on information about myself.

*5. When ___ was around me, I did not acknowledge her/his presence.*6. I mentally degraded ___, such as by seeing her/him as less than human,

less capable, or having fewer rights than others.7. When interacting with ___, I focused my attention away from her/him,

such as by paying attention to something else or just simply ‘‘zoning out.’’*8. When I was talking to ___, I would do things to make the interaction as

short as possible, such as pretending to agree or not asking questions.9. When interacting with ___, I avoided nonverbal cues that would

show closeness (e.g., I reduced the amount of smiling, touching,eye contact, etc.).

*10. I ignored ___’s thoughts, feelings, and intentions.11. I avoided doing things that I would normally do (e.g., joking

with someone) because they might have made ___ and myself closer.12. I tried to interact with ___ in the company of other people, rather

than interacting one-on-one.13. I tried to prevent myself from developing feelings about ___, or tried

to avoid acting on the negative feelings I already had.14. I treated ___ impersonally; that is, I treated her/him as I would

treat a stranger.*15. When in ___’s presence, I kept to myself and spoke less than

I would have if I liked her/him.16. I felt that ___ belonged to a different social group than I did.

*17. When ___ was speaking, I humored her/him and treatedher/him as being less capable of acting responsibly than other people.

Note. Items marked with an asterisk (*) are retained in the 8-item version. Strategy distribution from Hess (2002) is

as follows: avoidance (2, 5, 8, 12, 15), disengagement (1, 4, 7, 9, 11, 14, 17), cognitive dissociation (3, 6, 10, 13, 16).

202 J. A. Hess

Page 7: Measuring distance in personal relationships: The Relational Distance Index

Of the two versions (the 8-item and the17-item), the shorter measure is likely to bethe better choice for most researchers. Iteliminates some measurement redundancyand provides cleaner factor loadings thandoes the longer version, yet it still exhibitsadequate reliability. Furthermore, becauseit has fewer items, it is easier to administerin conjunction with other scales. The advan-tage of the 17-item version is that becausethe items were derived from a meaningfultaxonomy, rather than from a lengthy listmeant to be pared down, there is some con-ceptual value in retaining all the items. Itsfaithfulness to the existing construct, com-bined with a slightly higher reliability,would justify the use of the longer versionfor projects requiring a more detailed pic-ture of how people create distance in spe-cific relationships. However, researchersusing the longer version would be advisedto treat it as a single index and not docomputations with subscales because theitems in it do not separate as cleanly intothe two underlying factors.

Study 2: Factor Replication and Convergent

and Discriminant Validity

Hypotheses and research questions

The first study revealed a two-factor struc-ture of the RDI and provided evidence thatthe measure has good reliability. This studywas conducted to verify factor structurewith an independent sample and to assessconvergent and discriminant validity of theindex. Participants in Study 2 reported ontheir relationship with either a disliked or aliked relational partner, so as to assess theseissues within a broader range of relation-ships. The first hypothesis related to factorreplication:

H1: The items in the RDI fit a two-factorstructure of unfriendly and with-drawal tactics.

Validity was examined by administeringthe RDI in conjunction with other estab-lished measures. Because avoidance is one

Table 2. Factor loadings for the 17-item Relational Distance Indexwith 2-factor solution

Item Unfriendly Withdrawal

Ignore other’s feelings (10) .76 .53Inattention (7) .72 .66Mentally degrade the other (6) .72 .35Humoring the other (17) .68 .34Ignore the other (5) .65 .53Disregard message (3) .59 .42Deception (4) .52 .38Emotional detachment (13) .42 .41Cognitive detachment (16) .31 .20Being reserved (15) .47 .75Shorten interaction (8) .47 .68Nonimmediacy (9) .56 .68Impersonal (14) .62 .64Restraint (11) .54 .62Restrict topics (1) .26 .55Avoidance (2) .37 .54Interact in group (12) .17 .35

Note. Numbers in parentheses correspond to the item numbers in Table 1.

Relational Distance Index 203

Page 8: Measuring distance in personal relationships: The Relational Distance Index

of the common ways people create distance,it is reasonable to expect that avoidantconflict management would be associatedwith reports of distance. It is also reason-able to assume that reports of distance arepositively associated with people’s gestaltperception of distance in their relationship.In contrast, because distance is used tosignal the desire for decreased interactionwhereas affectionate communication isused to signal a desire for increasedcommunication, reports of distance shouldbe negatively correlated with reports ofliking and affectionate communication.Shortt and Gottman’s (1997) finding thatclose siblings show more positive affectthan do distant siblings provides supportfor this prediction. The prediction led totwo hypotheses:

H2: Subjective distance perception andavoidant conflict resolution are posi-tively associated with the RDI.

H3: Closeness, affectionate communica-tion, and liking are negatively asso-ciated with the RDI.

Because some of the items on the RDIare mildly antisocial (e.g., ignoring orhumoring the other), it was important toascertain that people’s reports were notbiased by a tendency to answer questionsin a socially desirable manner. Becausethe expected result was nonassociation,a research question was asked:

RQ3: Are responses on the RDI signifi-cantly correlated with the tendencyto answer questions in a sociallydesirable manner?

Previous research has indicated thatpeople report distancing themselves signifi-cantly more in affectively negative relation-ships than in affectively positive relationships(Hess, 2000). Therefore, another test of valid-ity is to compare reports of distance betweenthese ‘‘known groups.’’ The fourth hypoth-esis addressed this contrast.

H4: People in affectively negative relation-ships report greater use of distancingtactics than people in affectively posi-tive relationships.

Finally, it is pertinent to ascertain whetherany contextual or demographic factors relateto people’s reports of distancing. If peoplecan respond accurately, the current status ofthe relationship with either a liked or a dis-liked partner (ongoing or terminated) shouldnot impact people’s reports of distancingduring the relationship. And because dis-tance can be enacted immediately uponmeeting a person, the duration of the rela-tionship should also have little bearing onpeople’s reports of distance. Thus, anotherresearch question was asked:

RQ4: Are relationship status and rela-tionship duration associated withreports of distancing?

Table 3. Factor loadings for the 8-item Relational Distance Index

Item Unfriendly Withdrawal

Ignore other’s feelings .75 .46Mentally degrade the other .74 .31Ignore the other .66 .50Humoring the other .63 .29Being reserved .39 .72Shorten interaction .45 .63Restrict topics .23 .62Avoidance .33 .58

204 J. A. Hess

Page 9: Measuring distance in personal relationships: The Relational Distance Index

Method

Participants

Participants were 209 students recruitedfrom introductory communication classesat the University of Missouri–Columbia.They ranged in age from 18 to 42, with anaverage age of 20 years. Thirty-eight per-cent reported their sex as male, and 62%as female. Participants reported their raceas Caucasian (87%), African-American(8%), Asian (3%), Hispanic (1%), andmixed race (1%), and their year in schoolas freshman (24%), sophomore (43%),junior (19%), and senior (14%). Partici-pants were asked to report their behaviorsin a specific relationship. One hundredtwenty-nine of the participants reportedongoing relationships. These relationshipshad an average duration of 4.0 years and amedian duration of 2.0 years. Seventy-ninestudents reported relationships that hadpreviously been terminated. These relation-ships had an average duration of 2.4 yearsand a median duration of 2.0 years. Onestudent did not report whether he or shewas still in the relationship. Among therelationships that people reported, 61%were relationships arising from social net-work ties, 34% were romantic or formerlyromantic relationships, and 5% were familyrelations.

Instrument construction

The survey instrument began with the17-item RDI. For subjective distance per-ception a single item asked participants,‘‘How close do you think this relationshipwas/is?’’ This question was answered with a7-point scale anchored by very close andvery distant. Avoidant conflict resolutionstyle was measured with Christensen andSullaway’s (1984) 3-item mutual avoidancesubscale from their Communication Pat-terns Questionnaire.

Numerous instruments are available formeasuring closeness, liking, and affection-ate communication. To measure closeness,Ginsberg and Gottman’s (1986) 72-itemmeasure of closeness was selected. It was

chosen instead of Berscheid et al.’s (1989)Relational Closeness Inventory because theitems on Ginsberg and Gottman’s scalecould be generalized to a wider range ofrelational types. For affectionate communi-cation, two scales were chosen: Floyd andMorman’s (1998) 19-item AffectionateCommunication Index (ACI) and the6-item affection subscale of Foa and Foa’s(1974) Role Behavior Test. Finally, Rubin’s(1970, 1973) 9-item liking scale wasincluded to measure affection or disaffec-tion for the partner.

To examine whether people’s responseswere related to the tendency to answer insocially desirable ways, Crowne and Mar-lowe’s (1960) social desirability scale wasused. Relationship status and durationwere measured by asking participants toreport whether their relationship wasongoing or had been previously terminated,and asking them how long the relationshiphad lasted. The final questions on the sur-vey asked about the respondents’ demo-graphic characteristics.

Procedure

Students were offered extra credit in returnfor filling out the questionnaire. They wereread an announcement in class, then theysigned up for a research time. When theyarrived at the research site, they were giveninstructions and asked to fill out the ques-tionnaire. Participants were given as muchtime as needed. When they returned theircompleted survey to the researcher, the stu-dents were thanked for participation andgiven an information sheet explaining whatthe study was about.

Each of the constructs being measured inthis study was expected to score highly inconjunction with either liking or disliking ofthe relational partner (except for socialdesirability bias, which was expected to beunassociated with affect). Thus, people whorecalled affectively positive relationshipswere expected to report considerable affec-tive communication, but not much distan-cing or mutual avoidance. People whorecalled affectively negative relationships

Relational Distance Index 205

Page 10: Measuring distance in personal relationships: The Relational Distance Index

were expected to report the opposite. Thus,the instructions on the questionnaire variedas to whether they asked participants toanswer the questions with regard to astrongly liked or a strongly disliked partner.Which version each participant receivedwas assigned at random. One hundred fourparticipants responded to questions abouta liked partner, and 105 respondentsanswered with regard to someone theydisliked.

Although the items on the RDI weredeveloped to be applicable to both personaland work relationships, participants in thisstudy were asked to recall a personal rela-tionship. This was done because some ofthe other scales in the study (e.g., Floyd &Morman, 1998; Foa & Foa, 1974) weredeveloped primarily for the study ofpersonal rela-tionships. Restricting the rela-tionship to a personal relationship preventedfalsely reduced levels of closeness beingreported on those scales where workplacerole definitions might limit affective expres-sion (e.g., kissing on the lips, givingmassages).

Results

Hypothesis 1

The first hypothesis asserted that the factorstructure from Study 1 would be replicatedin an independent sample. Because previoussamples have indicated that people distancethemselves only minimally in affectivelypositive relationships, it is probable thatthe decreased variability in this half of thedata set would mask existing factor struc-ture. Thus, factor replication was testedonly with the data from affectively negativerelationships (matching the data fromStudy 1, which relied exclusively on affec-tively negative relationships). Althoughparticipants filled out all 17 distance itemsfor Study 2, only the items retained in theshort version were used for the factor ana-lysis. Factor structure was examined withconfirmatory factor analysis, using LIS-REL 8.50. The results showed that thefactor structure did replicate that foundin Study 1 (see Figure 1). The factor inter-

correlation was .64, and all diagnostics fellwithin appropriate parameters: w2(26,N¼ 209) ¼ 12.56, p¼ .86, RMSEA< .001,CFI¼ 1.00, RMSR¼ .045. All of theseindices point to an appropriate fit betweenthe model and the data (Hu & Bentler,1999).

To determine whether any factor struc-ture was evident in the affectively positiverelationships, an exploratory factor analysiswas run on the remaining half of the dataset using principal axis factoring and pro-max rotation as was done in Study 1. Forthe complete set of 17 items, five factorswere extracted, although the first factorhad an eigenvalue of 5.0 and the otherfour had eigenvalues below 1.50. These fac-tors bore some resemblance to theunfriendly withdrawal structure shown inthe data with disliked partners, althoughinterpretation was not perfectly clear. How-ever, when the 8-item index was entered forliked partners, a two-factor solutionemerged that closely matched the structurewith disliked partners. Mentally degradingthe other moved from the unfriendly factorto the withdrawal factor, but all the otheritems loaded onto the same factors. Aswould be expected for data with low vari-ance, many of the items cross-loaded. Theseresults offer some support for the notionthat the same patterns exist in people’sreported use of distancing tactics withrelational partners they like, just that thedifferences are much smaller in magnitude.

Hypotheses 2 and 3

H2 and H3 predicted high positive or highnegative correlations with other measures.The reliabilities for all scales were good. Formost of the scales, reported behaviors wereexpected to be very low in one of the condi-tions (e.g., affectionate communicationshould be minimal in relationships with dis-liked partners). This clustering of responsesaround one end of the scale should increasereliability in those conditions. Thus, reli-abilities were computed for both conditionsas well as overall. Three reliabilities arelisted for each of the scales: (i) with disliked

206 J. A. Hess

Page 11: Measuring distance in personal relationships: The Relational Distance Index

partner, (ii) with liked partner, and (iii)across all respondents. Coefficient alphaswere as follows: 8-item RDI (.75, .73, .92),4-item unfriendly subscale (.76, .56, .85),4-item withdrawal subscale (.71, .67, .92),17-item RDI (.84, .82, .95), 3-item mutualavoidance measure (.60, .77, .85), 72-itemcloseness inventory (.98, .95, .99), 19-itemaffectionate communication index (.95,

.91, .97), 9-item liking scale (.88, .89, .96),and 6-item affective role behavior subscale(.90, .92, .97).

The hypotheses predicted that the RDIwould vary positively with other measuresof distance or avoidance and vary nega-tively with measures of closeness or positiveaffect. As shown in Table 4, the datashowed that the RDI was positively

Ignorefeelings

Mentallydegrade

Humorthe other

Ignore theperson

Beingreserved

Restricttopics

Avoid theperson

Shorteninteraction

Unfriendly Withdrawal

1.00 .96 .83 .78 .80 .72 1.00 .98

.64

Figure 1. Confirmatory factor analysis of the 8-item Relational Distance Index.

Table 4. Correlations among the 8-item and 17-item RDIs and other scales (studentrespondents)

Distance (RDI)Unfriendlysubscale

Withdrawalsubscale

Subjectively judged distance .74** .59** .76**.77**

Mutual avoidance .75** .63** .75**(Christensen & Sullaway, 1984) .77**

Closeness �.80** �.65** �.81**(Ginsberg & Gottman, 1986) �.83**

Affectionate communication �.74** �.62** �.75**(Floyd & Morman, 1998) �.76**

Affective role behavior �.76** �.63** �.77**(Foa & Foa, 1974) �.79**

Liking �.82** �.71** �.80**(Rubin, 1970, 1973) �.83**

Social desirability bias .05 �.06 .12(Crowne & Marlowe, 1960) .07

Note. At each entry the top number refers to the 8-item version of the RDI and the bottom number refers to the

17-item version.

**p< .01.

Relational Distance Index 207

Page 12: Measuring distance in personal relationships: The Relational Distance Index

correlated with subjective distance percep-tion (.74) and mutual avoidance (.75) aspredicted in H2; and negatively correlatedwith the scales measuring closeness (�.79),affectionate communication (�.74), affec-tive role behavior (�.76), and liking (�.82)as predicted in H3. Thus, both hypotheseswere supported.

Research question 3

RQ3 asked whether the RDI covaried withtendency to answer questions in a sociallydesirable manner. The 33-item social desir-ability scale showed sufficient reliability (.79,.74, .77), and the RDI was unassociated withit (r¼ .05, p¼ .51; see Table 4). These resultssuggest that people’s reports of distance arenot associated with a tendency to reportbehavior in a socially desirable manner.

Hypothesis 4

The final hypothesis predicted that peoplein affectively negative relationships wouldreport distancing themselves from their part-ner significantly more than would peoplein affectively positive relationships. Thishypothesis was supported, using the 8-itemRDI, F(1, 205) ¼ 502.56, p< .001, Z2¼ .71.

Research question 4

RQ4 asked whether relational status(ongoing versus terminated) and durationaffected people’s reports of distancing. Bothof these questions were examined withineach condition because the differences acrossconditions might create artificial effects (ifthere was an uneven distribution of the inde-pendent variable across conditions). Usingthe 8-item RDI, relational status had nobearing on reports of distancing for eitherdisliked partners, F(1, 101)¼ .60, p¼ .44, orfor liked partners, F(1, 101) ¼ .67, p¼ .41.Likewise, the length that the relationshiphad lasted was not significantly correlatedwith reports of distancing for either dislikedpartners, r ¼ �.01, p¼ .94, or for liked part-ners, r¼ .11, p¼ .29. These results suggestthat relational parameters are not associatedwith people’s reports of distance.

Discussion

The results of this study lend support to thehypothesized factor structure of the RDIand to the claim that it actually taps into theconstruct it was designed to measure. Thesuccessful confirmatory factor analysis andthe appropriate correlations with associatedmeasures suggest that it does indeed measurewhat theory suggests it does. Furthermore,this data set included affectively positive rela-tionships as well as affectively negative rela-tionships, which Study 1 had used exclusively.So, these data suggest that the RDI has theexpected association with other scales regard-less of affect for partner.

Study 3: Temporal Stability

Hypothesis

Many theories of relationship processessuggest that distance should vary over timebecause dynamics influencing closeness anddistance are constantly in flux, such asthrough dialogic fluidity (e.g., Baxter &Montgomery, 1996), adjustments to main-tain equilibrium or congruence (e.g., Floyd,1999), or as part of recalibration in a system(e.g., VanLear, 1991). However, in theabsence of significant relational turningpoints, it is reasonable to assume thatchanges in distance would be moderateacross short intervals of time. Thus, if meas-ured at two points in close temporal prox-imity, reports of distance should bepositively correlated. This led to the follow-ing hypothesis:

H5: People’s reports of distance in thesame relationship are positivelycorrelated over short intervals oftime.

Method

Participant

Participants were 75 students recruited fromintroductory communication classes at the

208 J. A. Hess

Page 13: Measuring distance in personal relationships: The Relational Distance Index

University of Missouri-Columbia. Theywere offered extra credit in return for fillingout two questionnaires. They ranged inage from 18 to 23 years, with an averageage of 19. Thirty-nine percent reportedtheir sex as male, and 61% as female. Parti-cipants reported their race as Caucasian(97%), African-American (1%), and Asian(1%).

Procedure

An announcement was read in class, thenstudents signed up for a research time.When they arrived at the research site,they were given instructions and asked tofill out the questionnaire about their rela-tionship with a specific individual. Theywere also given a sheet to write down thename of the person they wrote about. Par-ticipants were given as much time asneeded, and then they returned their com-pleted survey to the researcher. Three weekslater, a second announcement in classreminded students to return to the researchsite to fill out a second questionnaire. Whenthe participants returned, they were giventhe second questionnaire and their originalreminder sheet so they accurately recalledthe person they answered questions about.

In effort to have participants respondabout relationships they maintained throughfrequent interactions, and at the same timehave them respond about relationships inwhich they might reasonably be expected todo at least moderate amounts of distancing,students were asked to think about theirrelationship with their supervisor at work.Such relationships could reasonably beexpected to display a variety of interactivequalities. If students did not maintain a jobor did not interact regularly with their super-visor, they were instructed to think abouttheir roommate. If they had neither a worksupervisor nor a roommate, they were askedto think about the person with whom theyinteracted most frequently. As a result, 38%of the respondents wrote about a supervisor,56% wrote about a roommate, and 6%wrote about another relationship.

Results

Hypothesis 5

H5 predicted a significant correlationbetween reported distancing at the twotimes students responded to the study. Ofthe 75 participants who filled out the firstsurvey, 64 (85%) returned to fill out thesecond survey. Reliability was acceptablein all conditions. At times one and tworespectively, reliability was .70 and .72with the 8-item version and .87 and .86with all 17 items. For the 8-item RDI, thecorrelation between measurements wasr¼ .68 (p< .01); for the 17-item RDI, thecorrelation was r¼ .75 (p< .01). Thus,H5 was supported. Corrected for attenu-ation (using the double correction,Muchinsky, 1996), these correlations roseto .96 and .87 respectively.

Discussion

These correlations suggest that the RDIexhibits good temporal stability. Assumingthat distance exhibits temporal fluidity,then we would expect some variation inresponses three weeks after the initial sur-vey. Thus, the correlations found in thisstudy fall within the expected range.

Study 4: External Validity

Hypotheses and research question

The data from the first three studies suggestthat the RDI is a reliable and valid index formeasuring people’s use of distancing tacticsinpersonal relationships.Becausethosestud-ies relied exclusively on college students,however, the possibility exists that theresponses are unique to that population.To determine whether the RDI can beused more widely, the same validation sur-vey used in Study 2 was administered to anonstudent sample. In this study, individ-uals who were married or divorcedresponded to a questionnaire about theirrelationship with their spouse or ex-spouse.

Relational Distance Index 209

Page 14: Measuring distance in personal relationships: The Relational Distance Index

The hypotheses and research question werethe same as those in Study 2:

H6: Subjective distance perception andavoidance are positively associatedwith the RDI.

H7: Closeness, affectionate communica-tion, and liking are negatively asso-ciated with the RDI.

RQ5: Are responses on the RDI signifi-cantly correlated with the tendencyto answer questions in a sociallydesirable manner?

Method

Participants

Students from communication classes wereasked to find one married or divorced per-son to fill out a questionnaire. The 100 par-ticipants thus obtained ranged in age from23 to 63 years, with a mean age of 46 and amedian age of 48. Ninety-three percent ofthe respondents reported on a relationshipwith their current spouse, and 78% of therespondents were female. They reportedtheir race as Caucasian (95%), Asian (4%),and African-American (1%). Among therespondents who were married, the lengthof marriage ranged from 5 months to 36years, with a mean 22 years and a median of25 years. Among those who were divorced,the length of time since divorce ranged from3 to 14 years, with a mean and a median of10 years.

Procedure

The questionnaire was the same as that usedin Study 2. For people who were divorced,the respondent had to communicate withtheir ex-spouse at least on a monthly basisin order to be eligible to participate. It wasexpected that participants would reportgreater distance with ex-spouses than withcurrent partners. Unfortunately, becauseonly 7% of the respondents reported on

their ex-spouse, this method was less effect-ive in adding variance to the data thanoriginally hoped (although there was stillconsiderable variance in distance reportedamong respondents who answered ques-tions about their spouse). After participantsfilled out the survey, they were instructed toseal their survey in a provided envelope andeither mail it back to the researcher or giveit to the student, who would return it. Onceall responses were returned, a random sam-ple of 10% were called to verify participa-tion. In all cases, the respondent indicatedthat he or she had indeed been the personwho completed the survey, so these phonecalls served as an opportunity for theresearcher to verbally thank these respond-ents for their participation.

Results

Hypotheses 6 and 7

The hypotheses in this study predicted thesame patterns of association as seen inStudy 2. As in Study 2, the reliabilitieswere very high. Because the majority ofthese respondents reported on marriages inwhich affect was positive, there was lessvariance in the data than there might be inother samples, and this likely increasedreliability in many of the scales. Coefficientalphas for the measures were as follows: .84for the 8-item RDI, .78 for the 4-itemunfriendly subscale, .80 for the 4-item with-drawal subscale, .93 for the 17-item RDI,.92 for the 3-item mutual avoidance meas-ure, .99 for the 72-item closeness inventory,.96 for the 19-item affectionate communica-tion index, .93 for the 9-item liking scale,and .96 for the 6-item affective role behav-ior subscale.

As shown on Table 5, the data showedthat the 8-item RDI was positively correl-ated with subjective distance perception(.72) and mutual avoidance (.58) as pre-dicted in H6, and was negatively correlatedwith the scales measuring closeness (�.67),affectionate communication (�.62), affect-ive role behavior (�.71), and liking (�.59)

210 J. A. Hess

Page 15: Measuring distance in personal relationships: The Relational Distance Index

as predicted in H7. Thus, both hypotheseswere supported.

Research question 5

RQ5 asked whether reports of distance wereassociated with a tendency to answer in asocially desirable manner. The 33-itemsocial desirability scale had a reliability of.82, and it was not significantly associatedwith reports of distance (r¼ .03, p¼ .81; seeTable 5). This suggests that the two con-structs are unassociated.

Discussion

The results of this study provide some evi-dence that the RDI is a valid measure ofrelational distance when used with samplesthat are heterogeneous in terms of age andmarital status. Thus, it appears to be anappropriate measure to use for researchwith nonstudent populations, or for assess-ment in married couples.

Most of the patterns of associationfound in this study are smaller in magnitudethan those recorded in Study 2. One plaus-ible explanation for this difference is thatlong-term marriages or relations withex-spouses may lead to ‘‘messier’’ patternsof interaction. That is, it is easy to be con-sistently close to a dating partner or a goodfriend in a voluntary association, and it iseasy to be consistently distant with a dis-liked partner in a nonvoluntary relation-ship. However, it is harder to beconsistently close in a long-term ongoingmarriage or extremely distant with a formerspouse, particularly when children and rela-tives are involved. It is also more likely thatindividuals adopt more realistic perceptionsof their partner (acknowledging both theirpartner’s positive attributes and their part-ner’s flaws) over the course of a long-termmarital relationship. Thus, people may judi-ciously distance themselves in the best ofmarriages, may be closer than preferredwith ex-spouses, and may have mixed

Table 5. Correlations among the 8-item and 17-item RDI and other scales (married ordivorced respondents)

Distance(RDI)

Antisocialsubscale

Withdrawalsubscale

Subjectively judged distance .72** .59** .70**.76**

Mutual avoidance .58** .47** .56**(Christensen & Sullaway, 1984) .55**

Closeness �.67** �.50** �.67**(Ginsberg & Gottman, 1986) �.69**

Affectionate communication �.62** �.49** �.62**(Floyd & Morman, 1998) �.65**

Affective role behavior �.71** �.60** �.68**(Foa & Foa, 1974) �.72**

Liking �.59** �.45** �.59**(Rubin, 1970, 1973) �.60**

Social desirability bias .03 �.11 .10(Crowne & Marlowe, 1960) �.02

Note. At each entry the top number refers to the 8-item version of the RDI and the bottom number refers to the

17-item version.

**p< .01.

Relational Distance Index 211

Page 16: Measuring distance in personal relationships: The Relational Distance Index

perceptions of the partner in any marriageor divorce. These ‘‘messier’’ interactionsand perceptions would likely introducemore variance unaccounted for by the mea-sures used in this study and thus reduce thecorrelations among them.

General Discussion

This study was undertaken to develop aninstrument capable of measuring people’sperceptions of distancing tactics used in apersonal relationship. Based on a recentlyderived typology, an index was developedto assess relational distance. In both 8- and17-item form, this index exhibits good reli-ability and performs well in numerous testsof validity. The 8-item version also has aclear and stable two-dimensional factorstructure.

The results of the factor analyses meritattention. Notably, they reveal differentstructural relations than were reported inthe previous study (Hess, 2002) fromwhich the items were derived. This differ-ence suggests an interesting conclusion.These two investigations relied on differenttypes of data. Data from the earlier researchwere based on conceptual similarity amongitems, whereas the data in the present stud-ies were based on reported frequency ofuse. Although it is possible that both typesof data could provide similar structures,such a result is by no means expected. Theconceptual similarity data from the earlierresearch show which tactics people see assimilar to each other based on features ofthe tactics themselves, such as whether theyinvolve overt actions directed at the partnerversus patterns of thought regarding thepartner. The frequency of use data fromthe present study show which tactics peopleuse in conjunction, and thus offer insightinto the motivation force controllingpeople’s use of distancing tactics. The differ-ence between these two data sets indicatesthat selection of distancing strategies is notbased solely on feature or functionalequivalence, it is also based on more globalsimilarities in the direction or intensity ofdistancing tactics. The factor analysis shows

that people either attempt to withdraw fromthe disliked person or they interact moremalevolently toward the other. The respon-dents’ tactical choices seem to be governedby a ‘‘fight or flight’’ scheme that is consist-ent with psychological approach-avoidancetheories of social behavior (e.g., Frijda,1986; Mehrabian, 1972).

The factor structure found in the presentstudy further clarifies the previouslyreported cluster structure of distancingbehaviors (Hess, 2000). In the previousexploration of relationships with dislikedpartners, three distancing profiles were sug-gested based on frequency of use: avoidinginvolvement, expressing detachment, andshowing antagonism. All but one of theitems on the avoiding involvement profileended up in the withdrawal factor in thepresent study, suggesting that this profile islargely composed of behaviors intended towithdraw from interactions. The items fromthe other two profiles in the previous study,expressing detachment and showing antag-onism, fell into the unfriendly factor in thisstudy, suggesting that those two profileswere comprised of behaviors that weremore overtly expressive of desire for dis-tance. The antagonistic profile in the pre-vious study included the most hostiledistancing tactics (e.g., rejection, dehuman-ization, explicitly stating desire for distance)as well as a hostility item that was antag-onistic, not just distancing. Consequently,that profile seems more venomous than theexpressing detachment profile. The latterprofile, although composed of behaviorsthat are more expressive than withdrawing,is more innocuous than overt hostility. Sothe results of this project suggest that theprofiles reported in the previous article maybe defined by their level of expressiveness(explicitly versus implicitly showing desirefor distance) and their level of hostility(antagonistic versus innocuous).

One interesting point not addressed inthis project that merits attention in futureresearch is the subject of analysis. In thisproject, only people’s perceptions of theirown distancing behaviors were studied.However, there are many other fruitful

212 J. A. Hess

Page 17: Measuring distance in personal relationships: The Relational Distance Index

avenues of inquiry regarding distance in arelationship. Each relational partner holdsperceptions not only of her or his ownefforts at distance regulation, but also ofthe partner’s. How closely matched arepartners’ perceptions of each other’s distan-cing tactics? Under what circumstances arepeople more accurate or less accurate atperceiving their partners’ intentions? Aresome people more skilled than others atdiscerning their partner’s distancingattempts? Kenny and his colleagues (e.g.,Kenny, Bond, Mohr, & Horn, 1996) havefound that people are reasonably accurateat judging others’ level of affect. Given theassociation between distance and negativeaffect, it is reasonable to speculate thatpeople might be fairly accurate, but at pre-sent there are no data to support this con-jecture. Another venue of exploration is toexamine interdependence between partners’distancing tactics. To what degree is theuse of distancing tactics by one personresponded to with a closeness-inducing or adistancing tactic by her or his partner? Lon-gitudinal research could shed light on thissituation. Research on marital satisfaction(e.g., Gottman, 1994) indicates that dissatis-fied couples are more likely to reciprocatenegativity than are satisfied couples. Whatrole does distance play in this process?

The fact that this index is based on sub-jects’ perceptions of distance instead ofresearchers’ observations of the enactedbehaviors is both a strength and limitation.The advantage of using self-reports is thatthey incorporate the meanings that peopleconstruct in their relationships. Becauseperceptions of distance result from people’sinterpretations rather than the behaviorsthemselves, a self-report instrument inher-ently includes information that observa-tional techniques cannot access. Thelimitation is that a person’s perceptionsmay not match the external reality of thesituation, so the assessment of felt distanceshould not be equated with the observablecommunication patterns between the twopartners. Use of stimulated recall would beoneway to account for both observed behav-ior and relational meaning, but such a small

sample of a person’s interaction would beunlikely to reveal all the different dimen-sions of distance within a relationship,especially because the sample would likelybe taped in a laboratory situation. So theRDI does provide important data thatcannot be measured in other ways.

The RDI offers the potential to facilitateadvances in relational theory and interven-tion. A reliable measure of distance can helpresearchers determine how distance relatesto other relational dynamics, such as con-flict, relational dissolution, or improvementof strained relations. Further research mayilluminate whether different causes promotedifferent types of distancing efforts. Forexample, it is plausible that distancing dueto strong negative affect will lead to moreantagonistic distancing behaviors than willdistancing when affect is positive, thoughfactors such as power discrepancies or cul-tural norms might moderate such a pattern.When distancing is motivated by forcesother then disliking (e.g., by a need for reju-venation or due to impending deadlines)then withdrawal may be the primarymechanism people use to create space.Researchers may also be able to learnmore about the relationship between distan-cing strategies and various outcomes. Forinstance, it may be possible to determinewhether certain distancing behaviors canincrease or decrease the chances that rela-tional partners will successfully resolve aconflict. And information about the actor’sdistancing behavior is essential to answer-ing questions such as whether the ability todistance oneself using particular tactics canreduce a person’s chances of committingviolent behaviors in an intimate relation-ship, as some therapists have suggested(e.g., Bartle & Rosen, 1994).

If one accepts that closeness-distance isone of the most fundamental of all relationalqualities, then a relational distance measureis an important tool for researchers studyingpersonal relationships. The RDI is one meas-ure to assess distance, and for that reason itseems to offer promise for helping to developa better understanding of the processes thatgovern personal relationships.

Relational Distance Index 213

Page 18: Measuring distance in personal relationships: The Relational Distance Index

References

Anderson, S. A., & Sabatelli, R. M. (1992). The differ-entiation in the family system scale (DIFS).American Journal of Family Therapy, 20, 77–89.

Bartle, S. E., & Rosen, K. (1994). Individuation andrelationship violence. American Journal of FamilyTherapy, 22, 222–236.

Baxter, L. A., & Montgomery, B. M. (1996). Relating:Dialogues and dialectics. New York: Guilford.

Berry, J. T., Hurley, J. H., & Worthington, E. L., Jr.(1990). Empirical validity of the Kvebaek FamilySculpture Technique. American Journal of FamilyTherapy, 18, 19–31.

Berscheid, E., Snyder, M., & Omoto, A. (1989). Therelationship closeness inventory: Assessing thecloseness of interpersonal relationships. Journal ofPersonality and Social Psychology, 57, 792–807.

Bethlehem, D. W. (1977). Validity of social distancescales: A Zambian study. Journal of Social Psycho-logy, 101, 157–158.

Christensen, A., & Sullaway, M. (1984). Communica-tion patterns questionnaire. Unpublished manu-script, Department of Psychology, University ofCalifornia, Los Angeles.

Crowne, D. P., & Marlowe, D. (1960). A new scale ofsocial desirability independent of psychopathology.Journal of Consulting and Clinical Psychology, 24,349–354.

DeVellis, R. F. (1991). Scale development: Theory andapplications. Newbury Park, CA: Sage.

Floyd, K. (1999). To match or not to match: Effects ofbehavioral congruence on interpersonal connected-ness. Journal of Social Psychology, 139, 309–322.

Floyd, K., & Morman, M. T. (1998). The measure ofaffectionate communication. Communication Quar-terly, 46, 144–162.

Foa, U. G., & Foa, E. B. (1974). Societal structures ofthe mind. Springfield, IL: Thomas.

Frijda, N. H. (1986). The emotions. Cambridge, UK:Cambridge University Press.

Gerber, G. L. (1973). Psychological distance in familyas schematized by families of normal, disturbed,and learning-problem children. Journal of Consult-ing and Clinical Psychology, 40, 139–147.

Ginsberg, D., & Gottman, J. M. (1986). Conversationsof college roommates: Similarities and differences inmale and female friendships. In J. M. Gottman &J. G. Parker (Eds.), Conversations of friends:Speculations on affective development (pp. 241–291). Cambridge, UK: Cambridge UniversityPress.

Gottman, J. M. (1994). What predicts divorce? Therelationship between marital processes and maritaloutcomes. Hillsdale, NJ: Lawrence Erlbaum.

Helgeson, V. S., Shaver, P., & Dyer, M. (1987).Prototypes of intimacy and distance in same-sexand opposite-sex relationships. Journal of Socialand Personal Relationships, 4, 195–233.

Hess, J. A. (2000). Maintaining nonvoluntary relation-ships with disliked partners: An investigation intothe use of distancing behaviors. Human Communi-cation Research, 26, 458–488.

Hess, J. A. (2002). Distance regulation in personal rela-tionships: The development of a conceptual modeland a test of representational validity. Journal ofSocial and Personal Relationships, 19, 663–683.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fitindexes in covariance structure analysis: Convent-ional criteria versus new alternatives. StructuralEquation Modeling, 6, 1–55.

Jacobson, N. S. (1989). The politics of intimacy. Behav-ior Therapist, 12, 29–32.

Kantor, D., & Lehr, W. (1975). Inside the family. SanFrancisco: Jossey-Bass.

Kelley, H. H., Berscheid, E., Christensen, A., Harvey,J. H., Huston, T. L., Levinger, G., McClintock, E.,Peplau, L. A., Peterson, D. R. (1983). Close rela-tionships. New York: W. H. Freeman.

Kenny, D. A., Bond, C. F., Jr., Mohr, C. D., &Horn, E. M. (1996). Do we know how much peoplelike one another? Journal of Personality and SocialPsychology, 71, 928–936.

Kline, R. B. (1998). Principles and practice of structuralequation modeling. New York: Guilford.

Kreilkamp, T. (1981). Psychological distance. In J. deRivera (Ed.), Conceptual encounter: A method forthe exploration of human experience (pp. 273–342).Washington, DC: University Press of America.

Mehrabian, A. (1972). Nonverbal communication.Chicago: Aldine.

Muchinsky, P. M. (1996). The correction for attenu-ation. Educational and Psychological Measurement,56, 63–75.

Napier, A. Y. (1988). The fragile bond: In search of anequal, intimate, and enduring marriage. New York:Harper & Row.

Nelson, B. J., & Frantz, T. T. (1996). Family inter-actions of suicide survivors and survivors of non-suicidal death. Omega: Journal of Death and Dying,33, 131–146.

Parks, M. R., & Floyd, K. (1996). Meanings for close-ness and intimacy in friendship. Journal of Socialand Personal Relationships, 13, 85–107.

Rubin, Z. (1970). Measurement of romantic love.Journal of Personality and Social Psychology, 16,265–273.

Rubin, Z. (1973). Liking and loving: An invitation tosocial psychology. New York: Holt, Rinehart, &Winston.

Ryder, R. G., & Bartle, S. (1991). Boundaries as dis-tance regulators in personal relationships. FamilyProcess, 30, 393–406.

Shortt, J. W., & Gottman, J. M. (1997). Closeness inyoung adult sibling relationships: Affective andphysiological processes. Social Development, 6,142–164.

Singh, N. N. (1977). Social distance between Fijiansand Indians in Fiji after four years of independ-ence. Journal of Social Psychology, 101, 147–148.

Spitzberg, B. H., & Cupach, W. R. (Eds). (1998).The dark side of close relationships. Mahwah, NJ:Lawrence Erlbaum.

Sullaway, M., & Christensen, A. (1983). Assessment ofdysfunctional interaction patterns in couples. Jour-nal of Marriage and the Family, 45, 653–660.

Vangelisti, A. L., & Maguire, K. (2002). Hurtful mes-sages in family relationships: When the painlingers. In J. H. Harvey & A. Wenzel (Eds.),A clinician’s guide to maintaining and enhancingclose relationships (pp. 43–62). Mahwah, NJ:Lawrence Erlbaum.

214 J. A. Hess

Page 19: Measuring distance in personal relationships: The Relational Distance Index

VanLear, C. A. (1991). Testing a cyclical model ofcommunicative openness in relationship develop-ment: Two longitudinal studies. CommunicationMonographs, 58, 337–361.

Weinstein, L. (1965). Social schemata of emotionallydisturbed boys. Journal of Abnormal Psychology,70, 457–461.

Wright, P. H., & Wright, K. D. (1995). Codependency:Personality syndrome or relational process? In.S. Duck & J. T. Wood (Eds.), Confronting relation-ship challenges (pp. 109–128). Thousand Oaks,CA: Sage.

Relational Distance Index 215