Journal of Personality and Social Psychology1980, Vol. 39, No. 3, 472-480

The Revised UCLA Loneliness Scale: Concurrent andDiscriminant Validity Evidence

Dan RussellUniversity of Iowa

Letitia A. Peplau and Carolyn E. CutronaUniversity of California, Los Angeles

The development of an adequate assessment instrument is a necessary pre-requisite for social psychological research on loneliness. Two studies providemethodological refinement in the measurement of loneliness. Study 1 presents arevised version of the self-report UCLA (University of California, Los Angeles)Loneliness Scale, designed to counter the possible effects of response bias in theoriginal scale, and reports concurrent validity evidence for the revised measure.Study 2 demonstrates that although loneliness is correlated with measures ofnegative affect, social risk taking, and affiliative tendencies, it is nonetheless adistinct psychological experience.

Social relationships are at the core of hu-man life. Not surprisingly, problematic as-pects of relationships have been a majorfocus of psychological research. Psychologistshave undertaken careful analyses of aggres-sion, competition, crowding, and other nega-tive factors in social relations. Some problemsof social relations have, however, been empha-sized to the neglect of others. Researchershave investigated instances where there are"too many" people, and individuals feel sub-jectively "crowded" (e.g., Freedman, 1975;Stokols, 1972). However, little attention hasbeen given to the other end of the continuumwhere social relationships are "too few," andpeople feel subjectively "lonely."

The importance of research on lonelinesslies not only in its potential for shedding lighton basic aspects of social relations but also in

The authors would like to thank Robin Lewis,Jeff Tanaka, Debbie McGrew, and Phillip Malamuthfor their help in collecting the data for Study 1;Warren H. Jones for providing subjects from theUniversity of Tulsa for Study 2 ; and Richard R.Lau for his comments on the manuscript and helpwith data analysis.

Requests for reprints should be sent to Dan Rus-sell, Division of Educational Psychology, Measure-ment, and Statistics, College of Education, Universityof Iowa, Iowa City, Iowa S2242.

the fact that loneliness is a common and dis-tressing problem for many people. In onenational survey (Bradburn, 1969), 26% ofAmericans reported having felt "very lonelyor remote from other people" during the pastfew weeks. Loneliness has been linked to avariety of other serious individual and socialproblems, including alcoholism (Nerviano &Gross, 1976), adolescent delinquent behavior(Brennan & Auslander, Note 1), suicide(Jacobs, 1971; Wenz, 1977), and physicalillness and overutilization of health care ser-vices (Lynch, 1976).

Empirical research on loneliness has beenhampered by a variety of problems (see re-views by Peplau & Perlman, 1979; Peplau,Russell, & Heim, 1978). A major hindranceis that loneliness, unlike aggression, competi-tion, and crowding, cannot be readily manipu-lated by researchers. Thus, the crucial taskfor investigators is not the development of anexperimental paradigm to produce lonelinessin differing degrees under controlled condi-tions but rather the development of instru-ments to detect variations in loneliness thatoccur in everyday life.

Our research on loneliness led initially tothe development of a 20-item, self-report mea-sure, the UCLA (University of California,Los Angeles) Loneliness Scale (Russell, Pep-

Copyright 1980 by the American Psychological Association, Tnc. 0022-3514/80/3903-0472$00.75

472

THE REVISED UCLA LONELINESS SCALE 473

lau, & Ferguson, 1978). In studies usingcollege samples, the scale showed high internalconsistency (coefficient alpha of .96). Con-current validity was indicated by relation-ships between scores on the loneliness scaleand other indicators of loneliness, social rela-tionships, and affective states.

Research by several other investigators hasalso supported the adequacy of the UCLALoneliness Scale as a measure and has begunto provide a more detailed description of theexperience of loneliness. Loneliness is relatedto a number of personal characteristics, in-cluding low self-esteem, shyness, feelings ofalienation, external locus of control, and beliefthat the world is not a just place (Jones,Freemon, & Goswick, in press). Lonely stu-dents report experiencing problems of inhib-ited sociability (Horowitz & French, 1979)and, in dyadic interactions, rate both them-selves and their partners more negativelythan do nonlonely students (Jones et al., inpress). Among both students and older adults,loneliness is linked to negative affects, includ-ing boredom, restlessness, and unhappiness,and to dissatisfaction with social relationships(Perlman, Gerson, & Spinner, 1978; Russellet al., 1978).

Although the UCLA Loneliness Scale is areasonably adequate measure, several poten-tial problems with the scale are apparent.First, all items on the scale are worded in thesame direction, with high scores reflectingfeelings of social dissatisfaction. Any sys-tematic response bias toward high or lowscores, irrespective of item content, wouldinfluence the total scale score. A second po-tential problem concerns the discriminantvalidity of the scale. Substantial correlations(ranging from .4 to .5) have been foundbetween loneliness scores and the Beck De-pression Inventory (Bragg, 1979) and theCoopersmith measure of self-esteem (Joneset al., in press). Conceptually, it is reasonablethat loneliness might co-occur with depres-sion and low self-esteem, and such findingssupport the validity of the UCLA LonelinessScale. At the same time, however, these find-ings indicate a need to demonstrate the dis-criminant validity of the scale by showingthat loneliness is distinct from related con-

structs. A final concern is the potential con-founding of loneliness scores with social de-sirability. Since there is a social stigmaattached to loneliness (Gordon, 1976), indi-viduals who want to appear in a positive lightmight underreport their experience of loneli-ness.

The two studies reported here address thesepotential problems with the original UCLALoneliness Scale and provide clarification con-cerning the nature of loneliness. In the firststudy, a revised version of the UCLA Lone-liness Scale is developed, incorporating newpositively worded items. The concurrent valid-ity of the revised scale is established by relat-ing loneliness scores to the experience of af-fects that have been linked both empirically(Russell et al., 1978) and theoretically(Weiss, 1973) to loneliness. A second studyprovides a further test of the concurrent valid-ity of the revised scale by examining relation-ships between loneliness and social behavior.In addition, Study 2 addresses the discrimi-nant validity of the revised scale by demon-strating that loneliness scores are distinctfrom measures of social desirability, socialrisk taking, negative emotional states, andaffiliative motivation.

Study 1

The first study had several goals. First, itwas designed to revise the UCLA LonelinessScale by incorporating new, positively wordeditems. The internal consistency of this revisedmeasure was assessed, and the correlationbetween scores on the original and the revisedscale was calculated. Finally, the concurrentvalidity of the revised loneliness scale wasinvestigated.

MethodThe 162 students (64 males and 98 females) who

voluntarily participated in Study 1 were all testedin the spring of their first year at UCLA. Theycompleted an extensive questionnaire; pertinent tothis research were the measures of loneliness andemotional states.

Loneliness measures. The original UCLA Lone-liness Scale (Russell et al., 1978) was given, fol-lowed by 19 new items written by the authors.These new items measured satisfaction with socialrelationships and represented, as nearly as possible,opposite wordings of the original scale items.

474 D. RUSSELL, L. PEPLAU, AND C. CUTRONA

A second set of measures assessed explicit self-labels of loneliness. Examples of such questions are"During your lifetime, how often have you feltlonely?" and "During the past two weeks, howlonely have you felt?" Six such questions wereasked, all involving the student identifying himselfor herself as lonely. Responses to each of these sixquestions were summed to form a single self-label-ing of loneliness index (coefficient alpha = .78).

Emotional state. The questionnaire containedthree mood measures assessing anxiety and depres-sion. The Beck Depression Inventory (BDI; Beck,1967) consists of 21 symptoms or attitudes charac-teristic of clinical depression. Each item has severalalternatives describing manifestations of each symp-tom that vary in intensity. Scoring on each itemranges from 0 to 3, depending on the severity ofthe symptom manifestation that is selected by re-spondents as being self-descriptive. The BDI wasfound to be quite reliable with clinical populations,with a Spearman-Brown split-half coefficient of .93being reported. Validity for the measure has beenindicated by relating BDI scores to clinical judg-ments of the severity of depression: Correlations of.65 and .67 were found in two studies. The BDI hasalso been shown to assess validly the severity ofdepression in college populations (Bumberry, Oliver,& McClure, 1978).

The Costello-Comrey Depression and Anxietyscales were also administered (Costello & Comrey,1967). To develop these scales, a variety of factoranalytic studies were conducted with both normaland clinical populations. The goal of these analyseswas the construction of factorially "pure" anxietyand depression scales, which minimized the inter-correlation of the two measures; the final anxietyand depression scales were found to correlate .40for males and .50 for females. Validity for the twoscales has been indicated by relationships withclinical diagnoses and other measures of anxiety anddepression. Split-half reliabilities of .90 for the de-pression scale (14 items) and .70 for the anxietyscale (9 items) have been reported.

In addition to these mood measures, students alsorated on 9-point scales the intensity of their currentexperience of 25 emotions. These included such af-fects as bored, empty, hopeless, and satisfied, alongwith a set of affects believed to be unrelated toloneliness, such as resigned, embarrassed, and con-fident.

Results

Data analyses had two goals: to develop arevised loneliness scale and to assess the con-current validity of the revised scale.

Revised UCLA Loneliness Scale. Prior todata analysis, it was decided that the newscale should consist of 20 items, half reflectingsatisfaction with social relationships and halfreflecting dissatisfaction. Accordingly, 10 ofthe new positively worded items were selected

for the final scale, along with 10 of the origi-nal negatively worded items. The criterion forselecting these items was their correlation withthe self-labeling loneliness index. The 10 posi-tively worded and 10 negatively worded itemsthat had the highest correlations (all greaterthan .40 in magnitude) with the self-labelingindex were selected for the final instrument.The revised UCLA Loneliness Scale items andscoring format are given in Table 1. Notethat the original and new items are randomlyintermixed in the revised scale.1

The internal consistency of the revisedmeasure was high (coefficient alpha of .94)and compared favorably with the alpha co-efficient of .96 obtained for the original scale.The correlation between the revised andoriginal scales for the present sample was .91.

Tests for sex differences were also con-ducted, comparing the scores of male andfemale students on the revised measure. Asignificant sex difference was found, £ ( 1 5 7 )= 3.20, p < .001, with men scoring signifi-cantly lonelier than women (Ms = 36.23 and31.12, respectively).- However, including sex

1 From the current data, we have also developeda 4-item survey version of the UCLA LonelinessScale, consisting of two positively worded and twonegatively worded items. Using optimal subset re-gression procedures, the set of four items that bestpredicted scores on the self-labeling loneliness indexwere selected. The items chosen were Numbers 1,13, 15, and 18 from Table 1. This four-item lone-liness scale had a coefficient alpha of .75 in thecurrent study. We recommend that investigators whowant a shortened version of the loneliness scale usethese four items. Normative data and a bibliographyof research using both versions of the scale are alsobeing compiled by the authors. Investigators usingthe measures are urged to send us summary datafrom their samples as well as a brief description oftheir research and findings.

2 Since previous research has not indicated anysex differences in loneliness, this result suggests apossible sampling bias. The participants in Study 1were originally tested in the fall of their first year atUCLA, and returning versus nonrcturning studentswere compared on their loneliness scores from thisearlier testing. No significant differences were foundfor women. In contrast, men who returned for re-testing were lonelier than nonreturning men, 2(143)= 2.45, p < .02, which could have created the ob-served sex differences. Because of this possible sam-pling bias, data from the participants in Study 1were not included in the normative statistics pre-sented in Table 2.

THE REVISED UCLA LONELINESS SCALE 475

Table 1The Revised UCLA Loneliness Scale

Directions: Indicate how often you feel the way described in each of the fol lowing statements. Circle onenumber for each.

Statement

1. I feel in tune with the people around me"2. I lack companionship3. There is no one I can turn to4. I do not feel alone"5. I feel part of a group of friends"6. I have a lot in common with the people around me"7. I am no longer close to anyone8. My interests and ideas are not shared by those

around me9. I am an outgoing person"

10. There arc people I feel close to"11. I feel left out12. My social relationships arc superficial13. No one really knows me well14. I feel isolated from others15. I can find companionship when I want it"16. There are people who really understand me"17. I am unhappy being so wi thdrawn18. People are around me but not w i th me19. There arc people I can talk to"20. There are people I can tu rn to"

Never Rarely Sometimes Often

1 21 21 21 21 2

21 2

1 2222222222222

3333333

3333333333333

Note. The total score is the sum of all 20 items." Item should be reversed (i.e., 1 = 4, 2 = 3, 3 = 2, 4 = 1) before scoring.

as a variable in the analyses indicated thatsex did not mediate any of the findings.

Loneliness and emotional state. To assessthe concurrent validity of the revised scale,the relationship of loneliness scores to mea-sures of emotional states was examined. Lone-liness scores were significantly correlatedwith scores on the Beck Depression Inventory(r — .62) and with the Costello-ComreyAnxiety (r — .32) and Depression (r = .55)scales. Loneliness scores were also significantlycorrelated (all rs, above .40) with feelingabandoned, depressed, empty, hopeless, iso-lated, and self-enclosed and with not feelingsociable or satisfied. Loneliness scores werenot significantly correlated with such concep-tually unrelated affects as feeling creative,embarrassed, sensitive, surprised, or thought-ful .

Study 2

Having obtained evidence concerning thepsychometric properties and concurrent valid-ity of the revised UCLA Loneliness Scale, we

undertook a second study to investigate issuesof validity in greater detail. The goal of Study2 was to demonstrate that loneliness, as mea-sured by the revised scale, is discriminablefrom measures of other, conceptually relatedconstructs such as depression, lack of affili-ative motivation, and low social risk taking.Study 2 also replicated the internal consis-tency analyses reported in Study 1 with anew sample and provided additional tests ofconcurrent validity.

Met/tod

To increase the generality of the findings, abroader sample of college students was recruited.One third of the students were enrolled in intro-ductory psychology classes at the University ofTulsa, one third were enrolled in introductory psy-chology classes at UCLA, and the remaining stu-dents were drawn from upper division psychologyclasses at UCLA. AH students received partial coursecredit for their participation. The total sample of237 students (107 males and 130 females) each com-pleted a questionnaire containing the followingmaterials, with the order of the materials randomlyvaried.

476 D. RUSSELL, L. PEPLAU, AND C. CUTRONA

Loneliness measures. The original UCLA Loneli-ness Scale was administered, along with the 10positively worded items included in the revisedscale. The self-labeling loneliness questions fromStudy 1 were also given, and a self-labeling loneli-ness index was again created by summing the re-sponses to these six items (coefficient alpha — . 7 2 ) .

Social activities and relationships. Students wereasked how frequently they had engaged in a varietyof solitary activities (e.g., eaten dinner alone) andsocial activities (e.g., done something with a friend)during the previous 2 weeks. Students also indicatedhow many close friends they had and the nature oftheir current dating or marital status.

Mood and personality measures. Seven measuresof mood and personality were administered. Toassess depression, the Beck Depression Inventory(Beck, 1967) was given. State anxiety was mea-sured using the State-Trait Anxiety Inventory(STAI; Speilberger, Gorsuch, & Lushene, 1970).This scale consists of a set of affective statements(e.g., "I feel calm") that are used by respondentsto describe their current feelings. Validity evidencefor this measure consists of correlations with otheranxiety measures and scores of individuals exposedto anxiety-provoking situations. Internal consistencyfor the STAI is high, with coefficient alphas rang-ing from .83 to .92 in different studies.

Self-esteem was measured using the Texas SocialBehavior Inventory (TSBI-Form A; Helmreich &Stapp, 1974). This scale is designed to assess socialself-esteem; items concern the respondent's feelingsof social self-confidence. The TSBI has been vali-dated by demonstrating relationships with inter-personal attraction in laboratory settings and withthe endorsement of positive self-descriptions on ameasure of masculinity and femininity. For the ver-sion of the scale used here, a coefficient alpha of.85 was found for both males and females.

Mehrabian's (1970) measures of Affiliative Ten-dencies and Sensitivity to Rejection were includedto assess approach and avoidance orientationstoward social relationships. Validity for the mea-sures has been indicated by low correlations with ameasure of social desirability and (as theoreticallypredicted) by nonsignificant correlations betweenscores on the two scales. In laboratory studies theAffiliative Tendency scale has been related to af-filiative behavior in social situations and to a mea-sure of dependency. Sensitivity to Rejection hasbeen related to dependency and to conformity insocial situations. The reliabilities for both measuresare sufficiently high; K-R 20 (Kuder-Richardsonformula) coefficients of .80 and .83 are reported byMehrabian for the Affiliative Tendency and Sensi-tivity to Rejection scales, respectively.

To measure social desirability, the Marlowe-Crowne Social Desirability Inventory (Crowne &Marlowe, 1964) was administered. This scale isdesigned to identify individuals who tend to describethemselves in an overly positive or desirable fashion.Individuals who respond in a socially desirablefashion on this measure have been found to be more

favorable in their attitudes toward a boring task,to be socially conforming, and to be very susceptibleto persuasion. A test-retest correlation of .88 forthis measure has been reported over a 1-monthperiod. Internal consistency for the measure is alsohigh, with a K-R 20 coefficient of .88 being found.

The measures of introversion-extroversion andlying developed by Eysenck and Eysenck (1975)were also administered. The Introversion-Extrover-sion scale is designed to assess whether the re-spondent is a sociable and friendly person versus aquiet and introspective person. Validation of thismeasure has involved demonstrations of relationshipsbetween scores on the scale and conditionability,level of aspiration, vigilance, and time judgment,based on Eysenck's personality theory. The Liescale is designed to assess whether individuals aredistorting their responses. This measure has beenvalidated by comparing scores on the scale underconditions where the respondents should be moti-vated to distort their responses versus neutral con-ditions. Test-retest reliabilities for both measuresare high, with a correlation of .89 found for theIntroversion-Extroversion scale and .84 for the Liescale over a 1-month period. A coefficient alpha of.85 is reported for the Introversion-Extroversionscale and .80 for the Lie scale.

A final measure that was administered was theAssertiveness scale developed by Rathus (1973). Onthis scale, respondents indicate how self-descriptivea set of assertive and nonassertive behaviors are.Validity for the measure has been indicated by sig-nificant correlations between scores on the scale andpeer ratings of assertiveness. Significant relation-ships were also found between scale scores andverbal reports of assertive behavior in different socialsituations. Test-retest reliability (over a 2-monthinterval) of .78 is reported by Rathus, along witha split-half correlation of .77. For the present study,half of the 30 items on the Rathus measure wereused, consisting of every odd item from the scale.

Results

Sex differences. Comparisons were madebetween the mean loneliness scores for malesand females. In contrast to the findings re-ported in Study 1, no significant differenceswere found, / ( 2 2 8 ) = .72, ns. In Table 2,normative loneliness data from this sampleare presented separately for males and fe-males. Including sex as a variable in theanalyses reported below indicated no medi-ation of the relationships by sex of subject.

Internal consistency. To cross-validate theinternal consistency findings from Study 1, thesame procedures were used to analyze datafrom the second study. An alpha coefficient of.94 was again found for the revised loneliness

THE REVISED UCLA LONELINESS SCALE 477

Table 2Normative Statistics for Male and FemaleCollege Students in Study 2

Statistic Males Females

nMSDMdnModeRange

10237.0610.9135.3830.0020-68

12836.0610.1134.1729.0020-66

measure; and once again the correlation be-tween scores on the original and revised lone-liness scales was .91. The similarity of thefindings from these two independent studiesis striking.

Concurrent validity. The relationship be-tween scores on the revised loneliness scaleand measures of social activities and relation-ships were examined as a test of concurrentvalidity. For solitary relationships, significantcorrelations (all />s < .001) were found be-tween loneliness scores and the amount oftime students spent alone each day (r = .41),the number of times they had eaten dinneralone during the previous 2 weeks (r = .34),and the number of times they had spent aweekend night alone during the previous 2weeks (r=.44). Lonely students also re-ported doing fewer social activities withfriends (r = — .28) and having fewer closefriends (r — —.44). The relationship betweenloneliness and dating or marital status wasexamined, using analysis of variance to com-pare three groups: those students who werenot dating at all, those students datingcasually, and those students either datingsteadily or married. Using scores on the re-vised UCLA Loneliness Scale as the depen-dent measure, significant differences werefound among these three groups, F(2, 187) =22.97, p < .001. Students who were not datingat all had a mean loneliness score of 43.1;students who were dating casually and thosewho were romantically involved had means of34.0 and 32.7, respectively. Post-hoc compari-sons (using the Scheffe procedure) indicatedthat students who were not dating were sig-nificantly more lonely than the other twogroups ,F( l , 189) = 35.23,p < .001; studentswho were dating casually did not differ sig-

nificantly from those who were dating steadilyor who were married.

Discriminant validity. The discriminantvalidity of the revised scale was assessed byexamining the relationship between lonelinessscores and scores on the other measures ofmood and personality. Initial analyses exam-ined the correlation of loneliness scores withthe mood and personality measures and theself-labeling loneliness index. An inspection ofthese correlations (see Table 3) indicates thatloneliness scale scores correlate more highlywith the self-labeling loneliness index thanwith any of the other measures. Nonetheless,loneliness was strongly correlated with manyof the mood and personality variables, raisingthe possibility that these variables might, ifcombined, account for much of the variancein loneliness scores. For example, lonelinessmight be a function of low social risk takingcombined with high levels of anxiety. There-fore, a second discriminant validity test wasconducted to examine whether scores on theUCLA Loneliness Scale were related moreclosely to the other measures of loneliness(i.e., the self-labeling loneliness index) thanto an optimal linear combination of the moodand personality variables.

Multiple regression analysis was used tocombine the mood and personality measuresas predictors of loneliness. To eliminate cor-relations among the predictors, a factor analy-

Table 3Correlation of Scores on the Revised UCLALoneliness Scale With Self-LabeledLoneliness and the Measures ofMood and Personality

MeasureRevised UCLALoneliness Scale

Self-labeling loneliness indexDepressionSelf-esteemI ntroversion-cxtroversionAffiliative tendencyAnxietyAssertivenessSensitivity to rejectionSocial desirabilityLying

.705

.505-.493-.457-.452

.359-.342

.276-.203-.001

Note. Based on 239 students in Study 2.

478 D. RUSSELL, L. PEPLAU, AND C. CUTRONA

Table 4Regression Analysis Predicting Scores on theRevised UCLA Loneliness Scale From Scoreson the Affiliative Motivation, Social RiskTaking, Negative Affect, and SocialDesirability Factors

PredictorStandardized

beta F value"

Affiliativemotivation

Social risktaking

Negative affectSocial

desirability

-.523

-.504.375

-.073

70.30*

59.32*41.50*

1.43

.12

.17

.14

.00

0 This is a test of the significance of each beta weight,with df = 1, 170.*p < .001.

sis was first conducted on the mood and per-sonality measures. Using principal factoringwith a varimax rotation, a four-factor struc-ture emerged.;i The first factor was labeledSocial Risk Taking; measures of introversion-extroversion, self-esteem, sensitivity to rejec-tion, and assertiveness loaded highly on it. Asecond factor was labeled Negative Affect; thedepression and anxiety scales defined thisfactor. The third factor was labeled SocialDesirability, since the measures of social de-sirability and lying loaded highly on thisfactor. A final factor was labeled AffiliativeMotivation; the affiliative tendency and in-troversion-extroversion scales had high load-ings.

In the subsequent regression analyses,scores on these four factors were used in lieuof the nine separate scales to predict scoreson the revised loneliness scale. All four factorscores were entered simultaneously into theregression equation. The regression results aresummarized in Table 4. Affiliative motivation,social risk taking, and negative affect wereall significant predictors of loneliness, whereasthe social desirability factor was unrelated toloneliness. In combination, these four factorspredicted 43% of the variance in lonelinessscores.

Although these factors were predictive ofloneliness, in combination they were able toaccount for less than half the reliable variancein loneliness scores. Using the alpha coefficient

as a lower bound estimate of the revised lone-liness scale's reliability, an additional 52%of the variance in loneliness scores is poten-tially explainable. This permitted a seconddiscriminant validity test for the lonelinessscale—a test of whether the unexplainedvariance could be accounted for by the self-labeling loneliness index.

To investigate this possibility, a hierarchi-cal regression analysis was performed. Thefour mood and personality factors were en-tered first into the regression equation, fol-lowed by the self-labeling loneliness index.After eliminating the variance explained bythe mood and personality factors, the self-labeling loneliness index was still a significantpredictor of loneliness, F(l, 169) = 81.01,p < .001. An additional 18% of the variancein loneliness scale scores was explained by theloneliness index. This result provides clearevidence of the discriminant validity of therevised UCLA Loneliness Scale.

A final discriminant validity test examinedwhether the previously reported concurrentvalidity evidence for the UCLA LonelinessScale was based solely on the influence of themood and personality variables. For example,is the relationship between loneliness and thenumber of close friends uniquely attributableto loneliness, or does it reflect the fact thatloneliness is correlated with low affiliativemotivation and low social risk taking? Thegeneral question is whether the relationshipsfound between loneliness and social behaviorsare due to the influence of a third set ofvariables (i.e., the mood and personality mea-sures) on both the person's loneliness andsocial relations. If scores on the revised lone-liness scale are related to the concurrentvalidity criteria independent of the mood andpersonality variables, then fur ther discrimi-nant validity evidence for the scale is pro-vided.

The concurrent validity analyses weretherefore redone, using partial correlations tocontrol statistically for the effect of the moodand personality variables. Significant relation-ships (p < .001) were again found between

•'' The factor analysis results arc available from theauthors.

THE REVISED UCLA LONELINESS SCALE 479

loneliness scores and the amount of timealone each day (r=.2T), the number oftimes dinner was eaten alone (r — .31), thenumber of weekend evenings alone (r = .31),and the number of close friends (r — — . 2 7 ) .However, the relationship between lonelinessand the number of social activities now onlyapproached significance (r = .12, p< .07).Analysis of covariance was used to test therelationship between loneliness and currentdating or marital status, controlling for theeffect of the mood and personality variables.This significant relationship also persisted,F(2, 144) = 7.3S, p < .001.

General Discussion

The results from Studies 1 and 2 indicatethat the UCLA Loneliness Scale has beensuccessfully revised. The revised lonelinessscale has high internal consistency, with a co-efficient alpha of .94 found in both studies.Concurrent validity for the new measure wasindicated by demonstrating that lonely peoplereport experiencing emotions theoreticallylinked to loneliness and do not report experi-encing emotions unrelated to loneliness.Lonely individuals also reported more limitedsocial activities and relationships. Discrimi-nant validity for the revised loneliness scalewas indicated by evidence that scores on themeasure were not confounded by social de-sirability. Scores on the scale were also foundto correlate more highly with other measuresof loneliness than with the measures of moodand personality variables that were examined.The revised loneliness scale passed a verystringent discriminant validity test, with thedemonstration that relationships between lone-liness and the concurrent validity criteriaexamined in Study 2 were independent of theinfluence of the other mood and personalityvariables on loneliness.

An important issue raised by altering theUCLA Loneliness Scale is whether empiricalrelationships found by researchers using theoriginal scale are still valid for the revisedversion. In Studies 1 & 2, the high correlationof .91 between scores on the original andrevised loneliness scales suggests that previousfindings would still hold true for the new scale.Future research replicating earlier findings

would provide empirical evidence on thisissue.

Although the results from the currentstudies support the utility of the UCLA Lone-liness Scale as a measure of loneliness, itshould be emphasized that the validity of ameasure is never "proven." The validity ofthe loneliness scale in other populations needsto be established, and the ability of the mea-sure to detect loneliness in such "at risk"groups as newcomers and the recently di-vorced should also be investigated. We hopethat the UCLA Loneliness Scale will providea starting point for a greater understandingof the widespread and distressing experienceof loneliness.

Reference Note1. Brennan, T., & Auslander, N. Adolescent loneli-

ness: An exploratory study of social and psy-chological pre-dispositions and theory. Unpub-lished manuscript, Behavioral Research Institute,Boulder, Colorado, January 1979.

ReferencesBeck, A. T. Depression. New York: Hoeber, 1967.Bradburn, N. The structure of psychological well-

being. Chicago: Aldine, 1969.Bragg, M. A comparative study of loneliness and

depression. Unpublished doctoral dissertation,University of California, Los Angeles, 1979.

Bumberry, W., Oliver, J. M., & McCIure, J. Validityof the Beck Depression Inventory in a universitypopulation using psychiatric estimate as the cri-terion. Journal of Clinical and Consulting Psy-chology, 1978, 46, 150-155.

Costello, C. G., & Comrey, A. L. Scales for measur-ing depression and anxiety. Journal of Psychology,1967, 66, 303-313.

Crowne, D. P., & Marlowe, D. The approval motive:Studies in evaluative dependence. New York:Wiley, 1964.

Eysenck, H. J., & Eysenck, S. B. G. Eysenck Per-sonality Questionnaire. San Diego, Calif.: Educa-tional and Industrial Testing Service, 1975.

Freedman, J. L. Crowding and behavior. New York:Viking Press, 1975.

Gordon, S. Lonely in America. New York: Simon& Schuster, 1976.

Helmrcich, R., & Stapp, J. Short forms of the TexasSocial Behavior Inventory (TSBI), an objectivemeasure of self-esteem. Bulletin of the Psycho-nomic Society, 1974, 4, 473-475.

Horowitz, L. M., & French, R. de Sales. Interper-sonal problems of people who describe themselvesas lonely. Journal of Consulting and ClinicalPsychology, 1979, 47, 762-764.

480 D. RUSSELL, L. PEPLAU, AND C. CUTRONA

Jacobs, J. Adolescent suicide. New York: Wiley,1971.

Jones, W. H., Freemon, J. R., & Goswick, R. A.The persistence of loneliness: Self and other re-jection? Journal of Personality, in press.

Lynch, J. J. The broken heart: The medical conse-quences of loneliness in America. New York:Basic Books, 1976.

Mehrabian, A. The development and validation ofmeasures of affiliative tendency and sensitivity torejection. Educational and Psychological Measure-ment, 1970, 30, 417-428.

Nerviano, V. J., & Gross, W. F. Loneliness andlocus of control for alcoholic males: Validityagainst Murray need and Cattell trait dimensions.Journal of Clinical Psychology, 1976, 32, 479-484.

Peplau, L. A., & Perlman, D. Blueprint for a socialpsychological theory of loneliness. In M. Cook &G. Wilson (Eds.), Love and attraction. Oxford,England: Pergamon Press, 1979.

Peplau, L. A., Russell, D., & Heim, M. Loneliness:A bibliography of research and theory. JSASCatalog of Selected Documents in Psychology,1978, *, 38. (Ms. No. 1682)

Perlman, D., Gerson, A. C., & Spinner, B. Lonelinessamong senior citizens: An empirical report.Essence, 1978, 2(4) , 239-248.

Rathus, S. A. A 30-item schedule for assessing as-sertive behavior. Behavior Therapy, 1973, 4, 398-406.

Russell, D., Peplau, L. A., & Ferguson, M. L. De-veloping a measure of loneliness. Journal of Per-sonality Assessment, 1978, 42, 290-294.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E.Manual for the State-Trait Anxiety Inventory.Palo Alto, Calif.: Consulting Psychologists Press,1970.

Stokols, D. On the distinction between density andcrowding: Some implications for future research.Psychological Review, 1972, 79, 275-277.

Weiss, R. S. Loneliness: The experience of emotionaland social isolation. Cambridge, Mass.: Massa-chusetts Institute of Technology Press, 1973.

Wenz, F. V. Seasonal suicide attempts and forms ofloneliness. Psychological Reports, 1977, 40, 807-810.

Received July 30, 1979