Translation and Cross Language Validation of Passionate Love Scale Among Adults in Lahore , Pakistan

Abstract Objective of the present study was to translate and cross language validate the Passionate Love Scale (PLS) from English language to Urdu language using forward-back translation (Brislin, 1976) procedure. It is developed by Hatfield and Sprecher (1986). It has 15 selfreported and uni-dimensional items, with 9 points rating scale ranges from 1 = not at all true to 9 = definitely true. Results showed high level of Cronbach’s alpha reliability coefficient α = 0.90, test retest reliability ranged from r = .73 to r = .96 (ps < .01), item total correlation varying from r = .50 to r = .74 (ps < .01) and factor loading ranged from .39 to .73. Item difficulty was determined through the Rasch scaling analysis and construct validity of the Urdu PLS secured the same one-dimensional factor structure of the English PLS by retaining 15 items. It showed that the Urdu PLS is reliable and valid tool to measure the cognitive, emotional and behavioral components of passionate love in Pakistani cultural context. Implications of the study were also discussed.

. They found that the love is a phenomenon of cultural universal as if it is emerged in all human groups with only slight variations across human history (Al-Krenawi & Jackson, 2014;Fisher, Brown, Aron, Strong, & Mashek, 2010).
Researchers accounted the universal nature of passionate love (Cacioppo, Bianchi-Demicheli, Hatfield, & Rapson, 2012;Fisher, Brown, Aron, Strong, & Mashek, 2010;Hatfield & Rapson, 2008b;Hatfield, Rapson, & Martel, 2007;Jankowiak & Fischer, 1992).They found that at least one type of love (passionate love) is expected to have emerged in all societies and in every era of human history.Intensity of passionate love can vary according to the experience of an individual, biological characteristics and cultural context (Kim & Hatfield, 2004;Landis & O'Shea, 2000;Yildirim, Hablemitoglu, & Barnett, 2014).Hatfield, Bensman, and Rapson (2012) reported in the article "A Brief History of Social Scientists' Attempts to Measure Passionate Love" that the concept of romantic love is as old as the existence of human being itself, but the idea of tool development on the passionate love was ignored till 1940's.Hatfield et al. enlisted 33 different scales of love with its necessary details in which Passionate Love Scale was selected to translate and validate from the source language (English) to target language (Urdu).Since, author claims the universality of the passionate love scale with the evidence that this scale has been translated into many languages such as Hernandez (2015) conducted research on the construct validity of Passionate Love Scale by using the data of 578 Brazilian individuals including both gender (men & women).Exploratory Factor Analysis retained the original one-dimensional factor structure of PLS which explained 64.2% of the common variance.Results indicated the significant level of internal consistency.Its scores were also positively associated with the Triangular Love Scale-R (Sternberg, 1997) and its subscales.Overall results showed that the Brazilian version of Passionate Love Scale is reliable and valid tool to measure the passionate love among the Brazilian population.Feybesse, Neto, and Hatfield (2011) Moreover, Yildirim, Hablemitoglu, and Barnett (2014) conducted a study on the reliability and validity of a Turkish version of the Passionate Love Scale by using the data of 150 undergraduate students.Results indicated the good level of internal consistency in form of Cronbach's alpha reliability coefficient (α = .89).Its factor loading ranged from .53 to .79 and item total correlation varied from .47 to .74.Confirmatory factor analysis retained the 15 items with single factor.Findings showed that the PLS is reliable and valid tool to measure the passionate love among Turkish population.They also found that 43.3% participants reported that they passionately loved at the time of the study, 33.3% that they had never passionately loved someone, and 23.3% indicated that they had previously been passionately in love with someone.According to their responses to the PLS, 30.7% of the participants reported that they were wildly, and even recklessly, in love, 29.3% felt less passion for their beloved, 21.3% occasionally felt passion for their beloved, 13.3% of the participants were tepid and felt passion infrequently, and 5.3% reported that the thrill of their love had faded away.Feybesse (2015) validated the Passionate Love Scale in France by using the data of 190 students (women 53.2% and men 46.8%).) of Paris Descartes University.Age ranges varied from 18 to 34 years old (SD = 2.99).
At the time of study 102 (53.7%) subjects were in love, 58 (35.8%) have been in love and 16 (8.4%)claimed that they have never been in love.Duration of the relationship varied from 23 to 47 months.Cronbach's alpha reliability coefficient for the 15 items version is .90.Therefore, it is the main objective of this study to translate and validate the passionate love scale from English language into Urdu language, to establish the cross language validation and determine the factor structure.Translation has made to avoid the linguistic barriers, cultural constraints and researcher's biasness.Moreover, to make it more comprehensible, functional, content and contextually equivalence according to the cultural context of Pakistan (Herdman, Fox-Rushby, & Badia, 1998;Laake, Benestad, & Olsen, 2007;Voracek, Fisher, Loibl, Tan, & Sonneck, 2008).

Objectives
The objectives of the current study were to translate and cross language validate the Passionate Love Scale (PLS) from source language (English) to target language (Urdu).It was achieved in three phases.In Phase I, forward back-translation (Brislin, 1976a) was carried out.In phase II, internal consistency was computed through test-retest reliability/cross language validation, Cronbach's's alpha reliability coefficient and item total correlation.In phase III, Factor structure of the Urdu PLS was determined through the Rasch Measurement Model (RMM) and Confirmatory Factor Analysis (Field, 2009).

The Passionate Love Scale (PLS)
The Passionate Love Scale (Hatfield & Sprecher, 1986) measures the cognitive, emotional and behavioral components of passionate love.It is a self-reported one-dimensional measure for young adults consisting of 15 items such as: "I would feel deep despair if he/she left me" by using 9 points rating scale; 1 = not at all true to 9 = definitely true.Its score calculated by mean or by sum of the ratings.All items are positively worded and not reverse scored.High scores on the PLS shows high level of passion and low score shows less intensity of passionate love.There are several ranges of scores such as 106-135 score indicate that the individual is wildly and recklessly involve in passionate love, 86-105 scores show the passion but comparatively less intense in nature, 66-85 scores depict infrequent gusts of desire, 45-65 scores reveal the tepid and infrequent passion, while 15-44 scores show that the thrill is gone.Cronbach's alpha reliability coefficient of the original PLS was reported α = .91.
Phase 1: Translation of Passionate Love Scale Brislin (1976a) forward-back translation method was used to translate the instrument from English language to Urdu language which was completed in the following four stages:

Procedure
Permission was taken from the original authors (Hatfield & Sprecher, 1986) for the translation of instruments which graciously granted.Stage I: It was requested to five bilingual experts to translate the instruments from English language to Urdu language independently by following the same guideline for translations (Brislin, 1976b).English language without exposing them the original items.Two Urdu language experts evaluated the appropriateness of the Urdu language and sentence structure of the instrument.Stage IV: Test try out and pilot study was conducted.

Stage I: Forward Translations
For forward translations five bilingual experts (three PhD holders of English department and two M. Phil degree holders of psychology department of GC University Lahore) were approached.They have the expertise in item writing and translating instruments.It was requested to them to translate the content of the scales from English to Urdu language, while keeping its contextual meaning intact (see Appendix for details).

Stage II: Committee Approach
Committee approach was used to check the adequacy and compatibility between the English and Urdu language of the instruments.Committee comprised of n = 5 (2 women and 3 men) bilingual experts.Committee members assessed the accuracy of translation, contextual meaning, concept clarity, language difficulty, grammatical validation, sensitivity to regional discrepancies and evaluated.Five forward translations were compared based on three criteria (1) cultural context (2) conceptual translation (3) removing any surplus and inadequate words from the items.

Stage III: Back Translation / Reformulation of Equivalence
Three back translations were carried out to assess the ambiguities, inconsistencies and differences between target and source language.Three bilingual experts (did M. Phil in English department from GC University Lahore) of Urdu and English language were contacted for the back translations.Urdu translated version of scales was given to them instead of original instruments.After these translations, the same committee members that evaluated the forward translations were approached.They compared three back translations by using the criteria of conceptual equivalence, language difficulty, deleting extra words and make it more comprehensible according to the cultural context of Pakistan.It was finalized after minimizing the discrepancies of conceptual meanings.Two Urdu language experts (did Master in Urdu, from the Government College University Lahore) improved the quality of item structure.After expert evaluation of the back translations the instruments were sending back to the original authors.They were agreed with the back translations and did not suggest major changes.

Stage IV: Pilot study
Once finalizing the Urdu translated version of Passionate Love Scale (PLS) test try out was conducted on n = 15 individuals to identify the problematic and difficult items.After modification and improvement of the items, pilot study was conducted on n = 50 participants.Results of test try out showed that 98% of students found the questionnaire easy while 2% asked to clarify the meaning of some items (for example item number 4 "I would rather be with him/her than anyone else" and item number 6 "I yearn to know all about him/her) which were improved in the pilot study after consultation of the original author.Findings of pilot study revealed that 100% students found it easy to mark and understandable.Original item order was followed in the Urdu translated version of PLS.
Phase 2: Internal Consistency of Urdu PLS Internal Consistency and Construct Validity of PLS were determined in the phase II.Internal consistency was included the cross language validation, Cronbach's alpha reliability coefficient and item total correlation respectively.

Method Sample
Purposive sampling technique was used.The sample of cross language validation consisted of 60 volunteer and bilingual participants (15 in each group) with equal distribution of gander (30 women & 30 men).They were matched on their demographic variables such as age (19-60 years), gender (men & women), education (BA/ BSc/BSc (Hons) to PhD), marital status (married or unmarried), profession (employee or unemployed) and area (urban or rural).

Procedure
After seeking permission from the relevant authority, volunteer participants were approached in group forms.
They were randomly divided into four groups (Urdu-Urdu, Urdu-English, English-English and English-Urdu) by taking all the ethical considerations (anonymity, confidentiality and right of withdraw) in account.Paper pencil method was used.They filled PLS along with demographic variables sheet.All these participants were approached again (for test-retest reliability) with the fifteen days gape by keeping all the conditions (same sitting arrangements, place, instructions, instruments, observer and objectives) standardized.It took average 10 to 15 minutes to complete both forms.Participants were debriefed and thanked for their cooperation.

Cross Language Validation
Cross language validation (Naqvi & Kamal, 2010) was conducted to check the compatibility between the Urdu translated version of PLS and original scale.Table 1 indicates the highly significant positive relationship among each condition of test retest reliability especially the Urdu-Urdu version of Passionate Love Scale shows r = .96,p < .01,which is highly significant.

Cronbach's Alpha Reliability
Results reported in

Item Total Correlation
Results reported in Table 3 shows highly significant positive relationship between each item to item total correlation which means that the Urdu Passionate Love Scale is an internally consistent measure of passionate love and every item positively contributed to improve the reliability of the instrument.

Rasch Analyses on Passionate Love Scale (PLS)
In-depth psychometric evaluation of the Urdu Passionate Love Scale was determined through the Rasch measurement model (Bond & Fox, 2015;Hagquist, Bruce, & Gustavsson, 2009;Smith, Conrad, Chang, & Piazza, 2002).It is an item response theory model which helps to describe the required scaling properties of linear interval measurement and the response pattern observed in the data corresponds to the theoretical pattern expected by the model (Embretson & Reise, 2000;Tennant & Conaghan, 2007).
Table 4 shows the results of the Rasch analyses for the 15-item passionate love scale.It was found that item difficult ranges from ±.55 to ±1.62.All of the items showed acceptable infit values ranging from .55 to 1.49 (WMS less than 2.0), and outfit values ranging from .57to 1.99.The entire scale reliability was .90, and person separation index was 2.33.It can be observed that all the statistics showed acceptable rating scale model fit.In addition, the threshold for rating scale (1 to 9) was between ±.01 to ±.65.

Factor Structure of the Urdu Passionate Love Scale
The Confirmatory Factor Analysis (CFA) was employed to determine the factor structure of the Urdu translated versions of Passionate Love Scale (UPLS).It was used to measure the predetermined hypothesis, theory or empirical evidence based research (Campbell & Fiske, 1959;Cronbach & Meehl, 1955;Preedy & Watson, 2010).Control conditions (Schermelleh-Engel, Moosbrugger, & Müller, 2003).Results interpretation uses the values of Chi-square test (χ 2 ) (Jöreskog & Sörbom, 1989).It shows the discrepancies between perceived and estimated covariance matrices and scores near to zero represents the little bit variations on the observed and estimated scores.Chi-square test is sensitive to sample size, small sample may reject the appropriate modal and large sample may retain the inappropriate model fit indices (Gatignon, 2010).This problem compensated with the Root Mean Square Error of Approximation (RMSEA).Its ranges varied from 0 to 1 and lesser values indicate the better model fit while < .06represents the acceptable level of model fit (Hooper, Coughlan, & Mullen, 2008).
Another criterion of goodness of fit index (GFI) is employed which determine the discrepancies between the assumed model and the observed covariance matrix.Its ranges varied from 0 to 1 and scores greater than .9 shows the satisfactory model fit (Baumgartner & Homburg, 1996).
The comparative fit index (CFI) deals with the difference between observations and hypothesized model.It also deals with the sample size problems existed in the chi-squared test normed fit index (Bentler, 1990).Its values varying from 0 to 1 and greater values (< .95)means better fit model (Hu & Bentler, 1999).Other criteria of Goodness of-fit index (GFI) was used which represented the latent factors.Its values ranged from 0 to 1 and greater than .9represent the acceptable level of model fit (Behling & Law, 2000).Analysis of Moment Structure (AMOS 18 software version) was used.Missing values were replaced with the mean before data was subjected to further analysis.In every model, basic factors were considered as latent variables and items were taken as observed values.Unseen factors were subjected as error of observed values.Path analysis was used to draw the diagrams and fit the variances and covariances matrix.Standardized regression weights were specified and factor loading < .35 was considered as acceptable value (Anastasi & Urbina, 1997;Field, 2009).
Results of Table 5 shows the Chi-Square value for Urdu Passionate Love Scale χ 2 (51, N = 300) = 73.02,p < .02,CFI = .96,RMSEA = .05was close fit model.A non-significant χ 2 makes a good fit model and could result rejection of the null hypothesis.According to the Bentler (1995), with large sample size χ 2 become significant while with the small sample the assumption of the χ 2 test reveals an inaccurate probability.Results strongly supports that the Urdu Passionate Love Scale is close fit model (see Figure 1 for the graphical representation of the model) Note.GFI = Goodness-of-Fit Index; CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of approximation.

Discussion
The main objective of the present study was to translate and cross language that validates the Passionate Love Scale (PLS) from English language to Urdu language.It was completed in three phases.In phase I, translation was done by using the standard procedure of forward back-translation (Brislin, 1976b).In phase II, internal consistency of the Urdu translated version of PLS was determined through test re-test reliability/cross language validation, Cronbach's alpha reliability coefficient and item total correlation.In Phase III, Item difficulty and factor structure of the Urdu PLS was calculated through the Rasch Measurement Model (RMM) and Confirmatory Factor Analysis (CAF).
First objective of the current study (phase I) was to translate the English version of PLS into Urdu Language by using the standard procedure of translation which was achieved successfully.Results of phase I (test try out) shows that participants interpret the item number 4 (I would rather be with him/her than anyone else) in terms of long term commitment like marriage.It means if they have given the chance to choose the partner/spouse they prefer to be married with the romantic partner rather than anyone else.While author develop this item in urge to spend preferable time with romantic partner in day to day life rather than anyone else like friends, family  Kagitcibasi, 1990).
Respondent also found the problem in item number 6 (I yearn to know all about him/her) as they were reluctant to answer it and perceived it comparatively irrelevant to investigate the personal life of the lover or they might be threaten to violate the personal boundaries of the lover or think it is irrelevant to know the love as they accept him/her with all of his/her strength and weakness.But the author clarify its meaning in a sense that in passionate love it is natural wish to try to know the liking and disliking of lover so that to understand each other well.Or it might be happened due to empirically supported aspects of Asians culture which includes the aspects of collectivism, joint family system, and conformity to social norms, Islamic culture, personal restraint, reservation, and suppression of emotional expression (Kim, Atkinson, & Umemoto, 2001).
Second objective of this study was to determine the internal consistency of the Urdu version of PLS through test re-tests reliability/cross language validation, Cronbach alpha reliability coefficient and item total correlation.
Findings showed the high level of internal consistency and test re-tests reliability/cross language validation of Urdu version of PLS.Findings were remains consistent with the previous literature (Feybesse, Neto, & Hatfield, 2011;Graham & Christiansen, 2009;Hatfield, Forbes, & Rapson, 2013;Hatfield & Sprecher, 1986;Yildirim, Hablemitoglu, & Barnett, 2014).Value of Cronbach's alpha reliability coefficient on scores of Urdu Passionate Love Scale was found quite high i.e. α = .90which considered the best (Tezbasaran, 1997).Split-half reliability coefficient of the PLS is .89for the part I and .90 for the part II.This reliability remained same when the Spearman-Brown formula was applied.The internal consistency was further determined by calculating the inter-item and item total correlation for the Urdu translated version of Passionate Love Scale.Highly significant positive relationship was found between item-total correlations which means each item significantly contribute in the total scores of Urdu PLS.
Third objective of the present study was to determine the item difficulty and factor structure of the Urdu PLS.
The Rasch model was used to determine the item difficulty and response pattern of the respondents on the current sample.It is an item response theory model which helps to make a linear interval scale, to fit the model.
It is correspond with the theoretical and proposed expected model (Conrad et al., 2012;Tennant & Conaghan, 2007).Findings of the Rasch model showed the appropriateness of the results which remain consisted with the previous literature (Lange, Houran & Li, 2015;McCreary et al., 2013) Factor structure of the Urdu Passionate Love Scale was determined through Confirmatory Factor Analysis (Cole, 1987).Results show the high level of factor loading.The criteria of >.35 for factor loading was followed.
Item which scored below to .35 was excluded from the further analysis.Results of the present study showed that the ranges of the factor loading varying from .39 to.73.Therefore, Confirmatory Factor analysis was run on the 15 items and the results of CAF retains the original uni-dimensional factor structure of the Urdu Passionate Love Scale as described by the Hatfield and Sprecher (1986).Same factor structure of the PLS was found while it was translated from English to Portuguese, French, and Turkish (Feybesse et al., 2011;Yildirim, Hablemitoglu, & Barnett, 2014).
conducted a study on the adaptation of the Passionate Love Scale into the Portuguese population by using the data of 204 Portuguese university students.Results of Factor Analysis showed the single dimension and good internal consistency of the PLS.Findings of this article remained consisted with the original English version of PLS.
Stage II: Moreover, five bilingual committee members (2 M. Phil degree holders and 3 PhD scholars had taken from the department of Psychology Government College University Lahore) were approached to finalize the Urdu version PLS.Stage III: Three bilingual experts translated the Urdu version of all items into Interpersona 2017, Vol.11(2), 188-206 doi:10.5964/ijpr.v11i2.230 Software of Analysis of Moment Structure (AMOS 18) were used to measure the "good model fit" which means the suitability of the model and absolute fit indices follow the criteria of (1) Prior theory, model or hypothesis; (2) Variations; (3) Specification of latent variable; (4) relationship among latent variables; and (5) Interpersona 2017, Vol.11(2), 188-206 doi:10.5964/ijpr.v11i2.230

Figure 1 .
Figure 1.Path diagram for the Urdu version of Passionate Love Scale (PLS).Note.Item 1-Item 15 in the Passionate Love Scale.In the figure, left sided arrows indicates the factor loading which is mentioned in Table 4 and right sided arrows show the R 2 while reciprocated arrows depicts the covariance between items.
Table2shows high level of Cronbach's alpha reliability coefficient on the scores of Urdu Passionate Love Scale (α = .90)which is quite good and high.

Table 2
Mean, Standard Deviation and Cronbach's Alpha Reliability Coefficient of the Urdu Passionate Love Scale(N = 300)

Table 3
Item Total Correlation of Urdu Passionate Love Scale(N = 300) Note.Urdu version of the Passionate Love Scale (see Appendix for details).**p <.01.

Phase 3: Factor Structure of the Urdu Passionate Love Scale Sample
engaged, 20% in relationship and 30% unmarried.70%experiencedlove and 30% never experienced it yet.186participantshave experienced passionate love at some point in life while 114 have never experienced it.Out of 300 respondents, 162 participants attempted the questionnaire by keeping in mind about their current beloved person, 24 responded according to their past experience of romantic love and 114 reported to never experience passionate love.188 individuals madly and recklessly experienced passionate love, 60 experienced Anjum & Batool 193 Interpersona 2017, Vol.11(2), 188-206 doi:10.5964/ijpr.v11i2.230less intensity of passionate love, 42 experienced occasional bursts of passion, 10 experienced halfhearted or infrequent passion and reported loss of thrill.

Table 5
Chi-Square, Degree of Freedom and Stepwise Model Fit Indices of CFA for Passionate Love Scale(N = 300) social gathering which was clear by the original author.It might be happened as an intimate relationship may differ in Pakistan due to collectivistic culture.Pakistani society differs from Western society in terms of sociocultural features and social structure despite of rapid westernization ( Interpersona 2017, Vol.11(2), 188-206 doi:10.5964/ijpr.v11i2.230or