The Translation and Preliminary Psychometric Validation of the Ghosting Questionnaire in Urdu

Round 1

Reviewer 1 Report

Thank you for the opportunity to review the manuscript. I find that the manuscript is well-written, and its content and structure are appropriate for articles reporting findings from psychometric research. The cultural adaptation of the Ghosting Questionnaire (GQ) into Urdu is deemed useful and significant. In cross-cultural research, especially when one of the countries involved is India or Pakistan, English versions of tools are often used to measure variables. Despite the prevalence of English in these countries, a lack of adaptation to local languages can exclude large social groups from studies. Despite my positive assessment of the manuscript, I believe that introducing or improving a few elements would enhance the article – my comments are outlined below.

Description of the Sample: It is necessary to provide additional information about the socio-demographic characteristics of the study participants. Currently, the article lacks descriptive statistics regarding gender, age, education, occupation, economic status, and relationship status of the participants. This information could be valuable in interpreting the study results. A limitation of the validation study is its restriction to the academic population – acknowledging this in the results discussion would be beneficial.

Varimax Rotation Method: The varimax rotation method is applicable in Exploratory Factor Analysis (EFA) for extracting factors, rather than in Confirmatory Factor Analysis (CFA), and certainly not for one-factor solutions. Clarification is needed.

Table 2: Table 2 pertains to Principal Component Analysis (PCA), not CFA, as previously announced in the article – these are two distinct analyses.

Data Presentation (Table 3): The data from Table 3 could be incorporated into the text, selecting only the most relevant fit indices – there is no need to report all of them.

Validity Evidence: The article lacks evidence of the validity of the tool based on associations with other variables or other instruments measuring similar phenomena – this is a clear research weakness.

Scale Construction Information: Information about the scale construction, currently in the Discussion section, should be moved to the Materials and Methods section.

Given the study's limitations, I suggest changing the title to "Translation and Preliminary Psychometric Validation of the Ghosting Questionnaire in Urdu" to emphasize the need for further validation studies.

Author Response

Thank you for the opportunity to review the manuscript. I find that the manuscript is well-written, and its content and structure are appropriate for articles reporting findings from psychometric research. The cultural adaptation of the Ghosting Questionnaire (GQ) into Urdu is deemed useful and significant. In cross-cultural research, especially when one of the countries involved is India or Pakistan, English versions of tools are often used to measure variables. Despite the prevalence of English in these countries, a lack of adaptation to local languages can exclude large social groups from studies. Despite my positive assessment of the manuscript, I believe that introducing or improving a few elements would enhance the article – my comments are outlined below.

Authors' response: Thank you for your thoughtful concerns and comments. We sincerely appreciate your valuable recommendations and have taken them into careful consideration. We have thoroughly addressed each and every comment you provided and incorporated the necessary revisions into the manuscript. To facilitate the revisions, we have highlighted our responses to your comments in yellow throughout the revised version of the manuscript. We hope these revisions meet your expectations and contribute to the overall quality of the paper.

Description of the Sample: It is necessary to provide additional information about the socio-demographic characteristics of the study participants. Currently, the article lacks descriptive statistics regarding gender, age, education, occupation, economic status, and relationship status of the participants. This information could be valuable in interpreting the study results. A limitation of the validation study is its restriction to the academic population – acknowledging this in the results discussion would be beneficial.

 Authors’ response: A new section (2.2) has been added to address this. As follow:

“2.2 Participants

The study included 540 participants, of which 200 (37%) were male and 340 (63%) were female. In terms of profession, the majority were students (n=469, 87%), followed by work-ing individuals (n=63, 12%) and unemployed participants (n=6, 1%). Regarding marital status, 461 (85%) were single and 79 (15%) were married. The mean age of participants was 22.37 (SD = 6.22) years, with a range from 14 to 56 years. The mean education level was 14.04 (SD = 2.24) years of schooling”.

In the discussion we provided the following limitation: “A limitation of the current validation study is the homogeneous sample recruited from a university population. While convenient sampling of students enabled initial evaluation of the Urdu Ghosting Questionnaire, it restricts generalizability of the findings. The sample lacked diversity in terms of age, relationship status, and other demographics that may influence experiences with ghosting. Measurement invariance across different subgroups needs to be examined in future research. Wider sampling from the community would provide a stronger evaluation of the factor structure and reliability of the Urdu Ghosting Questionnaire. The tool requires further validation using broader recruitment strategies to capture a more representative cross-section of the population. Though the student sample provides preliminary evidence of the questionnaire's adequacy, confirmation is needed of its psychometric properties and utility for measuring ghosting experiences in the general public. Ensuring cultural and demographic inclusiveness should be a priority going forward”.

Varimax Rotation Method: The varimax rotation method is applicable in Exploratory Factor Analysis (EFA) for extracting factors, rather than in Confirmatory Factor Analysis (CFA), and certainly not for one-factor solutions. Clarification is needed.

 Authors’ response: As the reviewer rightly pointed out, varimax rotation is a method used in Exploratory Factor Analysis whereas our analysis involved Confirmatory Factor Analysis of a one-factor model. Reference to varimax rotation was introduced mistakenly during revisions and does not reflect the actual confirmatory approach used. We have removed the statement about varimax rotation from the manuscript to eliminate this confusion. Please accept our sincere apologies for the confusion and any inconvenience caused.

We updated the results of our analyses for clarity.

Table 2: Table 2 pertains to Principal Component Analysis (PCA), not CFA, as previously announced in the article – these are two distinct analyses.

 Authors’ response: Table 2 was updated to include results of the CFA as follows:

CFA was conducted to evaluate the one-factor structure of the Urdu Ghosting Questionnaire. The hypothesized single factor model showed good fit with the data, as indicated by the CFI (0.991), TLI (.987), NFI (0.983) and other incremental fit indices exceeding 0.95. The RMSEA of 0.045 and SRMR of 0.018 were below recommended cutoffs, and the significant model chi-square (χ2(20) = 41.95, p = 0.003) could be attributed to the large sample size. All factor loadings were statistically significant and ranged from .812 to 1.070. The average variance extracted was 0.572, satisfying the recommended threshold for convergent validity. Discriminant validity was evidenced by the lack of correlation with any other factors. The unidimensional factor structure was supported, with high loadings suggesting scale items effectively measure the underlying construct of ghosting. The CFA provides empirical validation of the factorial validity and psychometric soundness of the Urdu Ghosting Questionnaire. Table 2 provides the CFA of the the Ghosting Questionnaire-Urdu (n = 504).

Table 2: Confirmatory factor structure for the Ghosting Questionnaire-Urdu (n = 504).

Notes: Confirmatory factor structure estimated using Maximum Lowlihood Extraction (MLE). Fitness measures Χ², (df) = 41.95, 20, Comparative Fit Index (CFI) = 0.991, Tucker-Lewis Index (TLI) = 0.987, Bentler-Bonett Non-normed Fit Index (NNFI) = 0.987, Bentler-Bonett Normed Fit Index (NFI) = 0.983, Parsimony Normed Fit Index (PNFI) = 0.702, Bollen's Relative Fit Index (RFI) = 0.976, Bollen's Incremental Fit Index (IFI) = 0.991, Relative Noncentrality Index (RNI) = 0.991, Log-likelihood = -5826.861, Number of free parameters = 24, Akaike (AIC) = 11701.721, Bayesian (BIC) = 11804.719, Sample-size adjusted Bayesian (SSABIC) = 11728.534, Root mean square error of approximation (RMSEA) = 0.045, RMSEA 90% CI lower bound = 0.026, RMSEA 90% CI upper bound = 0.064, RMSEA p-value = 0.638, Standardized root mean square residual (SRMR) = 0.018, Hoelter's critical N (α = .05) = 405.346, Hoelter's critical N (α = .01) = 484.59, Goodness of fit index (GFI) = 0.994, McDonald fit index (MFI) = 0.98, Expected cross validation index (ECVI) = 0.167.

Data Presentation (Table 3): The data from Table 3 could be incorporated into the text, selecting only the most relevant fit indices – there is no need to report all of them.

 Authors’ response: We included detailed description of the results of Table 3 in the main text as follow: “The convergent validity of the GQ-U was confirmed by the strong correlation between the Urdu and English versions of the GQ (Table 3; r = 0.947; p < 0.001). . Correlations be-tween ghosting and age (r = 0.386; p < 0.001) and education (r = 0.218; p < 0.001) were also found to be statistically significant. See Table 3”.

However, we opted to retaining Table 3 to facilitate the presentation of findings. Also, to allow future meta-analysts to find the data quickly.

Validity Evidence: The article lacks evidence of the validity of the tool based on associations with other variables or other instruments measuring similar phenomena – this is a clear research weakness.

 Authors’ response: We agree with the reviewer however we would like to highlight the fact that due to the lack of availability of other validated ghosting measures, we were delimited in our ability to assess convergent validity with independent instruments. The Ghosting Questionnaire is the only available too. To partially compensate, we examined the relationship between scores on the Urdu Ghosting Questionnaire and the original English language version.

We added to the discussion: “The Ghosting Questionnaire is the only existing tool assessing this construct. Due to this limitation, we were unable to examine convergence with unrelated instruments as would be ideal”.

Scale Construction Information: Information about the scale construction, currently in the Discussion section, should be moved to the Materials and Methods section.

Authors’ response: This information has been transferred to the method section. Which is: “The translated version comprises eight items and uses a 5-point Likert scale, i.e., never (1), rarely (2), sometimes (3), often (4), and always (5). A higher score indicates greater involvement in ghosting”

Given the study's limitations, I suggest changing the title to "Translation and Preliminary Psychometric Validation of the Ghosting Questionnaire in Urdu" to emphasize the need for further validation studies.

Authors' response: We greatly appreciate your valuable input and agree with your suggestion to change the title of our study. We recognize the limitations of our research and agree that it is important to emphasize the need for further validation studies. Based on your suggestion, we have revised the title to "Translation and Preliminary Psychometric Validation of the Ghosting Questionnaire in Urdu," which accurately reflects the scope and purpose of our study. We believe that this revised title will better convey the need for future validation efforts and align with the overall goals of our research. Thank you for your valuable contribution in improving the clarity and focus of our manuscript.

Author Response File: Author Response.docx

Reviewer 2 Report

Thank you for the opportunity to review the manuscript "Translation and Psychometric Validation of the Ghosting 2 Questionnaire in Urdu". 

The main objective of this study, i.e. the adaptation and validation of Ghosting Questionnaire (GQ), a tool originally in English, in Urdu.          

The absence of a tool to explore this construct in the cultural contest of South Asia, is the main strength of the present study. Furthermore, the paper introduction and the references given support the aim of the research and contribute to a clear understanding of it.      

Some annotations:     

- In Section 1: The difference between face-to-face communication and online communication is loosely stated in the introduction. This distinction turns out to be very important for the topic of work-as later described, so it could be deepened.   
- In Section 1: The interesting relationship between attachment theory and ghosting is reported. Although relevant references are given, this section could be enriched.           
- In section 1: The paragraph from "This study" to "sociocultural context" describes the objective of the study and could be included, enriching it, in a specific paragraph. Hypotheses that guided the exploration of the psychometric properties of the scale could also be added. 
- In the section 2.2 (line 146): Criteria for calculating the power of factorial analysis and subsequent analyses should be added.       
- In section 2.2 (line 158): The statistical reference relevant to the decision to use .75 as an indication of good internal consistency of the tool should be specified.
- In section 2.2: In lines 158 to 160, results about internal coherence are presented; for greater paragraph coherence, these results could be relocated to the dedicated section (Section 3. “Results”).
- In section 2.2; the cut-off considered for the goodness-of-fit indices of the model (CFI, TLI, RMSEA, SRMR) should be reported.    
- In section 3. (lines 179-180): preliminary results of the sample adequacy (KMO test) and Bartlett's test are presented. These are important preliminary statistical analyses in the validation process of a tool, so they should also be included these specific tests in the "Data Analyses" section.          

- Overall, confirmatory statistical analyses show that the tool can be considered valid and reliable; at the same time, beyond the correlation between the two versions of the tool in Urdu and English, in-depth divergent and convergent validity analyses have not been implemented.  
The results of these analyses are generally very important in understanding whether and to what extent the construct explored by the tool is related to other similar or different constructs. Therefore, this important limitation of the validation process should be included in the specific section "Strengths and limitations".        
- Although the study focuses on describing the process of adaptation and validation of the Ghosting Questionnaire, the discussion section could be enriched presenting the research areas and the intervention contexts in which this tool can be used.

Author Response

Thank you for the opportunity to review the manuscript "Translation and Psychometric Validation of the Ghosting 2 Questionnaire in Urdu". The main objective of this study, i.e. the adaptation and validation of Ghosting Questionnaire (GQ), a tool originally in English, in Urdu. The absence of a tool to explore this construct in the cultural contest of South Asia, is the main strength of the present study. Furthermore, the paper introduction and the references given support the aim of the research and contribute to a clear understanding of it.

Authors' response: Thank you for your thoughtful concerns and comments. We sincerely appreciate your valuable recommendations and have taken them into careful consideration. We have thoroughly addressed each and every comment you provided and incorporated the necessary revisions into the manuscript. To facilitate the revisions, we have highlighted our responses to your comments in yellow throughout the revised version of the manuscript. We hope these revisions meet your expectations and contribute to the overall quality of the paper.

Some annotations:

- In Section 1: The difference between face-to-face communication and online communication is loosely stated in the introduction. This distinction turns out to be very important for the topic of work-as later described, so it could be deepened.

Authors' response: We agree with the reviewer and thus added the following paragraph: “While ghosting can technically occur in any social interaction, it is far more prevalent and problematic in online communication spaces [11-13]. Face-to-face interactions of-ten provide more social cues and context, making it harder for one party to abruptly with-draw without explanation [11-13]. Online spaces allow more anonymity and distance, which can unfortunately make ghosting an easier option [6, 12, 13]. This distinction is important to understand, as the rise of ghosting corresponds directly with the expansion of online messaging, dating, and social platforms [15]. The lack of in-person accountability online may enable more ghosting behaviors, but that does not mean ghosting should be accepted or encouraged in online spaces either [15]”.

- In Section 1: The interesting relationship between attachment theory and ghosting is reported. Although relevant references are given, this section could be enriched.

Authors' response: We provided references as follow:

• Leckfor CM, Wood NR, Slatcher RB, Hales AH. From Close to Ghost: Examining the Relationship Between the Need for Clo-sure, Intentions to Ghost, and Reactions to Being Ghosted. J Soc Pers Relat. 2023 Aug 1;40(8):2422–44.
• Richardson E, Beath A, Boag S. Default defenses: the character defenses of attachment-anxiety and attachment-avoidance. Curr Psychol. 2023 Nov 1;42(32):28755–70.

- In section 1: The paragraph from "This study" to "sociocultural context" describes the objective of the study and could be included, enriching it, in a specific paragraph. Hypotheses that guided the exploration of the psychometric properties of the scale could also be added.

Authors' response: We agree with the reviewer and thus added the following paragraph: “The hypothesis of this study is that the translated Urdu version of the GQ will demonstrate adequate reliability and validity for measuring ghosting behaviors and experiences within Urdu-speaking populations, particularly in Pakistan and India. We predict that the factor structure and psychometric properties of the translated questionnaire will be com-parable to the original English version. This will indicate that the Urdu translation is a valid tool for examining ghosting in cultural contexts where Urdu is widely spoken”.

- In the section 2.2 (line 146): Criteria for calculating the power of factorial analysis and subsequent analyses should be added.

Authors' response: We agree with the reviewer and thus added the following paragraph: “The common guidelines for social surveys recommend 5 to 10 participants per questionnaire item [21]. The current study involved 540 participants in total. There was a two-week interval between the administration of the questionnaire to the subset of participants (n = 100), whose responses were used to assess test-retest reliability. The statistical power of the factor analysis and other psychometric analyses utilized in this study was approximately 90%”.

The paragraph was added to section 2.3 instead of 2.2 due to its relevance.

- In section 2.2 (line 158): The statistical reference relevant to the decision to use .75 as an indication of good internal consistency of the tool should be specified.

Authors' response: We agree with the reviewer and we added reference # [22] to specify the indication.  

- In section 2.2: In lines 158 to 160, results about internal coherence are presented; for greater paragraph coherence, these results could be relocated to the dedicated section (Section 3. “Results”).

Authors' response: We agree with the reviewer and we relocated the paragraph to suitable location.

- In section 2.2; the cut-off considered for the goodness-of-fit indices of the model (CFI, TLI, RMSEA, SRMR) should be reported.

Authors' response: We agree with the reviewer and thus added the following paragraph: “For CFI and TLI, values above 0.90 indicate acceptable fit and values above 0.95 indicate excellent fit [22]. For RMSEA, values less than 0.08 suggest reasonable fit and values less than 0.05 suggest close fit [22]. SRMR values less than 0.08 are considered favorable. In addition to these overall model fit indices, individual factor loadings above 0.4 were considered satisfactory [22]”.

- In section 3. (lines 179-180): preliminary results of the sample adequacy (KMO test) and Bartlett's test are presented. These are important preliminary statistical analyses in the validation process of a tool, so they should also be included these specific tests in the "Data Analyses" section.

Authors' response: We agree with the reviewer and thus added the following paragraph: “The English and Urdu versions of the sample underwent rigorous evaluation to assess their adequacy for sampling [22]. The Kaiser-Meyer-Olkin (KMO) test is a statistical measure used to assess the sampling adequacy in factor analysis [22]. It determines whether the data collected for a particular sample is suitable for conducting further analysis [22]. Thus, both the English and Urdu versions of the sample were subjected to the KMO test [22]. A KMO value ranges between 0 and 1, with higher values indicating better sampling adequacy [22]. Additionally, Bartlett's test of sphericity was performed to assess whether the variables in the data were correlated [22]. This test helps determine if the data is suitable for conducting factor analysis [22]. In this study, Bartlett's test confirmed that both the English and Urdu versions had sufficient data (p < 0.001), indicating a significant correlation between the variables [22]”.

- Overall, confirmatory statistical analyses show that the tool can be considered valid and reliable; at the same time, beyond the correlation between the two versions of the tool in Urdu and English, in-depth divergent and convergent validity analyses have not been implemented. The results of these analyses are generally very important in understanding whether and to what extent the construct explored by the tool is related to other similar or different constructs. Therefore, this important limitation of the validation process should be included in the specific section "Strengths and limitations".

 Authors’ response: Table 2 was updated to include results of the CFA as follows:

CFA was conducted to evaluate the one-factor structure of the Urdu Ghosting Questionnaire. The hypothesized single factor model showed good fit with the data, as indicated by the CFI (0.991), TLI (.987), NFI (0.983) and other incremental fit indices exceeding 0.95. The RMSEA of 0.045 and SRMR of 0.018 were below recommended cutoffs, and the significant model chi-square (χ2(20) = 41.95, p = 0.003) could be attributed to the large sample size. All factor loadings were statistically significant and ranged from .812 to 1.070. The average variance extracted was 0.572, satisfying the recommended threshold for convergent validity. The unidimensional factor structure was supported, with high loadings suggesting scale items effectively measure the underlying construct of ghosting. The CFA provides empirical validation of the factorial validity and psychometric soundness of the Urdu Ghosting Questionnaire. Table 2 provides the CFA of the the Ghosting Questionnaire-Urdu (n = 504).

Table 2: Confirmatory factor structure for the Ghosting Questionnaire-Urdu (n = 504).

Notes: Confirmatory factor structure estimated using Maximum Lowlihood Extraction (MLE). Fitness measures Χ², (df) = 41.95, 20, Comparative Fit Index (CFI) = 0.991, Tucker-Lewis Index (TLI) = 0.987, Bentler-Bonett Non-normed Fit Index (NNFI) = 0.987, Bentler-Bonett Normed Fit Index (NFI) = 0.983, Parsimony Normed Fit Index (PNFI) = 0.702, Bollen's Relative Fit Index (RFI) = 0.976, Bollen's Incremental Fit Index (IFI) = 0.991, Relative Noncentrality Index (RNI) = 0.991, Log-likelihood = -5826.861, Number of free parameters = 24, Akaike (AIC) = 11701.721, Bayesian (BIC) = 11804.719, Sample-size adjusted Bayesian (SSABIC) = 11728.534, Root mean square error of approximation (RMSEA) = 0.045, RMSEA 90% CI lower bound = 0.026, RMSEA 90% CI upper bound = 0.064, RMSEA p-value = 0.638, Standardized root mean square residual (SRMR) = 0.018, Hoelter's critical N (α = .05) = 405.346, Hoelter's critical N (α = .01) = 484.59, Goodness of fit index (GFI) = 0.994, McDonald fit index (MFI) = 0.98, Expected cross validation index (ECVI) = 0.167.

We added to the discussion: “The Ghosting Questionnaire is the only existing tool assessing this construct. Due to this limitation, we were unable to examine convergence with unrelated instruments as would be ideal”.

We added to the discussion/limitations: “Finally, the current validation study involved a homogeneous sample recruited from a university population. While convenient sampling of students enabled initial evaluation of the Urdu GQ, this approach restricts the generalizability of the findings. The sample lacked diversity in terms of age, relationship status, and other demographic information that may influence experiences with ghosting. Measurement invariance across different subgroups needs to be examined in future research. Wider sampling from the community would provide a stronger evaluation of the factor structure and reliability of the Urdu GQ. This tool requires further validation using broader recruitment strategies to capture a more representative cross-section of the population. Although the student sample provides preliminary evidence of the questionnaire's adequacy, confirmation of its psychometric properties and utility for measuring ghosting experiences in the general public is needed. Ensuring cultural and demographic inclusiveness should be a priority”.

- Although the study focuses on describing the process of adaptation and validation of the Ghosting Questionnaire, the discussion section could be enriched presenting the research areas and the intervention contexts in which this tool can be used.

Authors' response: We agree with the reviewer and thus added the following paragraph: “The implications of GQ to practice are many. First, it provides a common measurement tool for researchers across Pakistan and India to systematically examine the prevalence, predictors, and impacts of ghosting in the region’s diverse populations. Both quantitative surveys and qualitative studies could employ the questionnaire to uncover cultural patterns and insights into ghosting behaviors and experiences not previously explored. Second, the availability of the Urdu questionnaire also opens possibilities for evaluating interventions aimed at reducing ghosting or mitigating its negative effects. Programs teaching healthier relationship communication skills to youth could use the questionnaire as a pre-post assessment of their impact. The scale could also assess the outcomes of public awareness campaigns or online platform policies designed to curb ghosting practices. Third, mental health professionals in the region now have access to a validated tool for helping clients cope with ghosting experiences in counseling contexts. The questionnaire provides a framework for therapists to better understand and address clients’ feelings and behaviors around ghosting”.

Author Response File: Author Response.docx

Reviewer 3 Report

The main focus of this study is to create a new language version of an existing psychometric scale that can be used with Urdu speakers in Pakistan and India. As my expertise is in quantitative methodology (the term is used here generically to refer to the related areas of research methodology, psychometrics, and statistics), I will only comment on the methodological aspects of the study.

It is commonly acknowledged that there is no single methodological approach to instrument validation that is always applicable and will yield a satisfactory measuring device. For this reason, the question of instrument validation cannot typically be answered with absolute certainty. What can be done, however, is develop a strong case for the validity of a given instrument by accumulating evidence in support of the inferences to be drawn from scores obtained with it. Hence, quality measurement requires that multiple pieces of evidence be accumulated and, for this reason, typically needs more than a single study be conducted. Certain methodological approaches can be used for this purpose such as a correlational analysis study and exploratory and confirmatory factor analyses of the scale.

As the above mentioned instrument validation approaches were well detailed in the proposed project, the research design and methodology used in this study are deemed to be solid. There are just a few specific issues included in the Results section that need to be corrected before the paper can be recommended for publication.

First, provide in the main body of the text some details on the scale of the items used in the questionnaire. Based on the range values provided in Table 1, this is clearly a Likert scale, but no details are provided about the actual range (e.g., 1 to 5 to denote ‘strongly disagree’ to ‘strongly agree’?). This will help the reader better interpret the provided mean and standard deviation values in Table 1. 

Second, it is not clear to me what the listed percentage ( %) values in Table 1 are meant to represent. Also in the “Notes” given in Table 1, the provided “Components extracted = 1; Variance explained = 62.348” are meaningless in terms of both an exploratory factor analysis (EFA) and a confirmatory factor analysis (CFA) - they should be removed from the table. These are terms used in a principal component analysis (PCA), which has absolutely noting to do with a factor analysis. The same goes for the text in line 184 , page 4 - there is no such a thing as “62.348% of the variance explained” in a EFA. The same changes and corrections need to be made in Table 2. 

Also, it is incorrect to state that “….Confirmatory factor analysis of the GQ Urdu version (GQ-U) was carried out by applying the varimax rotation method… (pg. 182). There is no such thing as a rotation in a CFA, rotation only occurs in an EFA.

Finally, in Table 3, the listed “Number of free parameters = 24” do not make sense if indeed a 1-factor model was analyzed for the data in this study. There are p = 8 items in the scale, so the number of non-redundant elements in the data correlation matrix are p (p + 1)/ 2 = 36. There are 8 factor loadings and 8 error variances or parameter estimates to be determined, so the number of free parameters should 36 - 16 = 20. Were any values for some reason constrained? If so, for what reason? Either way, the listed value 24 does not coincide with the provided analyses descriptions.

- Needs some minor revisions -

Author Response

The main focus of this study is to create a new language version of an existing psychometric scale that can be used with Urdu speakers in Pakistan and India. As my expertise is in quantitative methodology (the term is used here generically to refer to the related areas of research methodology, psychometrics, and statistics), I will only comment on the methodological aspects of the study.

Authors' response: Thank you for your thoughtful concerns and comments. We sincerely appreciate your valuable recommendations and have taken them into careful consideration. We have thoroughly addressed each and every comment you provided and incorporated the necessary revisions into the manuscript. To facilitate the revisions, we have highlighted our responses to your comments in yellow throughout the revised version of the manuscript. We hope these revisions meet your expectations and contribute to the overall quality of the paper.

It is commonly acknowledged that there is no single methodological approach to instrument validation that is always applicable and will yield a satisfactory measuring device. For this reason, the question of instrument validation cannot typically be answered with absolute certainty. What can be done, however, is develop a strong case for the validity of a given instrument by accumulating evidence in support of the inferences to be drawn from scores obtained with it. Hence, quality measurement requires that multiple pieces of evidence be accumulated and, for this reason, typically needs more than a single study be conducted. Certain methodological approaches can be used for this purpose such as a correlational analysis study and exploratory and confirmatory factor analyses of the scale.

As the above mentioned instrument validation approaches were well detailed in the proposed project, the research design and methodology used in this study are deemed to be solid. There are just a few specific issues included in the Results section that need to be corrected before the paper can be recommended for publication. First, provide in the main body of the text some details on the scale of the items used in the questionnaire. Based on the range values provided in Table 1, this is clearly a Likert scale, but no details are provided about the actual range (e.g., 1 to 5 to denote ‘strongly disagree’ to ‘strongly agree’?). This will help the reader better interpret the provided mean and standard deviation values in Table 1. 

 Authors’ response: These details have already been provided as follow: “For the purpose of this study, ghosting was defined as abruptly ending contact with-out providing a reason. The study translated and validated the English Ghosting Questionnaire (GQ) into Urdu. The translated version comprises eight items and uses a 5-point Likert scale, i.e., never (1), rarely (2), sometimes (3), often (4), and always (5). A higher score indicates greater involvement in ghosting”.

Second, it is not clear to me what the listed percentage ( %) values in Table 1 are meant to represent. Also in the “Notes” given in Table 1, the provided “Components extracted = 1; Variance explained = 62.348” are meaningless in terms of both an exploratory factor analysis (EFA) and a confirmatory factor analysis (CFA) - they should be removed from the table. These are terms used in a principal component analysis (PCA), which has absolutely noting to do with a factor analysis. The same goes for the text in line 184 , page 4 - there is no such a thing as “62.348% of the variance explained” in a EFA. The same changes and corrections need to be made in Table 2. 

 Authors’ response: The percentage values in Table 1 refer to the prevalence of ghosting experience among the participants. This clarification has been incorporated in the notes of Table 1. Components extracted and variance explained have been removed everywhere. The same changes were made to Table 2.

Also, it is incorrect to state that “….Confirmatory factor analysis of the GQ Urdu version (GQ-U) was carried out by applying the varimax rotation method… (pg. 182). There is no such thing as a rotation in a CFA, rotation only occurs in an EFA.

 Authors’ response: This has been removed. We updated CFA to as following:

CFA was conducted to evaluate the one-factor structure of the Urdu Ghosting Questionnaire. The hypothesized single factor model showed good fit with the data, as indicated by the CFI (0.991), TLI (.987), NFI (0.983) and other incremental fit indices exceeding 0.95. The RMSEA of 0.045 and SRMR of 0.018 were below recommended cutoffs, and the significant model chi-square (χ2(20) = 41.95, p = 0.003) could be attributed to the large sample size. All factor loadings were statistically significant and ranged from .812 to 1.070. The average variance extracted was 0.572, satisfying the recommended threshold for convergent validity. Discriminant validity was evidenced by the lack of correlation with any other factors. The unidimensional factor structure was supported, with high loadings suggesting scale items effectively measure the underlying construct of ghosting. The CFA provides empirical validation of the factorial validity and psychometric soundness of the Urdu Ghosting Questionnaire. Table 2 provides the CFA of the the Ghosting Questionnaire-Urdu (n = 504).

Finally, in Table 3, the listed “Number of free parameters = 24” do not make sense if indeed a 1-factor model was analyzed for the data in this study. There are p = 8 items in the scale, so the number of non-redundant elements in the data correlation matrix are p (p + 1)/ 2 = 36. There are 8 factor loadings and 8 error variances or parameter estimates to be determined, so the number of free parameters should 36 - 16 = 20. Were any values for some reason constrained? If so, for what reason? Either way, the listed value 24 does not coincide with the provided analyses descriptions.

 Authors’ response: We believe the reviewer was referring to Table 2 not Table 3.

Table 3 is the correlation between English and Urdu Ghosting Questionnaire. Table 2 described the estimates of loading.

Table 2 has been updated accordingly.

Table 2: Confirmatory factor structure for the Ghosting Questionnaire-Urdu (n = 504).

Notes: Confirmatory factor structure estimated using Maximum Lowlihood Extraction (MLE). Fitness measures Χ², (df) = 41.95, 20, Comparative Fit Index (CFI) = 0.991, Tucker-Lewis Index (TLI) = 0.987, Bentler-Bonett Non-normed Fit Index (NNFI) = 0.987, Bentler-Bonett Normed Fit Index (NFI) = 0.983, Parsimony Normed Fit Index (PNFI) = 0.702, Bollen's Relative Fit Index (RFI) = 0.976, Bollen's Incremental Fit Index (IFI) = 0.991, Relative Noncentrality Index (RNI) = 0.991, Log-likelihood = -5826.861, Number of free parameters = 24, Akaike (AIC) = 11701.721, Bayesian (BIC) = 11804.719, Sample-size adjusted Bayesian (SSABIC) = 11728.534, Root mean square error of approximation (RMSEA) = 0.045, RMSEA 90% CI lower bound = 0.026, RMSEA 90% CI upper bound = 0.064, RMSEA p-value = 0.638, Standardized root mean square residual (SRMR) = 0.018, Hoelter's critical N (α = .05) = 405.346, Hoelter's critical N (α = .01) = 484.59, Goodness of fit index (GFI) = 0.994, McDonald fit index (MFI) = 0.98, Expected cross validation index (ECVI) = 0.167.

Comments on the Quality of English Language

- Needs some minor revisions -

 Authors’ response: We have revised the manuscript for linguistic errors.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Dear authors, 

thank you for considering all the suggestions reported. 

Overall, the additions of the references requested in the introduction make the theoretical framework of your study even stronger; the added hypotheses clarify the subsequent analyses reported. 

All suggestions regarding the methodology and results section have been considered.

Overall, all thoughts and comments have been adequately considered. The changes made have certainly enriched the quality of your work. 

Reviewer 3 Report

I have read through the revised manuscript and the author responses to my initial comments on the submission. I have no further criticisms to offer and would support publication. 

Just as a note, to clarify my initial comment with regards to the listed “Number of free parameters = 24” that did to make sense to me with only having p = 8 items in the scale. I now realize that the authors were using the term “Number of Free Parameters” relative to the Information Criteria and not in relation to “Degrees of Freedom” of the tested model, which is the value that is traditionally provided in CFA analyses. One of the reasons the “Number of free parameters” is rarely used when reporting CFA results is because each SEM program computes them differently, as opposed to the degrees of freedom that are always computed in the same way. For example, in Mplus versus AMOS or lavaan, the differences are due to the fact that the intercepts are estimated in Mplus but not in Amos or lavaan.

Just a couple of minor sentence structures need editing.