|ORIGINAL RESEARCH ARTICLE
|Year : 2017 | Volume
| Issue : 1 | Page : 19-25
Patients' perceptions of argentine physicians' empathy based on the jefferson scale of patient's perceptions of physician empathy: Psychometric data and demographic differences
Raúl A Borracci1, Hernán C Doval2, Leonardo Celano3, Alejandro Ciancio4, Diego Manente5, José G. E Calderón5
1 Biostatistics, School of Medicine, Austral University, Pilar, Buenos Aires, Argentina
2 Bioethics Committee, Argentine Society of Cardiology, Buenos Aires, Argentina
3 Department of Cardiology, “Cosme Argerich” Hospital, Buenos Aires, Argentina
4 Department of Medicine, UCES University, Buenos Aires, Argentina
5 Research Area, Argentine Society of Cardiology, Buenos Aires, Argentina
|Date of Web Publication||13-Jul-2017|
Raúl A Borracci
La Pampa 3030, 1428 Buenos Aires
Source of Support: None, Conflict of Interest: None
Background: The aim of this study was to evaluate the validity of a modified Spanish version of the Jefferson Scale of Patient's Perceptions of Physician Empathy (JSPPPE) in Argentine patients and to explore how local demographic characteristics influence patients' perceptions of their physicians' empathy.
Methods: A survey was conducted in March 2013 among 400 Spanish-speaking outpatients attending three different public or private hospitals of Buenos Aires. A principal component analysis (PCA) was used to identify the JSPPPE factor structure, and a confirmatory factor analysis (CFA) was employed to evaluate its construct validity. Demographic variables including age, gender, geographic origin, education, health coverage, regular physician-established and patient-perceived health status were used to find what factors may influence empathy rating.
Results: The PCA yielded a one-factor model that accounted for 77.5% of the variance, and an adequate model fit was observed with CFA indices. Male and elderly patients, South American descendants, less educated people, and public hospital attendants were associated with a higher JSPPPE score. Patients perceived a lower interest of physicians in their daily problems and a poorer capacity “to stand in their shoes.”
Discussion: The JSPPPE provides a valid score to measure patients' perceptions of physician empathy in Argentina. These findings afford insight into Argentine patients' awareness of their doctors' empathic concern; however, JSPPPE scores may be alternatively interpreted in terms of patients' satisfaction or likeability.
Keywords: Argentina, empathy, Jefferson Scale of Patient's Perceptions of Physician Empathy, patients, physicians
|How to cite this article:|
Borracci RA, Doval HC, Celano L, Ciancio A, Manente D, Calderón JG. Patients' perceptions of argentine physicians' empathy based on the jefferson scale of patient's perceptions of physician empathy: Psychometric data and demographic differences. Educ Health 2017;30:19-25
|How to cite this URL:|
Borracci RA, Doval HC, Celano L, Ciancio A, Manente D, Calderón JG. Patients' perceptions of argentine physicians' empathy based on the jefferson scale of patient's perceptions of physician empathy: Psychometric data and demographic differences. Educ Health [serial online] 2017 [cited 2017 Sep 21];30:19-25. Available from: http://www.educationforhealth.net/text.asp?2017/30/1/19/210513
| Background|| |
Recent research suggests that empathy between patient and physician positively influences healing and seems to be associated with better clinical outcomes., Specifically, cognitive empathy is a well-known concept associated with these communication skills which has been assessed with different quantitative tools. The Jefferson Scale of Physician Empathy (JSPE) is a worldwide validated instrument used to measure physicians' self-reported empathy., Through a multidimensional approach, this scale allows physicians to estimate their capacities to understand patients' experiences, concerns, and perspectives, combined with the skill to communicate this understanding with an intention to help. However, since physicians may overestimate their own empathic engagement, patients' perceptions about physician empathy must also be considered. With this purpose, the Jefferson Scale of Patient's Perceptions of Physician Empathy (JSPPPE) was developed and validated by Kane et al. in 2007. This type of instrument is important to compare and examine the agreement between patient and physician perception of empathy, and preliminary research has reported a moderate correlation between JSPE and JSPPPE scores. Lack of correlation between patient and doctor empathic understanding could be explained by a real difference of perceptions or by the inaccuracy of the tool.
Most researches on empathy have been focused on the physicians' point of view, whereas the patients' perspective on the topic has been scarcely investigated, especially the differences in perception due to age, gender, health status, education, and other social factors.,
The aim of this study was to evaluate the validity of a modified Spanish version of the JSPPPE in Argentine patients and to explore how local demographic characteristics influence patients' perceptions of their physicians' empathy.
| Methods|| |
Setting and population
A survey was conducted among a nonrandomized sample of 400 Spanish-speaking outpatients attending three different public or private hospitals of Buenos Aires in March 2013. All patients consulting during the study period were considered; the exclusion criteria were dementia, aphasia, no consent or unwillingness to participate, or any difficulty with the Spanish language that would prevent their understanding or ability to answer the questionnaire. Patients invited to participate were informed that this study would explore doctors' quality of care and that their responses would have no impact on their medical care. In addition, they were assured of the complete confidentiality of their responses and their verbal consent was obtained. The participation in the study was anonymous and voluntary; there was no economic incentive for completing the questionnaire, which took < 10 min to answer.
Measurements and instruments
The JSPPPE instrument includes 5 items, each answered on a 5-point Likert scale, ranging from 1 = strongly disagree to 5 = strongly agree. Two researchers (RAB, HCD) independently translated the original English version of the JSPPPE into Spanish. They discussed and resolved the differences in their translations and reached consensus on the best initial wording. The product was back-translated by one independent bilingual translator who was unaware of the original English version. Back translation confirmed the original meaning, and for many statements, the exact original wording was recovered. Items were left in the same order as in the original, and the same 5-point Likert response structure was used. In addition, two conceptually relevant items were added to the Jefferson scale: one as an extra criterion of empathic patient–physician relationship measurement and the other as a validation question. The two additional questions had the following structure:
- Does the doctor devote enough time to me? (extra criterion)
- Would you recommend the doctor to your family or friends? (validation question).
We named the original Jefferson scale as the “5-item JSPPPE” and the one including the previously defined extra criterion question as the “6-item JSPPPE.” The validation question was excluded from the scales. Patients were provided with a form containing the 6-item JSPPPE version and the validation question to be answered. Since the 5-item and 6-item JSPPPE have a response set that ranges from 1 to 5, a summed score was calculated with a maximum of 25 and 30, respectively (the higher the score, the greater the empathy). The sample included patients attending an urban public hospital, an urban private medical center, and a suburban clinic. The medical services were selected to compare the most heterogeneous population that may represent the social structure of Buenos Aires.
The survey was completed with different demographic variables to characterize the sample, including age, gender, geographic origin, education, health coverage, attending hospital, and regular doctor. These variables were used to find what factors may bias empathy rating. In addition, perceived health status was tested through a 5-option Likert scale ranging from bad (1) to excellent (5).
Data obtained were anonymously included and processed in a database. The project was reviewed and approved by the Ethical Review Board of the Argentine Society of Cardiology (ERB: #SAC-016).
The validity framework for the present study was based on Messick , and Kane  assumptions. First, validity is not a test property; rather, it refers to the test use for a particular purpose. Second, evaluation of the usefulness and appropriateness of a test for a particular purpose requires multiple sources of evidence. Though Messick argued that all forms of validity should be considered the aspects of test construct validity, in the present study, criterion-related validity was also used as a traditional source of evidence. Two additional plausible assumptions were adopted in this study: (1) it was assumed that JSPPPE scores were relatively invariant over subgroups and (2) that the alpha coefficient could be used to evaluate the scale generalizability. On the other hand, construct validity was assessed by describing an underlying latent variable, known as perceived empathy, that the JSPPPE tried to measure. This latent variable was included in the path diagram of the confirmatory factor analysis (CFA). Construct validity is also based on any evidence that bears on the meaning of the test scores. Consequently, in the present study, score meaning was explored by studying the expected performance differences across groups based on demographic characteristics. Unintended social consequences are a basic concern in the validation of testing programs, especially under the unified validity perspective. An interpretation-and-use model was adopted to explain the role of unintended consequences in validity. Since we suggested that the JSPPPE may be useful for evaluating empathy perception, the differential impact against particular groups was taken into account.
Due to the questionnaire's structure, the occurrence of ceiling effect was expected, and the non-Gaussian distribution of empathy scores was assessed with the Kolmogorov-–Smirnov goodness-of-fit test. The median score (i.e., the sum) was adopted as a representative value; nevertheless, the mean and standard deviation (SD) were included in factor analysis. The principal component analysis (PCA) was used to explore the dimensionality of the JSPPPE. Only factors with eigenvalues >1.25 were preserved, and factor coefficients >0.40 were required for the interpretation of factor structure, using Varimax rotation. The PCA criteria to identify factor structure were examined with the Kaiser–Meyer–Olkin (KMO) analysis. A KMO index >0.50 was adopted to establish whether the data set was suitable for factor analysis., To evaluate the appropriateness of PCA for nonnormal distributions, PCA was performed on raw response data, on transformed data with log10 (1/x) for negative-skewed distributions, and with the Templeton two-step approach. Furthermore, raw response data were also analyzed with multiple correspondence analyses (MCAs). MCA is a factorial method related to PCA for multiple nominal data analysis. Since MCA is not restricted to normal distributions, this tool may be used to estimate PCA appropriateness for evaluating the present data. LISREL 9.20 (Scientific Software International, Inc., Skokie, IL, USA) software was used to test the original one-factor structure of the JSPPPE by a CFA. A CFA investigates how the data fit into a predetermined and constructed model by presenting the relationship between the data in the model and the estimation of errors. Assessment of model data fit was done using model Chi-square goodness-of-fit and approximate fit indices. A nonsignificant Chi-square test (P > 0.05) indicates model fit; nevertheless, Chi-square values are very sensitive to sample size, especially when the sample size is below 100 or above 200. The additional approximate fit indices employed included: goodness-of-fit index (GFI), adjusted GFI (AGFI), normed fit index (NFI), non-NFI (Tucker–Lewis index) (NNFI), relative fit index (RFI), incremental fit index (IFI), and comparative fit index (CFI). Values >0.9 arising from the GFI, AGFI, NFI, NNFI, RFI, IFI, and CFI indicate model fit; conversely, values ≥0.85 represent acceptable model fit. Other indices calculated were the root mean square error of approximation (RMSEA) and the root mean square residual (RMR), in which values <0.08 indicate a reasonable model fit., The internal consistency of the scale was assessed with Cronbach's alpha, and a value >0.70 was considered acceptable. To counteract the skewed distribution and the potential ceiling effect, a nonparametric analysis was adopted; then, mean per item or median sum-score and mean rank (nonparametric test results) were included in the tables. Two-tailed Mann–Whitney U-test and Kruskal–Wallis nonparametric tests were used to compare nonnormal score distributions. Effect size for differences between mean ranks by demographic factors was assessed with Cohen's d index. Usual standards of small, medium, and large effect sizes measured with Cohen's d index were, respectively, 0.2, 0.5, and 0.8. Item-total correlation and correlation between the 5- and 6-item JSPPPE scores and the validation question were calculated with Spearman's rho. Demographic factors associated with total JSPPPE scores were assessed with multivariate linear regression analysis. Nonmetric independent variables were transformed and included in the linear analysis as fictitious variables. Sample size reached a ratio of 50 cases (individuals) per variable (item). All statistical analyses, except CFA, were performed using SPSS Statistics for Windows, Version 17.0 (SPSS, Inc., Chicago, IL, USA).
| Results|| |
Out of the 400 surveyed patients, 310 (77.5%) completed the questionnaire with no missing data. The population characteristics are shown in [Table 1]. Mean population age was 49.6 (SD: 17.2) years. Sample distribution according to family origin, educational attainment, and health coverage was representative of the Buenos Aires city population.
|Table 1: Baseline characteristics of 310 surveyed patients answering the Jefferson Scale of Patient's Perceptions of Physician Empathy in Argentina (percentages)|
Click here to view
The distribution of JSPPPE scores was J shaped, and 20.3% (n = 63) of patients provided the maximum score. The KMO measures of sampling adequacy (0.89 and 0.91 for the 5- and 6-item JSPPPE, respectively) and Bartlett's tests of sphericity were significant (χ2(10) = 1222, P = 0.0001 and χ2(15)= 1591, P = 0.0001, respectively), establishing the suitability of the data set for factor analysis. The 5- and 6-item PCA yielded a one-factor model that accounted for 77.5 and 76.5% of the variances, respectively [Table 2]. PCA performed on raw and transformed data, as well as MCA performed on raw data demonstrated equivalent outcomes (comparative data not shown); hence, only PCA for raw data has been reported. The item-total score correlation ranged from a low of 0.698 for the statement “asks about what is happening in my daily life” to a high of 0.864 for “is an understanding doctor,” indicating a substantial contribution of all items to the total score [Table 2]. Inter-item correlations ranged from 0.622 to 0.820 with a mean value of 0.716. Spearman's correlation coefficient (rho) between the total score of the 5-item JSPPPE and the score of the validation question was 0.749 (P = 0.0001), while that between the 6-item JSPPPE and the validation question was 0.774 (P = 0.0001). These results support the criterion-related validity of the JSPPPE in our population. Finally, Cronbach's alpha coefficients were 0.92 and 0.94 for the 5- and 6-item scales, respectively.
|Table 2: Principal component analysis, internal consistency, and validation of items in the Jefferson Scale of Patient's Perceptions of Physician Empathy|
Click here to view
A CFA was conducted on the original 5-item JSPPPE questionnaire to determine its construct validity. The present data fitted the one-factor JSPPPE structure. Though a low goodness-of-fit statistic was indicated by a significant maximum likelihood ratio Chi-square (P = 0.001), it is well known that Chi-square values are very sensitive to sample size and are not appropriate for analyzing when the sample size exceeds 200. On the contrary, an adequate fit was observed with the other indices: GFI (0.974), AGFI (0.922), NFI (0.988), NNFI (0.982), RFI (0.976), IFI (0.991), and CFI (0.991), all with values >0.90. In addition, RMSEA (0.098) and RMR (0.014) also indicated a fairly acceptable fit. For a comparative purpose, CFA was also conducted on the 6-item JSPPPE instrument. Again, a significant maximum likelihood ratio Chi-square (P< 0.0001) showed a low goodness-of-fit, whereas the rest of indices showed an adequate or acceptable model fit: GFI (0.913), AGFI (0.797), NFI (0.963), NNFI (0.946), RFI (0.939), IFI (0.968), CFI (0.967), RMSEA (0.158), and RMR (0.084).
[Table 3] summarizes the median score and mean rank per item obtained with both JSPPPE versions. The item “asks about what is happening in my daily life” showed the lowest rank; while the statement “is an understanding doctor” got the highest position in both JSPPPE models. The 5-item JSPPPE mean rank score for women was 144.7 versus 174.8 for men (P = 0.004) (Cohen's d index: 0.58) while the 6-item JSPPPE was 144.5 versus 175.3 for women and men, respectively; P = 0.003 (Cohen's d index: 0.60). The comparison of the mean rank score according to age (quartiles) showed a significant association between older age and higher empathy score [Table 4]a; nevertheless, effect size assessed with Cohen's d index ranged from low to medium/large. In contrast, higher educational attainment was associated with a lower JSPPPE score and an overall medium effect size [Table 4]b. Patients attending a public hospital showed higher perception of physician empathy than those attending a private center (P = 0.0001) with a medium/large effect size. This trend remained when data were adjusted by gender, age, and other covariates [Table 4]c. The individuals who had a regular personal doctor reported a 5-item JSPPPE mean rank of 173.3 versus 135.5 for those without a regular doctor (P = 0.0001) (Cohen's d index: 0.73) and 174.1 versus 134.6; P = 0.0001 in the 6-item JSPPPE (Cohen's d index: 0.76). People of South American ancestry had a 5-item JSPPPE mean rank of 174.6 versus 148.7 for people of European ancestry (P = 0.045) (Cohen's d index: 0.50) and 174.2 versus 148.8; P = 0.050 in the 6-item JSPPPE (Cohen's d index: 0.49). Finally, self-reported health status and health coverage were not significantly associated with different JSPPPE scores.
|Table 3: Mean rank and mean score per item obtained with the Jefferson Scale of Patient's Perceptions of Physician Empathy|
Click here to view
|Table 4: Comparison of responses with the Jefferson Scale of Patient's Perceptions of Physician Empathy according to (a) age-quartiles, (b) educational attainment, and (c) medical center (the latter adjusted by age, gender, and having a regular doctor)|
Click here to view
Multivariate analysis found that the variables male (odds ratio [OR]: 6.4, 95% confidence interval [CI]: 1.93–21.2), South American ancestry (OR: 5.5, 95% CI: 1.15–26.2), having a regular doctor (OR: 11.8, 95% CI: 3.69–37.5), and attending a public urban hospital (OR: 7.3, 95% CI: 2.05–25.8) were associated with higher total 5-item JSPPPE scores. Similarly, the variables associated with higher 6-item JSPPPE scores were male (OR: 9.6, 95% CI: 2.33–39.9), South American ancestry (OR: 7.1, 95% CI: 1.12–45.4), having a regular doctor (OR: 20.7, 95% CI: 5.25–81.8), and attending a public urban hospital (OR: 10.6, 95% CI: 2.37–47.6).
| Discussion|| |
The results of this study support the JSPPPE reliability and validity for measuring patients' perceptions of Argentine physicians' empathy. The resulting factor structure of this Spanish version answered by our patients was very similar to that reported by the Jefferson scale developers. The PCA did not result in the exclusion of any item and identified the one-factor model previously described. Likewise, psychometric data such as KMO indexes, sphericities, variances, and eigenvalues assigned to the 5- and 6-item scales were similar to those previously reported. The additional question of the 6-item model did not improve the validity and reliability of the JSPPPE original questionnaire, especially considering the nonsignificant difference in the percentage explained variance and Cronbach's alpha coefficient. However, this extra item yielded the highest correlation coefficient regarding the validation question. Some concerns may exist on the criterion-related validity since the response to the validation item may have been highly influenced by the answers to the first 5 items. Nevertheless, the construct validity was confirmed by the performance of most goodness-of-fit indices determined by the CFA.
It is crucial to emphasize that the JSPPPE measures patients' perceptions of physician empathy and not the real physician's empathy. Furthermore, the perceived empathic level could also vary as a function of physician characteristics. As Messick, Kane, and others have observed, it is essentially impossible not only to prove that test interpretations are valid but also to prove that the researcher is measuring what he/she thinks he/she is measuring. Hence, though the original JSPPPE was built to assess empathy, its scores may be alternatively interpreted in terms of patients' satisfaction or likeability.
The surveyed patients assigned the lowest scores to the items “the doctor asks about what is happening in my daily life” and “the doctor can view things from my perspective.” These results showed that respondents perceived the physicians' lower interest in patients' daily problems and a reduced skill “to stand in their shoes” or see things as patients see them [Table 3]. In contrast, a report on Argentine physicians' empathy based on the JSPE indicated that doctors overestimated their capacity to handle both these aspects. Lin et al. also observed that emergency physicians and their patients perceive physician empathy differently. Specifically, when patients expressed their feelings, doctors usually did not resonate with their concerns.
Females, younger patients, and individuals with higher educational attainment seem to be more demanding when evaluating physician empathy. Male patients assigned higher JSPPPE scores than women, and this difference remained when values were adjusted by age. Perhaps, this gender difference may be explained by the fact that empathy expectations are lower in males than in females. In addition, a close positive association between age and JSPPPE scores was observed. Older patients provided higher empathy scores than younger individuals. On the other hand, higher educational attainment was associated with lower empathy scores. After adjusting for age, gender, and other covariates, patients attending an urban public hospital had higher perception of doctors' empathy than those surveyed in private centers. These findings could be explained by a more demanding attitude of patients toward doctors in the private health-care system or a better professional engagement of public hospital physicians.
As presumed, patients who had a regular family doctor showed better perception of their doctors' empathy. In contrast, although individuals perceiving a better health status could be expected to have higher awareness of physician empathic engagement, an association between the patients' self-reported health status and the JSPPPE scores was not found. From the ethnic viewpoint, people of South American ancestry seemed to be less demanding when scoring physician empathy. Compared with European descendants, South American patients assigned higher JSPPPE scores, even when data were adjusted by covariates. While some significant differences between JSPPPE scores seemed to be marginal, most effect sizes assessed with Cohen's d index showed a medium effect.
Finally, multivariate analysis summarized that variables such as male, South American ancestry, having a regular doctor, and attending a public urban hospital were significantly associated with higher total 5- and 6-item JSPPPE scores.
Pollak et al. examined patients' perceptions of oncologists' empathy and trust. They found that patients with low economic security were most likely to rate their doctors as more empathic than patients with high economic security. Furthermore, patients knowing their oncologist longer than a year gave higher trust ratings than individuals who knew their doctor for less time. Although health coverage was not significantly associated with JSPPPE scores in our study, patients attending an urban public hospital had, on an average, less economic security than individuals attending a private health center. These features could also explain the different perceptions of empathy according to the place of attendance. Recently, Al Onazi et al. studied the perception of health-care providers' empathy in hemodialyzed Arabian patients. As in our work, they found that older patients and individuals with lower educational attainment perceived higher empathy levels than younger patients and more highly educated individuals. In contrast, female patients' perception of empathy was significantly higher than that of male individuals. Cultural features of Arabian females could help explain differences with our findings.
One limitation of this study was that the self-reported data of the JSPPPE may limit the interpretation of the findings since respondents may over or underestimate their own perception of empathy. Moreover, no attempt was done to compare patients' versus physicians' empathy perceptions in the same sample. Since variance due to physicians was not calculated, JSPPPE reliability could not be properly established. Regarding reliability, this study only measured internal consistency of the scale scores for a single patient–physician encounter; on the contrary, the relationship between scores of patients visiting the same doctor was not evaluated. Furthermore, time spent by a doctor with a patient could generate some meaningful difference between patient groups, especially for the extra criterion answer (“does the doctor devote enough time to me?”); this possibility, however, was not evaluated. In the case of the instrument limitations, empathy should probably be represented as a multidimensional structure, though the developers of the original JSPPPE found that factor analysis indicated that the scale was measuring a single factor of emphatic engagement.
| Conclusions|| |
This study showed that the Spanish version of the JSPPPE provides a valid score to measure patients' perceptions of physicians' empathy in Argentina. In general, patients perceived that physicians had lower interest about patients' daily problems and a lower capacity “to stand in their shoes.” The higher JSPPPE scores given by males, elderly patients, South American descendants, less educated patients, and public hospital attendants might simply reflect their lower empathy expectation. In short, these findings provide an insight into Argentine patients' awareness of their doctors' empathic concern though JSPPPE scores may be alternatively interpreted in terms of patients' satisfaction or likeability. Positive social consequences of the demographic differences found with the JSPPPE scores may be considered as a warning to pay more attention to groups with lower expectations in terms of physician empathy. On the contrary, unintended consequences could arise from the inequity of offering a lower quality of care to those patients, based on their known limited demands.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Rakel DP, Hoeft TJ, Barrett BP, Chewning BA, Craig BM, Niu M. Practitioner empathy and the duration of the common cold. Fam Med 2009;41:494-501.
Hojat M, Louis DZ, Markham FW, Wender R, Rabinowitz C, Gonnella JS. Physicians' empathy and clinical outcomes for diabetic patients. Acad Med 2011;86:359-64.
Hojat M, Mangione S, Nasca TJ, Cohen MJ, Gonnella JS, Erdmann JB, et al
. The Jefferson Scale of Physician Empathy: Development and preliminary psychometric data. Educ Psychol Meas 2001;61:349-65.
Hojat M, Gonnella JS, Nasca TJ, Mangione S, Vergare M, Magee M. Physician empathy: Definition, components, measurement, and relationship to gender and specialty. Am J Psychiatry 2002;159:1563-9.
Hojat M. Ten approaches for enhancing empathy in health and human services cultures. J Health Hum Serv Adm 2009;31:412-50.
Kane GC, Gotto JL, Mangione S, West S, Hojat M. Jefferson Scale of Patient's Perceptions of Physician Empathy: Preliminary psychometric data. Croat Med J 2007;48:81-6.
Glaser KM, Markham FW, Adler HM, McManus PR, Hojat M. Relationships between scores on the Jefferson Scale of physician empathy, patient perceptions of physician empathy, and humanistic approaches to patient care: A validity study. Med Sci Monit 2007;13:CR291-4.
Lin CS, Hsu MY, Chong CF. Differences between emergency patients and their doctors in the perception of physician empathy: Implications for medical education. Educ Health (Abingdon) 2008;21:144.
Hojat M, Louis DZ, Maxwell K, Markham F, Wender R, Gonnella JS. Patient perceptions of physician empathy, satisfaction with physician, interpersonal trust, and compliance. Int J Med Educ 2010;1:83-7.
Messick S. Meaning and values in test validation: The science and ethics of assessment. Educ Res 1989;18:5-11.
Messick S. Validity. In: Linn R, editor. Educational Measurement. 3rd
ed. Washington, DC: American Council on Education/Macmillan; 1989, p. 13-103.
Kane MT. Explicating validity. Assessment in Education: Principles, Policy & Practice 2016; 23:198-211.
Sireci SG. On validity theory and test validation. Educ Res 2007;36:477-81.
Henson R, Roberts J. Use of exploratory factor analysis in published research: Common errors and some comment on improved practice. Educ Psychol Meas 2006;66:393-416.
Hogarty K, Hines C, Kromrey J, Ferron J, Mumford K. The quality of factor solutions in exploratory factor analysis: The influence of sample size, communality, and over determination. Educ Psychol Meas 2005;65:202-26.
Templeton GF. A two-step approach for transforming continuous variables to normal: Implications and recommendations for IS research. Commun Assoc Inf Syst 2011;28:41.
Albright JJ, Hun MP. Confirmatory Factor Analysis Using Amos, LISREL, Mplus, and SAS/STAT CALIS. Working Paper. The University Information Technology Services (UITS) Center for Statistical and Mathematical Computing, Indiana University; 2009. Retrieved June, 6, 2016, from http://www.indiana.edu/~statmath/stat/all/cfa/index.html
Vaughan B, Mulcahy J, McLaughlin P. The DREEM, part 2: Psychometric properties in an osteopathic student population. BMC Med Educ 2014;14:100.
Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate Data Analysis. 5th
ed. New Jersey: Prentice Hall International; 1998.
Schermelleh-Engel K, Moosbrugger H, Müller H. Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods Psychol Res Online 2003;8:23-74.
Sullivan GM, Feinn R. Using effect size-or why thePvalue is not enough. J Grad Med Educ 2012;4:279-82.
Borracci RA, Doval HC, Nuñez C, Samarelli M, Tamini S, Tanus E. Measurement of empathy among Argentine cardiologists: Psychometrics and differences by age, gender, and subspecialty. Cardiol J 2015;22:52-6.
Pollak KI, Arnold R, Alexander SC, Jeffreys AS, Olsen MK, Abernethy AP, et al.
Do patient attributes predict oncologist empathic responses and patient perceptions of empathy? Support Care Cancer 2010;18:1405-11.
Al Onazi M, Al Jondeby M, Azeem M, Al Sayyari A. Factors affecting Saudi hemodialysis patients' perception of healthcare providers' empathy. Arab J Nephrol Transplant 2011;4:71-6.
[Table 1], [Table 2], [Table 3], [Table 4]