|Year : 2008 | Volume
| Issue : 1 | Page : 49-54
Psychometric evaluation of a hindi version of positive-negative affect schedule
R Pandey1, N Srivastava2
1 Reader, Department of Psychology, Banaras Hindu University, Varanasi-221005, India
2 Research Scholar, Dept. of Psychology, M. G. Kashi Vidyapith, Varanasi, India
|Date of Web Publication||13-May-2010|
Reader, Department of Psychology, Banaras Hindu University, Varanasi-221005
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background - The present paper reports the development and psychometric evaluation of a Hindi version of the Positive-Negative Affect Schedule (PANAS) originally developed in English by Watson, Clark, and Tellegen (1988). The PANAS is widely used tool for assessment of positive and negative affect in clinical as well as non-clinical setting and has also been used as a differential diagnostic tool for distinguishing the clinical depression from anxiety. Material & Method - A Hindi version of the PANAS (PANAS-H) was developed using the contemporary psychometric standards for developing transliteral equivalents and cross-cultural adaptation of psychological tests/scales. In order to evaluate the psychometric properties, the PANAS-H was administered on a relatively heterogeneous sample of 179 participants. The obtained data was subjected to an exploratory factor analysis (principal component analysis) which identified two theoretically significant orthogonal factors. The mood adjectives reflecting the positive affect loaded significantly on factor-1 whereas the affective lexicons representing the negative emotional engagement loaded significantly on factor-2. On the basis of this pattern of factor loading the first factor was labeled as 'Positive Affect' (PA) and the second factor as 'Negative Affect' (NA). Results : A significant but low negative correlation was observed between PA and NA which suggests that PA and NA are not independent of each other. Item analysis done for each subscales revealed that the Hindi affective lexicon used for tapping the dimensions of PA and NA are reliable and valid and form a homogeneous item-pool. Conclusion : The reliability of the PA and NA subscales as well as that of the whole scale was found to be highly satisfactory (0.804 for PA, 0.776 for NA, and 0.658 for full scale). Overall, the findings suggest that 1) the PANAS-H can reliably and validly measure the PA and NA of Hindi speaking individuals, and 2) the PANAS-H measures two distinct (PA and NA) but negatively related dimensions of affect.
|How to cite this article:|
Pandey R, Srivastava N. Psychometric evaluation of a hindi version of positive-negative affect schedule. Ind Psychiatry J 2008;17:49-54
The Positive and Negative Affect Schedule (PANAS) is a 20-item selfreport measure of positive and negative affect developed by Watson, Clark, and Tellegen (1988) and has been widely used for diverse purpose in both clinical and non-clinical setting. The positive affect (PA) reflects the pleasurable engagement and subjective experience of happiness whereas the negative affect (NA) subscale measures the level of subjective distress and unpleasurable engagement. Emotions such as enthusiasm and alertness are indicative of high PA, whilst lethargy and sadness characterize low PA (Watson & Clark, 1984).
The PANAS has been used as a clinical tool in the area of depression and most popularly to distinguish between clinical anxiety and depression. Researchers have pointed out that it is difficult to distinguish between anxiety and depression through the use of conventional self-report scales of anxiety and depression (Clark & Watson, 1991a, b) because such scales predominantly measure negative affect, a factor common to both anxiety and depression (Watson & Clark, 1984). The tripartite model of depression (Clark & Watson, 1991b) suggest that since negative affect is common to both anxiety and depression, the two clinical states can be differentiated on the basis of some factors or components specific to anxiety or depression. According to this model, the physiological arousal is the specific component of anxiety and the anhedonia (low PA) is the specific component of depression. Thus, as a measure of PA the PANAS can be reliably used to distinguish anxiety and depression.
The clinical use of PA and NA scales in differentiating clinical anxiety and depression has been demonstrated in several researches. For example, Dyck, Jolly, and Kramer (1994) using a sample of psychiatric patients (N=162) found that NA factor, and not the PA factor, significantly contributed to predicting anxiety, but both factors significantly predicted depression. Similarly, Jolly, Dyck, Kramer, and Wherry (1994) observed that when NA was controlled, depression, but not anxiety, was related to PA scores. These observations suggest that NA is a stronger predictor of anxiety and PA a better predictor of the depression. These findings also imply that the PANAS along with the measures of anxiety and depression can be used for differential diagnosis of the clinical states of anxiety and depression. Besides its use for clinical depression and anxiety, the PANAS has been applied in other psychiatric population and for understanding various types of behavioral problems such as eating disorder (Holt & Ricciardell, 2002).
Apart form using the PANAS as a differential diagnostic tool for differentiating anxiety and depression it has also been used to measure the affective component (PA and NA) of subjective well-being (SWB) in various researches (Diener, Suh, & Oshi, 1997). The most popular conceptual model of SWB suggests that there are two primary components of SWB- the cognitive component and the affective component. The former deals with the individual's subjective evaluation of his/her life (life satisfaction) whereas the latter consists of the prevalence of PA and relative absence of NA in one's life (Diener, Scollon, & Lucas, 2003; Diener, Suh, Lucas, & Smith, 1999).
The clinical and non-clinical utility and the brevity of the PANAS has made it a widely popular tool in the field of psychology and psychiatry. Given its wide popularity, attempts have been made to adapt it in non-western culture and languages. For example, recently Yamasaki and associates (Yamasaki, Katsuma, & Sakai, 2006) have reported the Japanese adaptation of the children version of the PANAS. The wide applicability of the PANAS lead researchers to develop short versions (Mackinnon, Jorm, Christensen, Korten, Jacomb, & Rodgers, 1999) and versions suitable for children (Laurent, Catanzaro, Joiner, Rudolph, 1999).
Apart form adaptation, and development of short versions, attempts have also been made to evaluate its psychometric properties. For example, several studies have been conducted to verify the factor structure of the PANAS using both exploratory (Watson et al., 1988) and confirmatory factor analyses (Crocker, 1997). These studies, in general, have reported that the two factor structure of the PANAS is robust and stable across various types of samples. However, some researchers have found that alternative factor solution best fit the data as opposed to the two factor solution of Watson and associates (1988). For example, Mehrabian (1998) found that a complex hierarchical structure represented a superior fit in comparison to Watson and associates' (1988) hypothesized two-factor structure. Recently, an attempt was made by Crawford and Henry (2004) to resolve this controversy relating the factor structure of the PANAS and the independence of its two subscales (PA and NA). The findings of confirmatory factor analysis revealed that the PA and NA scales of the PANAS measure two distinct but negatively correlated factors.
The PANAS has been found to be highly reliable in various clinical and non-clinical trials. For example, Watson and associates (1998) examined the reliability of the PANAS using different time frames ranging from 'right now' to 'during the last year'. They reported that the reliability of PA subscale ranged from .86 to .90 whereas those of the NA scale from .84 to .87. Attempts to evaluate the reliability of this scale in clinical setting have also been made. For example, Ostir and associates ( Ostir, Smith, P.M., Smith, D., & Ottenbacher, 2005) examined the reliability of the PANAS in a sample of patients receiving medical rehabilitation and found that the PANAS has excellent reliability. Similar findings of reliability has been reported by others also using clinical and non-clinical samples ( e.g., Crawford & Henry, 2004; Jolly et al., 1994; Mehrabian, 1998; Roesch, 1998)
The foregoing review suggests that the PANAS is a psychometrically sound and useful tool not only for the diagnosis/differential diagnosis of various clinical conditions involving or having implication of pathology of affect but also for assessing the happiness and well-being of clinical and non-clinical population. However, as psychometric evaluation and refinement of measures is an ongoing process (Nunally, 1967) and the present review suggest that there is some controversy regarding the factor structure of the PANAS, attempts should be made to examine the psychometric properties and factor structure of the PANAS in other cultures and languages.
Such an attempt becomes important for two reasons. First, as a measure of affective engagement the PANAS scores and its psychometric properties are likely to be influenced by cultural and linguistic variations. Numerous crosscultural studies have demonstrated that people from different cultures vary in terms of emotional perception, experience and expression. For example, it has been found that people from individualistic culture (like America) most frequently experience such 'ego focused' emotions as pride, anger frustration, whereas those from collectivistic culture most frequently engage themselves in such 'other focused' emotions as sympathy and shame (Markus & Kitayama, 1991, 1994). Similarly, it has been noted that in reporting one's happiness and positive emotions people from East Asian culture show a hesitant attitude as compared to Western culture (Suh, 2000). As the PANAS involves rating of positive and negative emotions most of which have a selfreference (self focused emotions), it would be interesting to explore whether the two factor structure (including PA and NA) emerges with a sample of East Asian culture (India in the present context) using affective adjectives in Hindi. The second reason to examine the psychometric properties (particularly the factor structure) of the PANAS is the existing controversy regarding the nature and number of factors.
In the light of the wide applicability of the PANAS and the need to evaluate its psychometric properties the present study makes an attempt to develop a Hindi version the scale and explore its psychometric properties and factor structure on an Indian sample. Apart from demonstrating the cultural fairness and linguistic independence of the PANAS, the Hindi version of the PANAS will help the psychiatrists and psychologists in making diagnosis and assessment of Hindi speaking individuals.
| Method|| |
Sample: The psychometric evaluation of the PANAS was done on an incidentally sampled group of 179 Hindi Speaking participants. Out of the 179 participants, 112 were males and the remaining 67 were females. All the participants belonged to middle class family and had an education level of undergraduate or above. The sample was relatively heterogeneous with regard to age and occupational background. The age of the participants ranged from 20 to 60 years. As far as the occupation is concerned, the present sample included, students, engineers, doctors, bank employees, clerical staff and officers from various public and private sectors, as well as self employed and businessmen. The attempt to get a relatively heterogeneous sample was made in order to ensure wide variations in response so that reliability and validity of the scale is not artificially truncated because of restricted response variance.
The Positive and Negative Affect Schedule (PANAS, Watson et al., 1988) consists of two 10 item subscales which provide a brief measure of PA and NA. The items consist of mood adjectives that tap affective lexicon related to PA and NA. Respondents are asked to rate the extent to which they have experienced each particular emotion within a specified time period on a five point scale ranging from 1 to 5. The scale points are: 1 'very slightly or not at all', 2 'a little', 3 'moderately', 4 'quite a bit' and 5 'very much'. A number of different time-frames have been used with the PANAS, but in the current study the time-frame adopted was 'during the past month'.
Before evaluating the psychometric properties of the PANAS its Hindi version was developed. The items of the English version of the PANAS were translated in Hindi and the Hindi version of the scale was submitted to a panel of three subject experts for evaluating the adequacy of the translation. The experts were asked to rate the adequacy of translated item content on a 3-point scale ranging from 1 (not adequate at all), 2 (moderately adequate) and 3 (adequate) and to give suggestion to improve the content of those items that received a rating of less than 3. The comments of the experts were reviewed. All the experts suggested that it will be more appropriate to the present the mood adjectives of the PANAS in a Hindi phrase reflecting the mood in question rather than using its single word trnsliteral equivalents. For example, using the Hindi equivalents of such phrases as 'feeling interested', 'feeling distressed' etc. rather than using Hindi equivalents of the words 'interested', 'distressed' etc. Similarly, for certain mood adjectives (e.g., upset, alert etc.), the experts recommended that few word definition or synonyms should be used. Due attention was given to these and any other suggestions of the experts and necessary modifications/revisions were done in the light of their suggestions. The instruction was also translated in Hindi.
The revised item pool was then administered to a small group of participants (N=25) with the instructions (written in Hind) to mark their responses on each item. The purpose of this administration was to identify any gross difficulty in understanding of either the instructions or the item content by the participants. Review of the responses revealed that the scale is free from item-ambiguity and there was no evidence of difficulty in understanding either the instructions or the intended meaning of any item. This revised and the final Hindi version of the PANAS (here after referred as PANAS-H) was then administered on the final sample (N=179) in order to examine its psychometric properties.
| Results and Discussion|| |
To examine the factor structure of the PANAS-H the obtained data was subjected to an exploratory factor analysis (principal component analysis. Rather than using a hypothesis testing confirmatory factor analytic approach an exploratory factor analysis has been done in the present research in order to cross-validate the two factor solution of the PANAS reported by Watson and associates (1998).
The principal component analysis applied to 20 X 20 inter-item correlation matrix initially extracted 20 unrotated components. In order to identify the theoretically significant factors (components) we first used the criterion of eigenvalues greater than 1 which yielded 5 factors explaining 57.8 percent of the total variance. However, the examination of loading patterns suggested some redundant factors (factors having significant loadings on less than 3 items) and therefore the scree-test was used to identify the number of significant factors as per the suggestions of Gorsuch (1983). In the scree-test the eigenvalues of each factor is graphically plotted see [Figure 1] and those factors are retained for subsequent rotation that significantly depart from the scree. As it is evident from the scree plot that the initial two factors clearly departing from the scree. These two factors (components) explained 37.85 percent of the total variance.
The unrotated two factor component matrix was subsequently rotated using varimax (orthogonal) method with Kaiser normalization criterion. The rotated component matrix has been presented in [Table 1]. The cut-off loading for deciding the significance of items was set to 0.40 (see Gorsuch, 1983).
As evident from the [Table 1], all the mood adjectives reflecting positive demotional state have significant and positive loading on Factor-1 which reflects the presence of PA factor in Hindi affect lexicon. Thus, the first factor was labeled positive affect. Similarly, the 10 negative affect adjectives loaded significantly and positively on factor-2 and therefore this factor was labeled as negative affect (NA). None of the items loaded significantly (compared with the pre-established criteria of significant loading) on both factors except one item of factor-2 (nervous) which has a marginally significant loading of -.355 on factor-2. However, as this item loaded more strongly and positively on factor-2 and its item content was also congruent with the remaining items of this factor, this item was considered as member of factor-2.
Overall, the findings of factor analysis indicate the existence of two orthogonal factors of PA and NA in Hindi affect lexicon and suggest that these two factors can be validly measured using Hindi affect words. Further, it provides factorial (construct) validity of the PA and NA subscales of the PANASH.
The observation of two orthogonal factors, though, may be considered evidence for the independence of these factors, a significant negative correlation obtained between these two factors (r = -.176, p<.05) suggest that they are related with each other.
Taken together, the findings suggest that the PANAS-H measures two distinct but negatively related factors of PA and NA. This finding corroborates the findings of Crawford and Henry (2004) who also reported similar results using the confirmatory factor analytic approach. The findings are also in tune with earlier factor analytic research which reported a two factor solution with the PANAS (e.g., Crocker, 1997; Watson et al., 1988).
The finding of a low but significant negative correlation between PA and NA, though, do not support the widely held view of that these two affect components are independent of each other (see reviews by Diener et al. 1997, 1999, 2003), the observed low parentage of shared variance between these two factors (3.05 %) suggest that the two factors can be considered largely independent or less dependent of each other.
The internal consistency of the PANAS-H and its subscales was determined using the Chronbach's alpha method. The reliability of the positive affect scale was found to be .804 whereas that of negative affect subscale as .776. The reliability of the full scale was found to be .685 which is lower than the reliability of the either subscales. The observed low reliability of the full scale may be because of the independence or low correlation of the PA and NA subscales. To further ensure the reliability of items of the PANAS-H, the conventional item analysis was also done for the two subscales separately. The results [Table 2] indicated that none of the items were psychometrically poor. The corrected item-total correlation ranged from .379 to .571 for positive affect scale and from .279 to .643 for negative affect scale. None of the 'alpha-if item deleted' values exceeded the overall alpha which provides further evidence of the reliability of the items of the PANAS-H. The squared multiple correlation (SMC) is another index of item reliability (item homogeneity and validity) which indicates the amount of variance explained in a given item by the remaining items. The SMC ranged from .214 to .428 for items of PA subscale and from .125 to .482 for items of NA subscale. The SMC values are also highly satisfactory and provide additional evidence for the reliability of the items of PA and NA subscales of the PANAS-H.
Overall, the findings of reliability and item analyses suggest that 1) the PA and the NA subscales of the PANAS-H are composed of internally consistent and homogeneous items measuring the respective affects, and 2) the PANAS-H is satisfactorily reliable to measure the dimensions of PA and NA. The findings of the factor analysis provide support for the construct validity of the PANAS-H and suggest the existence of two orthogonal factors of PA and NA. However, the observation of a low but significant negative correlation between PA and NA do not support the notion of independence of these two affective dimensions. Thus, on the basis of the present findings it can be concluded that the PANASH reliably and validly measures two distinct but negatively related dimensions of PA and NA.
The findings of the present study, though, suggest that the PANAS-H is a psychometrically sound measure of the affective dimensions of PA and NA, further empirical evidence for the reliability and validity of the scale should be collected. For example, other aspects of the reliability such as temporal consistency should be examined in future research as the affect and mood are likely to vary over time. Moreover, the effect of different time frame (used for evaluating one's affective state) on the reliability of the PANAS-H should also be investigated in future research. The present study also do not report the other aspects of validity of the PANAS-H (such as convergent- discriminant validity, criterion related validity), and therefore these psychometric aspects of the PANAS-H should also be addressed by researchers. However, despite these limitations of the present study, the findings do provide some evidence for the psychometric adequacy and satisfactory reliability (internal consistency) and validity (factorial validity) of the PANAS-H.
| References|| |
|1.||Clark, L. A., & Watson, D. (1991a). Theoretical and empirical issues in differentiating depression from anxiety. In J. Becker & A. Kleinman (Eds.), Psychosocial aspects of depression (pp. 39-65). Hillsdale, NJ: Erlbaum. |
|2.|| Clark, L. A., & Watson, D. (1991b). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316-336. |
|3.|| Crawford, J.R., & Henry, J.D. (2004). The Positive and Negative Affect Schedule (PANAS) : Construct validity, measurement properties and normative data in a large non-clinical sample. British Journal of Clinical Psychology, 24, 245-265. |
|4.|| Crocker, P. R. E. (1997). A confirmatory factor analysis of the Positive and Negative Affect Schedule (PANAS) with a youth sport sample. Journal of Sport and Exercise Psychology, 19, 91-97. |
|5.|| Diener, E., Scollon, C. N.& Lucas, R. E. (2003). The evolving concept of subjective well-being : the multifaceted nature of happiness. Advances in Cell Aging and Gerontology, 15 , 187-219. |
|6.|| Diener, E., Suh, E., Lucas, R. E., & Smith, H. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276-302. |
|7.|| Diener, E., Suh, E., & Oishi, S. (1997). Recent findings on Subjective well-being. Indian Journal of Clinical Psychology, 24, 25-41. |
|8.|| Dyck, M. J., Jolly, J. B., & Kramer, T. (1994). An evaluation of positive affectivity, negative affectivity, and hyperarousal as markers for assessing between syndrome relationships. Personality and Individual Differences, 17, 637-646. |
|9.|| Gorsuch, R.L. (1983). Factor Analysis. Lawrence Erlbaum Associates: Hillslade, NJ. |
|10.|| Holt, K., & Ricciardell, L.A. (2002). Social comparisons and negative affect as indicators of problem eating and muscle preoccupation among children. Journal of Applied Developmental Psychology, 23, 285-304. |
|11.|| Jolly, J. B., Dyck, M. J., Kramer, T. A., & Wherry, J. N. (1994). Integration of positive and negative affectivity and cognitive content specificity: Improved discrimination of anxious and depressive symptoms. Journal of Abnormal Psychology, 103, 544-552. |
|12.|| Laurent, J., Catanzaro, S.J., Joiner, T.E., Rudolph, K.D., Potter, K.I., Lambert, S., Osborne, L. and Gathright, T. (1999). A measure of positive and negative affect for children : scale development and preliminary validation. Psychological Assessment, 11, 326-338. |
|13.|| Mackinnon A.1, Jorm A.F., Christensen H., Korten A.E., Jacomb P.A., & Rodgers B (1999). A short form of the Positive and Negative Affect Schedule: evaluation of factorial validity and invariance across demographic variables in a community sample. Personality and Individual Differences, 27, 405-416. |
|14.|| Markus, H., & Kitayama, S. (1991). Culture and self: Implications for cognition, emotion and motivation. Psychological Review, 98, 224-253. |
|15.|| Markus, H.R., & Kitayama, S. (1994). The cultural construction of self and emotion : Implications for social behaviour. In S.Kitayama & H.R. Markus (Eds.), Emotion and culture: Empirical studies of mutual influence, pp.89-130. Washington, Dc: American Psychological Association. |
|16.|| Mehrabian, A. (1998). Comparison of the PAD and PANAS as models for describing emotions and for differentiating anxiety from depression. Journal of Psychopathology and Behavioural Assessment, 19, 331-357. |
|17.|| Nunally, J.C. (1967). Psychometric theory. McGraw Hill: New York |
|18.|| Ostir, G.V. Smith, P.M., Smith, D., & Ottenbacher, K.J. (2005). Reliability of the Positive and Negative Affect Schedule (PANAS) in medical rehabilitation. Clinical Rehabilitation, 19, 767- 769. |
|19.|| Roesch, S. C. (1998). The factorial validity of trait positive affect scores: Confirmatory factor analyses of unidimensional and multidimensional models. Educational and Psychological Measurement, 58, 451-466. |
|20.|| Suh, E. M. (2000). Self, the hyphen between culture and subjective well-being. In E. Diener & E. M. Suh (Eds.), Culture and subjective well-being , ( pp . 63-86). Cambridge, MA: MIT Press. |
|21.|| Watson, D., & Clark, L. A. (1984). Negative affectivity : The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465-490. |
|22.|| Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS Scales. Journal of Personality and Social Psychology, 47, 1063-1070. |
|23.|| Yamasaki, K., Katsuma, R., & Sakai, A. (2006). Development of a Japanese version of the positive and negative affect schedule for children. Psychological Reports, 99, 535-546. |
[Table 1], [Table 2]