Cultural Differences in How People Deal with Ridicule and Laughter: Differential Item Functioning between the Taiwanese Chinese and Canadian English Versions of the PhoPhiKat-45

The PhoPhiKat-45 measures three dispositions toward ridicule and laughter, including gelotophobia (i.e., the fear of being laughed at), gelotophilia (i.e., the joy of being laughed at), and katagelasticism (i.e., the joy of laughing at others). Despite numerous cultural adaptations, there is a paucity of cross-cultural studies investigating measurement invariance of this measure. Undergraduate students from a Canadian university (N = 1467; 71.4% females) and 14 universities in Taiwan (N = 1274; 64.6% females) completed the English and Chinese PhoPhiKat-45 measures, respectively. Item response theory and differential item functioning analyses demonstrated that most items were well-distributed across the latent continuum. Five of 45 items were flagged for DIF, but all values had negligible effect sizes (McFadden’s pseudo R2 < 0.13). The Canadian sample was further subdivided into subsamples who identified as European White born in Canada (n = 567) and Chinese born in China, Hong Kong, or Taiwan (n = 180). In the subgroup analyses, no evidence of DIF was found. Findings support the utility of this measure across these languages and samples.


Introduction
Humour is a multidimensional phenomenon that encompasses a function (e.g., pro-social or mean-spirited) and fulfills complex needs for the individual (e.g., engage with others, mock others) [1]. Ruch [1] proposed that humour measures should have both humour and humourlessness represented, such that some individuals enjoy humorous interactions while others do not and thus avoid it. In response, Ruch and Proyer [2] proposed the measurement of three dispositions toward ridicule and laughter, including gelotophobia, gelotophilia, and katagelasticism. Gelotophobia is defined as the fear of being laughed at, a construct theoretically distinct from social anxiety, in which good-natured laughter may be perceived as a threatening form of ridicule directed towards the target [3,4]. Gelotophobia is highly correlated with fear of social situations, feeling perturbed, and greater stress in the workplace [5]. In contrast to the clinical categorization, a gelotophobic response is distinguished by an increased sensitivity to laughter [5]. Gelotophilia is the joy of being laughed at, in which joint laughter is perceived as a sign of appreciation [2]. Individuals who are high in gelotophilia do not mind being laughed at, as they see the role such as "class clown" desirable. Katagelasticism is the joy of laughing at others, in which the individual high in this trait may feel entitled to laugh at others at their expense if the opportunity presents itself [2].
The PhoPhiKat across cultures. Though prior adaptations of the PhoPhiKat questionnaire are solid contributions to the scientific literature, none of these studies have investigated the cross-cultural validity of this instrument. One might wonder, for instance, whether the information derived from the PhoPhiKat scores might be equally generalized to Eastern and Western countries. The psychometric properties of a revised PhoPhiKat questionnaire for Chinese individuals have been supported across several studies [8,13], where it demonstrated strong construct and structural validity, internal consistency, and test-retest reliability. Similarly, Chen et al. [8] showed promising results for the reliability and validity of the traditional Chinese version, titled PhoPhiKat-TC. Translated versions of the PhoPhiKat were translated from English, to a different language, and then back to English for further analysis [8]. These measures were developed to enable: (1) global psychological investigation and experimental design with psychometrically sound tools; and (2) crosscultural comparisons between individualistic and collectivistic cultures in order to further study cultural differences in humour and humour-related traits [8].
In the United States, the English version has been utilized in the cross-cultural study of gelotophobia in particular [14][15][16]. With focus on European and Asian Americans, Asian Americans scored higher compared to other cultures on the gelotophobia factor, which was interpreted as a result of the societal norm of collectivism among Asian cultures [8,15,17]. It is critical that measurement equivalence is established across cultures prior to concluding that the overall scale and its individual items have the same meaning across cross-cultural samples, however [18][19][20]. Measurement invariance analyses provide evidence that differences in test scores reflect true latent variable differences than group differences based on measurement bias [21,22]. The difference among cultural norms within a similar geographical location provide rationale for further cross-cultural research on the potential psychometric and observed means-differences in the PhoPhiKat-45.
How culture shapes humour and laughter. The presence of humour is commonplace, however, it may be perceived in different ways depending on one's cultural norms [23]. As outlined in Jiang, Li, and Hou [23], being humorous is viewed as a positive character trait and often used in everyday life in Western cultures. In Eastern cultures, it is largely not viewed as positively, despite occasional utilization to promote belongingness in attempt to raise social status [23,24]. Numerous studies have demonstrated how humour characteristics (i.e., appreciation, production, correlates) can vary among cultures such as Hong Kong and Taiwan [25,26], pushing forward its importance for further study.
Established in Taiwanese and Chinese samples, previous studies found that those among a collectivistic culture experience humour differently, with Taiwanese and Chinese individuals reporting to be more afraid of being laughed at due to the threat of group harmony [27,28]. Comparatively, as demonstrated in European and North American samples, those in individualistic societies may express less fear of being laughed at as it would not pose a threat to a group environment [28]. As Chinese culture stresses conformity and withholding of particular emotional reactions, aligned with the collectivist view, this information is not unprecedented, similarly to the results for the individualistic [28,29].
Amidst individuals with gelotophobia, there is greater avoidance of social situations with laughter-related emotions, comparable to previous findings revealing evidence associating gelotophobia with low extraversion [28,30,31]. Gelotophobia was linked to a lower life of engagement in a cross-cultural comparison, in addition to a lower life of meaning and pleasure among those in China, which lastly aids the explanation of why some aspects of humour-related emotions and traits are often avoided [11]. Concentrating on gelotophilia, those who are high in this dimension may 'expose' themselves to get laughed at, as demonstrated by English North Americans within the same study [28,31]. The comparison between Chinese individuals and English North Americans is important to understand how cultures can vary within as well as across geographical regions [32].
In summary, there are differences within and across cultures in how individuals perceive aspects of humour and humour-related traits. This distinction appears greatest between individualistic and collectivistic cultures, and may determine how a particular cultural group may define humour and its associations with key outcomes [32]. Further, there is stronger background on cross-cultural research relating to gelotophobia, with less information on gelotophilia and katagelasticism. The current investigation undertakes to further extend this literature, and to respond to this important evidence gap, in an evaluation of differences among these factors and their associations with perceptions and cultural beliefs across cultural groups.
The present research. Presently, there are a lack of cross-cultural studies on the PhoPhiKat-45 measure. The current study will respond by using item response theory (IRT) and differential item functioning (DIF) to be the first to investigate the psychometric properties of the measure and its relation to different samples. The present study aims to investigate: (1) the item response parameterization of the English and Chinese version of the PhoPhiKat-45; (2) evaluate uniform and non-uniform DIF for individual items with these two groups; and (3) further subdivide Canadian individuals to those who identified as European White and Chinese to evaluate whether DIF represents differing cultural perceptions within the same geographical region.
The first objective employs IRT, a technique to determine how much information a particular item on a measure or questionnaire may provide to the researcher [29,[33][34][35][36][37]. In using an item characteristic curve to determine the relationship between a participant and a given item, IRT can define the precision of items in both the English and Chinese measures. DIF is often utilized to interpret differences between groups within a measure in a meaningful way, such as determining group differences [34,38,39]. When considering previous works, it is hypothesized that the Chinese sample may have a higher likelihood of selecting an answer high in gelotophobia in comparison to the English sample based on measurement biases rather than cultural differences [15,17]. Thus, questions can be asked about biases due to cultural differences, similar to Lampert, Isaacson, and Lyttle's [15] study, or if another inference can be made. DIF can threaten test validity and highlight the need to thoroughly investigate each measure and account for any dangers to its strength [37,38,[40][41][42][43][44]. By employing IRT and DIF in the present study, the aim is to provide evidence for the psychometric properties and cross-cultural measurement invariance of the PhoPhiKat-45 in English and Chinese.
In the Taiwanese Chinese sample, data were collected using paper and pencil from undergraduate students from 14 universities located in Northern, Central, Southern, and Eastern areas in Taiwan (N = 1274; 64.6% females) between the ages of 18 to 29 years (M age = 19.74; SD = 1.48). Students participated voluntarily and were debriefed after participation.

Materials and Procedure
PhoPhiKat-45. The PhoPhiKat-45 is a reliable and valid measure with 45 items that assesses gelotophobia (i.e., the fear of being laughed at), gelotophilia (i.e., the joy of being laughed at), and katagelasticism (i.e., the joy of laughing at others) [4,45,46]. Each respondent rates the items on a four-point Likert type scale ranging from strongly disagree to strongly agree. The measure has demonstrated evidence of strong internal consistency, test-retest reliability, and structural, convergent, and discriminant validity [4,5,8,45,46].

Analytic Strategy
The marginal maximum likelihood estimation method was used to evaluate gelotophobia, gelotophilia, and katagelasticism through Samejima's graded response model (GRM) [47]. Orlando and Thissen's [48] S-X 2 statistics were reported for item fit [49]. Local dependence (LD) is evaluated using Yen's Q 3 statistics for excess item covariation when the latent trait is controlled for [50]. Marais and Andrich [51] recommended critical residual correlation values between 0.10 to 0.30. For this study, a value of 0.30 was used to test the violation of the local independence assumption. Yen's Q 3 values showed no evidence for the violation of the assumption of local item independence for all items in the Canadian English Sample. In the Taiwanese Chinese sample, most items showed no evidence for the violation of the assumption of local item independence; however, items 1 and 7 showed a residual correlation of 0.46. Specifically, these two items shared overlap in content regarding being paranoid about being the person laughed at when overhearing strangers laughing. As such, method variance may lead to residual correlations that the latent trait cannot explain [52].
In Samejima's graded response model, a single item discrimination value (a) and three category threshold (b i ) function values were produced for every item. Discrimination values were categorized as follows: ≤0.24 as very low, 0.25 to 0.64 as low, 0.65 to 1.34 as moderate, 1.35 to 1.69 as high, and ≥1.7 as very high [53]. The threshold parameters (b i ) were presented as z-scores (M = 0, SD = 1) and indicate the amount of latent trait required for a 50% probability of endorsing the following response category.
DIF was used to determine whether biases across the Canadian English and Taiwanese Chinese versions existed. Crane et al. [54] recommended an ordinal regression approach to DIF that follows Samejima's [47] graded response model (GRM) [55,56]. Based on the latent trait, each item incorporates three separate ordinal regression models. Model 1 (scores) includes the model at baseline with response probabilities regressed on the latent trait. Model 2 (scores and group) is a uniform DIF model with the main effect specified by group. Model 3 is a non-uniform DIF model that includes an interaction term between the participants' latent trait and group-specific covariate [57]. Uniform DIF is identified with the comparisons of models 1 and 2 and non-uniform DIF is established with comparisons between models 2 and 3 [57]. Total DIF effects are established with comparisons between models 1 and 3 [54,58,59].

DIF Subgroup Analyses
To ascertain whether DIF biases existed when Chinese participants completed the English version of the PhoPhiKat-45, subgroup analyses were conducted within the English Canadian sample. This group was further divided into a group who identified as European White born in Canada (N = 567; 77.4% females; Mage = 18.40; SD = 1.98) and a group who identified as Chinese born in China, Taiwan, or Hong Kong (N = 180; 72.8% females; Mage = 18.73, SD = 1.17). The same criterion threshold of McFadden's R 2 > 0.035 and 10% change in proportional β was used to examine DIF. No significant DIF was identified across the three subscales of gelotophobia, gelotophilia, and katagelasticism.

DIF and Effect Sizes
The results from the ordinal logistic regression for DIF detections when comparing the Canadian English and Taiwanese Chinese versions are displayed in Table 1. Among the 14 gelotophobia items, 2 (i.e., items 37 and 43) were flagged as DIF items with the

Item Response Theory
Unidimensionality for each of the gelotophobia, gelotophilia, and katagelasticism dimensions was established, such that the first factor should account for >20% of the variance met by both samples [39]. However, the Canadian English sample shows a negative slope for item 13 (i.e., I believe that I make involuntarily a funny impression on others), suggesting misinterpretation for the translation of this item from the original German measure. Thus, it was removed from subsequent analyses [45]. For the Canadian sample, discrimination parameters for the three subscales are as follows: gelotophobia Mdn(a) = 1.

Test Information Function (TIF)
The evaluation of TIFs for the Canadian English ( Figure 1) and Taiwanese Chinese ( Figure 2) samples showed similar measurement precision across the latent trait (i.e., θ). In the Canadian sample, the information and associated standard errors of measurement (SEs) showed precise measurement for respondents falling approximately within −1.5 to +3.0, −2.0 to +1.5, and −1.0 to +3.0 of the gelotophobia, gelotophilia, and katagelasticism dimensions, respectively, in the latent trait continuum. This is indicated by the maximum I and minimum SE (Figures 1 and 2 for the graphical representation of the TIF). In the Taiwanese Chinese sample, the TIF for gelotophobia, gelotophilia, and katagelasticism showed precise measurement of θ from roughly −1.5 to +3.0, −3.0 to +1.5 and −1.5 to +3.0, respectively, as evidenced by the maximum I and minimum SE.

DIF Subgroup Analyses
To ascertain whether DIF biases existed when Chinese participants completed the English version of the PhoPhiKat-45, subgroup analyses were conducted within the English Canadian sample. This group was further divided into a group who identified as European White born in Canada (N = 567; 77.4% females; M age = 18.40; SD = 1.98) and a group who identified as Chinese born in China, Taiwan, or Hong Kong (N = 180; 72.8% females; M age = 18.73, SD = 1.17). The same criterion threshold of McFadden's R 2 > 0.035 and 10% change in proportional β was used to examine DIF. No significant DIF was identified across the three subscales of gelotophobia, gelotophilia, and katagelasticism.

DIF and Effect Sizes
The results from the ordinal logistic regression for DIF detections when comparing the Canadian English and Taiwanese Chinese versions are displayed in Table 1

Discussion
The purpose of this study was to provide evidence for the psychometric properties of the PhoPhiKat-45 in English and Chinese in two independent samples collected from  Note. χ 12 , χ 23 , and χ 13 represents the χ 2 likelihood ratio statistic between Models 1 and 2, Models 2 and 3, and Models 1 and 3, respectively. R 12 , R 23 , and R 13 represent McFadden's pseudo R 2 from a comparison Models 1 and 2, Models 2 and 3, and Models 1 and 3, respectively; ∆β represents the change in β as a percentage

Discussion
The purpose of this study was to provide evidence for the psychometric properties of the PhoPhiKat-45 in English and Chinese in two independent samples collected from Canada and Taiwan. Using IRT parameters for the Canadian English and Taiwanese Chinese versions, items were well-distributed across the latent continuum and showed high discrimination parameters allowing differentiation across different levels of gelotophobia, gelotophilia, and katagelasticism. Item characteristic curves for each individual item demonstrated high discrimination parameters that were well spread across the latent continuum for gelotophobia, gelotophilia, and katagelasticism, respectively measured. Notably, one item (i.e., item 13 "I believe that I make involuntarily a funny impression on others") was excluded from the gelotophobia latent trait analysis due to a negative slope identified that precluded further analyses. The English PhoPhiKat-45 measure was translated from the German version and this item may be read by English speakers as making a funny impression or appearing hilarious or amusing in front of others naturally. Although the Taiwanese Chinese version was translated from the English version, the item performed well in the Taiwanese sample. Future studies should investigate whether this item may be removed or modified for the English measure of the PhoPhiKat-45.
The results showed that single test reliability values were acceptable for all three dimensions of gelotophobia, gelotophilia, and katagelasticism. The measurement precision of the test was evaluated using the TIF [67][68][69]. Interestingly, both the English and Chinese versions of the measure show that the scale does not capture extremely low ends of gelotophobia and katagelasticism and the extremely high end of gelotophilia. It is worth noting that both gelotophobia and katagelasticism are quasi-traits, such that both traits are unipolar constructs intended to measure only the presence or absence of the trait [2,70]. As such, the low end of the gelotophobia spectrum suggests the absence of fear of being laughed at rather than the presence of gelotophilia. These findings corroborate Ruch and Proyer's [2] initial conceptualization of gelotophobia, that it is not a bipolar trait which would entail both extreme ends of the spectrum as separate entities (i.e., gelotophobia and gelotophilia). These results suggest that the absence of gelotophobia (i.e., no fear of being laughed at) does not indicate presence of gelotophilia (i.e., the joy of being laughed at). As such, both measures are needed for accurate measurement of these separate constructs.
Measurement invariance using DIF was also conducted to investigate whether meaningful interpretation of mean group comparisons may be made without significant biases. At the single-item level, five items were flagged for DIF in the Taiwanese Chinese version against the original English version. Only one item (i.e., item 21 "some people set themselves up for one to make fun at them") analyzed has 10% change in β and over 0.13 in McFadden's pseudo R 2 . Following an approach by Lau, Chiesi, and Saklofske [32] and Li et al. [71], the Canadian sample who completed the measure was subdivided into those who identified as European White born in Canada and those who identified as Chinese born in China, Hong Kong, or Taiwan. In the subgroup analyses, no evidence for DIF was found, suggesting that the DIF identified in the Taiwanese Chinese and English versions are more likely linked to subtle shifts in meaning with the translation process rather than the cultural concepts covered by the item content. Cultural expressions of affect, emotions, and behaviours vary cross cultural groups, which may lead to cross-cultural comparison biases [72]. Overall, there is support for using this scale to assess gelotophobia, gelotophilia, and katagelasticism with Chinese university students in either Chinese or English languages. DIF findings were mostly negligible and do not appear to reflect attainable differences in research or clinical practice.
Several limitations from the present investigation should be addressed in future studies. First, both the Canadian and Taiwanese samples of undergraduate university students represent young and well-educated individuals. Future research should replicate these findings to investigate whether findings are generalizable across age groups (e.g., youth, older adults), clinical status (e.g., those diagnosed with physical illnesses or mental health disorders), and socioeconomic background. Second, although the present study included an extensive assessment of individual item performance of the PhoPhiKat-45, future studies will benefit from investigating whether these variables have the same predictive validity across cultures. Specifically, Chinese individuals conceptualize the value of humour in everyday life differently than those from Western cultures given the early historical and cultural influences of Confucianism, Taoism, and Buddhism (for a review see Yue [17]). With Chinese individuals respecting authority and seniority more than Canadians, Chinese individuals tend to be more careful and conservative when using humour [17]. Whereas a serious disposition may be associated with lower well-being in Western cultures, seriousness is positively associated with well-being in China [73][74][75][76]. For example, Singaporean students in Nevo, Nevo, and Yin's [77] study represented more conventional traits in relation to jokes and humour, such as aggressiveness, contrast to students in the United States, who joked more about sexual items. In addition, they did not use humour as a way to cope unlike other countries studied, such as Israel, demonstrating that a culture's values can impact an individual's humour [77]. With this in mind, future studies should examine whether some levels of gelotophobia may be protective for the individual in a culture where humour may be avoided or seen as distasteful [17,78]. Lastly, the present study reported the proportion of DIF items that may affect group means, but there are other methods such as the reporting of linking errors (i.e., robust linking methods [79]) and DIF variance. Future studies should explore these avenues to enable the assessment of systematic biases between countries.

Conclusions
Overall, IRT and DIF analyses conducted on the English and Taiwanese Chinese versions of the PhoPhiKat-45 demonstrated strong psychometric properties for assessing gelotophobia, gelotophilia, and katagelasticism. With the exception of one item, crosscultural comparisons of the measure did not present biases with significant effect sizes for the individual item characteristics. Through a subgroup analysis, it is suggested that translations may alter item meanings rather than cultural concepts. However, variance across cross cultural groups may result in comparison biases, such as differences in the conceptualization of humour, including cultural influences, trait representation, and personal coping methods. Findings from this investigation support the utility of the PhoPhiKat-45 English and Chinese versions, and support the generalizability of gelotophobia, gelotophilia, and katagelasticism measures across different languages.

Informed Consent Statement:
This study was conducted according to the American Psychological Association (APA) and National Association of Psychology ethical standards for the treatment of human subjects. Data collection was anonymous and involved no identifying information and no medical treatment. Participants were informed that their participation was voluntary, that they could leave the study at any time, and that their data would be treated anonymously. In addition, they were informed that by starting the survey they would be considered to have read and accepted the informed consent. Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available.
Acknowledgments: The first author would like to thank Canadian Institutes of Health Research, Mitacs, Centre for Addiction and Mental Health Discovery and Womenmind program, and Mental Health Research Canada for supporting the first author's postdoctoral research program. The scholarship awards did not participate in study conceptualization, data collection, data analysis, or manuscript preparation. The authors would like to thank Meaghan Elizabeth Mosier-Farquharson for her help in organizing the appendices.

Conflicts of Interest:
The authors declare no conflict of interest.