On kitsch and kic: Comparing kitsch concepts from Bavaria, Serbia and Slovenia

The German word kitsch has been internationally successful. Today, it is commonly used in many modern languages including Serbian and Slovenian (kic)-but does it mean the same? In a pilot study, thirty-six volunteers from Bavaria, Serbia and Slovenia rated two hundred images of kitsch objects in terms of liking, familiarity, determinacy, arousal, perceived threat, and kitschiness. Additionally, art expertise, ambiguity tolerance, and value orientations were assessed. Multilevel regression analysis with crossed random effects was used to explore crosscultural differences: Regardless of cultural background, liking of kitsch objects was positively linked to emotionally arousing items with non-threatening content. Self-transcendence was positively linked to liking, while ambiguity of the parental image was concordantly associated with kitschiness. For participants from Serbia and Slovenia, threatening content was correlated with kitschiness, while participants from Bavaria rated determinate items as kitschier. Results are discussed with regard to literature on kitsch and implications for future research.


Highlights:
• Kitsch objects were rated as non-threatening, determinate but hardly arousing.
• Particularly non-threatening and mildly arousing kitsch objects were preferred.
• Liking of kitsch objects was positively associated with self-transcendence.
• Female participants tended to prefer familiar and non-threatening kitsch objects.• In the Serbian and Slovenian sample, kitsch was linked to threatening content.

Introduction
Out of the manifold expressions for bad taste from various modern languages, only the German word kitsch has been internationally successful (Menninghaus, 2009): As an antithesis to high art and a synonym for "tasteless mass produced trash" (Pazaurek, 1912(Pazaurek, /2012, p. 349, p. 349, translation by first author) kitsch counts among the most controversial aesthetic concepts of 20th century art theory (Simon-Schäfer, 1980).The term itself can be traced back to the late 1800s when it started circulating among artists and art dealers from Munich as a derogatory label for "cheap artistic stuff" (Cǎlinescu, 1987, p. 234). 1 Meanwhile, kitsch has also entered many modern languages including Serbian and Slovenian (kič).This raises two questions: Firstly, is the colloquial use of the term kitsch consistent with basic assumptions from art theory?Secondly, when people from different cultures use the word kitsch colloquially, do they refer to the same aesthetic idea or is kitsch just an umbrella term for rather culture-specific notions of bad taste?In a first pilot study, we sought preliminary answers to these questions based on a comparison of kitsch concepts from Bavaria, Serbia and Slovenia.For this study, a set of comparative dimensions was identified from the extensive literature on kitsch and art.Practically all authors who have expressed their thoughts about kitsch are agreed that the term serves as a derogatory label of bad taste, which may be used either for nominal categorisation (e.g.A is kitsch, but not B) or continuous evaluation (e.g.A is kitschier than B) of anything subject to aesthetic judgement.Recent accounts of kitsch dwell on the contradictory relationship between liking and kitsch (Kulka, 1996): Although kitsch is despised, it is proving commercially successful.Especially with Pop Art, tension between high art and kitsch has eased considerably.In the present study, liking and kitschiness were therefore considered as two separate, but not independent, variables.
What variables predict kitsch judgements?In Kitsch and Art, the philosopher Tomaš Kulka (1996) elaborates on three characteristics of kitsch: Above all, he claims that kitsch requires a familiar, heart-warming subject matter that elicits an immediate emotional response.Secondly, this subject should be depicted in a perfectly conventional way allowing for instant identifiability.Thirdly, context and manner of representation must not enrich the observer's "associations relating to the depicted objects or themes" (p.37).Kulka concludes that the more clearly and unambiguously an object complies with these "three conditions, the more paradigmatic an example of kitsch it is" (p. 38).This definition advocates a continuous assessment of kitschiness.Moreover, it links kitsch to several variables that figure prominently in empirical aesthetics: familiarity, determinacy, positive emotional valence, and arousal.
Familiarity seems highly relevant for kitsch since it relies on themes to which the beholder has been positively predisposed.This claim is shared by several authors who observe that nostalgia in particular engenders kitsch.Norman (2004), for instance, argues that keepsakes are mainly cherished for their ability to instil positive emotions and pleasant memories: As relics of the good old times they radiate what Titchener (1910) once described as "a glow of warmth, a sense of ownership, a feeling of intimacy, [and] a sense of being at home" (p.408).Hence, we expect that an object will not be regarded as kitsch, unless it appears highly familiar.For kitsch to elicit a spontaneous affective response, its emotion-laden subject must be immediately and unambiguously identifiable (Kulka, 1996).Research literature even suggests that lack of ambiguity is what distinguishes kitsch from art (Muth, Hesslinger, & Carbon, 2015;Ortlieb & Carbon, 2014).We therefore surmise that kitsch has to be highly determinate in order to be enjoyable.Apart from being familiar and determinate, kitsch must not disturb or question our basic sentiments and beliefs.Only if its subject matter is charged with positive emotions, it will be able to comfort us.Thus, we expect that kitsch has to be perceived as non-threatening in order to yield a spontaneous heart-warming response.Apart from Kulka several authors have asserted that the appeal of kitsch lies in its "instantaneous emotional gratification without intellectual effort" (Menninghaus, 2009, p. 41).Since the concept of arousal plays an important role in both empirical aesthetics (Berlyne, 1971) and emotional appraisal theories (James, 1884;Lange, 1887Lange, /2013;;Schachter & Singer, 1962), it appears particularly relevant for the study of kitsch.According to literature, we expect that cognition plays a completely different role in the appreciation of art and kitsch: Art is exciting because it acquaints us with something new or enigmatic (Berlyne, 1971).Thus, art perception mostly involves cognitive mastery (Leder, Belke, Oeberst, & Augustin, 2004).By contrast, kitsch offers immediate emotional gratification without reasoning (Cǎlinescu, 1987;Kulka, 1996;Menninghaus, 2009) in that it draws on common life experience, shared beliefs and representational conventions.Hence, the consoling properties of kitsch-familiarity, determinacy and positive valence-suggest a negative relationship between kitschiness and arousal.
So what is wrong with immediate emotional gratification?Why are people repulsed rather than attracted by kitsch?So far we have addressed objectrelated aspects of kitsch.In the following, we will focus on five variables that might account for variance between individuals in terms of liking: art expertise, ambiguity tolerance, value orientation, age, and gender.Kitsch is usually contrasted with applied art (Pazaurek, 1912/2012), avant-garde art (Greenberg, 1939), or art proper (Kulka, 1996;Simon-Schäfer, 1980).Thus, it seems likely that one's kitsch concept depends on one's previous knowledge about art.This assumption accords with theories from sociology of art relating taste judgements to culture capital (Bourdieu, 1984).Since kitsch is seen as an antithesis to high art, we speculate that a high level of expertise will be positively correlated with kitsch ratings and negatively associated with liking.Ambiguity tolerance is another personality trait that is relevant in this respect.It moderates a person's openness to the enigmatic, contradictory and even disturbing properties of contemporary artworks (Muth et al., 2015).Regarding kitsch, we expect that individuals with a low tolerance for ambiguity are particularly susceptible to its familiar and clear-cut qualities.Dissanayake (1990) writes that it is from our emotional responses to art that we learn about our personal values.So what does our affective response to kitsch tell us about our values?For Greenberg (1939) and Broch (1933Broch ( /2002))-two supporters of the avant-garde-kitsch stands for "Evil in the Value-System of Art" (p.137): hedonism and conservatism.Accordingly, hedonistic and conservative value orientations should be positively correlated with liking, but negatively with kitsch ratings.Even demographical aspects such as gender and age have been associated with kitsch: The writer and theatre critic H. W. Fischer (1919), for example, claimed that kitsch reflects female taste, while two authors relate kitsch to intergenerational conflict (Avenarius, 1920;Stemmle, 1931).Avenarius (1920) reports that young artists of the early 1900s made fun of their well-established predecessors and their oldfashioned ways by calling them "Kitschiers."From these early accounts of kitsch one might infer that liking ratings should be higher among female participants and positively associated with age, while kitsch ratings should be higher among male participants and negatively related to age.
Is the colloquial use of the word kitsch today still compatible with literature on kitsch?And when people from different cultures use the term, do they have the same aesthetic concept in mind?In the following we report on a cross-cultural pilot study that explored these questions by comparing kitsch concepts from Bavaria, Serbia and Slovenia.

Stimulus material
Being a very broad aesthetic concept, kitsch of course raises the question of appropriate stimulus material.Although kitsch may apply derogatorily to music, literature, filmmaking, architecture, fashion, furnishing, and interior decoration, it was originally limited to artistic production in the visual domain (Cǎlinescu, 1987).Hence, we decided to address kitsch in everyday objects: A set of 200 high-resolution digital images of kitsch objects such as keepsakes (e.g.miniature Eiffel tower), give-aways, toys, or collectibles (Fig. 1 D), as well as merely decorative (Fig. 1, A and C), and devotional objects (Fig. 1, A and E) was used.2Apart from these 200 images of kitsch objects, eight images of plain household objects (see Figure 1, F) were included as base stimuli.

Variables and measures
The 208 images were rated in terms of liking, familiarity, determinacy, arousal, perceived threat, and kitschiness.For each rating, a seven-point Likert-scale was used (see Table 1).The German, Serbian and Slovenian versions of these scales were created by the first, second and third author based on an English reference version. 3Based on these translations three otherwise identical digital questionnaires were created using the ExperimentBuilder© (Version 1.10.165)software.A self-devised questionnaire was applied to estimate the participants' level of expertise in the visual arts domain (see Table 2).The original German version was created by the senior author.The Serbian and Slovenian versions of this questionnaire were translated from an English version which was also provided by the senior author.Table 3 shows how scores for art expertise were calculated.Ambiguity tolerance was measured using the German Inventar zur Messung der Ambiguitätstoleranz (IMA) by Reis (1996).According to the IMA-manual, persons scoring high on scales for ambiguity tolerance are likely to "(a) seek out ambiguity, (b) enjoy ambiguity, and (c) excel in the performance of ambiguous tasks" (MacDonald, 1970, quoted from Reis, 1996, p. 7), whereas persons with low scores tend "to perceive or interpret information marked by vague, incomplete, fragmented, multiple, probable, unstructured, uncertain, inconsistent, contrary, contradictory, or unclear meanings as actual or potential sources of psychological discomfort or threat" (Norton, 1975, quoted from Reis, 1996, p. 7).The IMA comprises 40 items which discriminate between ambiguity tolerance regarding openness to experience (OE), role stereotypes (RS), parental image (EB), unsolvable problems (PR), and social conflict (SK).For each participant, six sum scores-five sub scores and a global IMA-score-were calculated in accordance with the IMA-manual: Twenty-four items featured a six-point Likertscale ranging from 1=very true to 6=not true at all.The original German version of the IMA has been validated and normative data is reported by Reis (1996).The Serbian and Slovenian versions of the IMA-scales were translated by the second and third author from an English translation by the first author.All versions of the IMA-scales are reported in Table 4.
(1) I like to know in advance what to expect when I go on holiday.
(2) When dancing, women should give men the lead.
(7) One should be able to identify gender by the kind of clothing a person is wearing.
(13) A woman should decide between having a career or having children.
(18) I appreciate that there is an increasing number of so-called "househusbands".
(5) My father and I have an ambivalent relationship.
(6) I will raise my children exactly the way my parents raised me.
(10) I have always had a good relationship with my father.
For the assessment of basic value orientation, the Short Schwartz's Value Survey-German (SSVS-G) by Boer (2014) was applied (see Table 5).The underlying model by Schwartz (1992) describes commonalities in content and structure of values based on empirical research in twenty countries.It postulates two universal value dimensions: Selfenhancement versus Self-transcendence and Conservation versus Openness-to-Change.The first dimension relates to "the conflict between concern for the welfare of other people (high Self-Transcendence) and concern for individual outcomes and personal interests (low Self-Transcendence)" (Lindeman & Verkasalo, 2005, p. 171), whereas the latter reflects "the conflict between the motivation to preserve the status quo and the certainty that conformity to norms provides (high Conservation), on one hand, and the motivation to follow one's own intellectual and emotional interests (low Conservation) on the other hand" (p.171).Participants rated the importance of ten abstract values-Power, Achievement, Hedonism, Stimulation, Self-direction, Universalism, Benevolence, Tradition, Conformity, and Security-on a six-point rating scale (1=not important at all; 2=not important; 3=not really important; 4=somewhat important; 5=important; 6=very important).For each participant four different scores were derived from these ratings: Self-enhancement (mean from scores on power and achievement), Self-transcendence (mean from scores on benevolence and universalism), Conservation (mean from scores on tradition, conformity and security), and Openness-to-Change (mean from scores on hedonism, stimulation and self-direction).In the case of the SSVS-G, the English version of the Short Schwartz's Value Survey (SSVS) by Lindeman and Verkasalo (2005) served as a basis for translations into Serbian and Slovenian.In both versions, however, rating scales were adjusted to the six-point scheme of the SSVS-G.All translations were checked for translation correctness based on back translations by three native speakers from each country.2005) features a nine-point Likert-scale ranging from 0=not important at all to 8=very important.In the present study, response options were adjusted to the six-point scale of the SSVS-G.

Procedure
The study was conducted at the Universities of Bamberg (Bavaria), Belgrade (Serbia) and Ljubljana (Slovenia).In every location, the study was conducted according to the same procedure (Figure 2): Initially, participants signed an informed consent form and completed the art expertise questionnaire, the IMA-questionnaire (Reis, 1996) and the SSVS-G (Boer, 2014).Subsequently, 208 images of everyday objects were presented and rated in terms of liking, familiarity, determinacy, arousal, perceived threat, and kitschiness.In order to minimize loss of vigilance, stimuli were rated in two sessions (k 1 = 92 and k 2 = 116).Within each session images were presented in a random sequence, while the order of ratings was identical for every stimulus.

Results
Initially, descriptive statistics for age, art expertise and personality scales were examined (see Table 6).Since some of the variables (age, IMA Problems, SSVS Self-transcendence, SSVS Self-enhancement, SSVS Openness-to-Change) had non-normal distributions, measured using the Shapiro-Wilk test, the nonparametric Kruskal-Wallis test was used for all variables to test differences among the three cultures.Significant differences between the three subsamples were obtained for age, H (2) = 18.21, p <.01 and IMA Parental image, H (2) = 7.34, p <.05.Mann-Whitney tests were used to follow up these findings.Bonferroni correction was applied so effects are reported at a .0167level of significance.It appeared that the Slovenian sample was older than the Serbian (U = 5, r = .83,p <.0167) and the Serbian sample had higher scores on IMA Parental image than the Bavarian (U = 31.5,r = .48,p <.0167).Note.Bolded values are significant differences between the three groups at the level p <.0167, df = 2 The medians for averaged values for the six rating dimensions of 200 kitsch images are shown in Table 7 (eight non-kitschy objects were excluded from analysis).In general, kitsch objects were disliked and described as non-threatening, non-arousing and highly determinate.Regarding perceived threat, 80% (!) of all ratings were equal to or less than 2 on a seven-point Likert-scale ranging from 1=not threatening at all to 7=very threatening.Similarly, 54% of arousal ratings (1=not exciting at all; 7=very exciting) and 47% of liking ratings (1=I don't like it at all; 7=I like it very much) were equal to or less than 2. Besides, 54% of determinacy ratings (1=not determinate at all; 7=very determinate) were equal to or higher than 6.Again, normality of distribution was assessed using Shapiro-Wilk test, and for some scales (kitschiness, liking, arousal and perceived threat) non-normal distributions were confirmed.Therefore, the Kruskal-Wallis test was used for all variables to test the cross-cultural differences in terms of these ratings.In three cases, ratings differed significantly between the three cultures: familiarity, H (2) = 14.42, p <.01, determinacy, H (2) = 15.61,p <.01, and kitschiness, H (2) = 7.67, p <.05.In the Serbian sample, images were rated as more familiar than in the Slovenian (U = 25, r = .55,p <.01) and the Bavarian (U = 7, r = .77,p <.01) samples, and more determinate than in the other two samples (Slovenian: U = 10, r = .73,p <.01, Bavarian: U = 27, r = .53,p <.0167).The Bavarian sample rated the images as kitschier than the Slovenian (U = 28.5, r = .51,p <.01).Note.Bolded values are significant differences between the three groups at the level p <.0167, df = 2 For further exploratory analyses, multilevel modelling of crossed random effects was used (Heck, Thomas, & Tabata, 2013;Hox, 2010).This method has two assets: Firstly, the advantage of multilevel analysis over traditional univariate or multivariate analysis on one level with mean ratings of individuals or mean ratings of images (Muth et al., 2015;Silvia, 2007).Secondly, crossed effects modelling has an advantage over hierarchical multilevel models in that it avoids the problem of whether to analyse data on the level of ratings done by individuals within images or images within individuals (Fischer, Carbon, Rutar, Stojilović, & Ortlieb, 2016).Models of crossed random effects are able to take multiple sources of variation into consideration-variations between participants, between images and their interaction, and individual replies (level 1) are nested within participants and within images (level 2).
The restricted maximum likelihood method was used to evaluate the model.The significance of fixed effects was assessed using p values from the Wald test, and the significance of random effects was estimated using -2ΔLL likelihood ratio tests and informative criteria (AIC and BIC) between two models including the same fixed effects (Hoffman, 2015).The values for denominator degrees of freedom were obtained by a Satterthwaite approximation.
Liking and kitschiness were defined as dependent variables.Ratings of familiarity, determinacy, arousal, and perceived threat, which varied with every new judgement, served as predictors on the level of kitsch stimuli, while constant values for art expertise, ambiguity tolerance, value orientation, gender, and age were included as predictors on the level of participants.All continuous predictors were centred on the grand mean.

Best predictors for liking
Initially, empty models (i.e.models without predictors) were modelled to split the total variance of liking.At both levels the diagonal matrix of random effects was defined providing an estimate of the variance for each random effect, and covariance between the two random effects was restricted to be zero.As a result, 17.7% of the total variation of liking is due to differences between participants, while 13.4% reflects variation between images.The remaining 68.9% of total variation is residual variance which cannot be explained by variation either at the level of participants or images (Table 8, Model I).
For the next model, both individual ratings of familiarity, determinacy, arousal, and perceived threat and their interactions were used.Three level-1 covariates were significantly related to liking judgements: arousal, familiarity and perceived threat.In Model II (Table 8) predictors on the level of participants are included.Predictors and interactions which were not significant were left out.Only for the SSVS-scale Self-transcendence a significant interrelation with liking was obtained: Higher scores on Self-transcendence were positively correlated with liking of kitsch objects.Relative to the empty model, predictors of Model II reduced the residual variance by pseudo-R 2 res = .31,image random variance by pseudo-R 2 img = .38and subject random variance by pseudo-R 2 sub = .49.Model II reduced total variance from 3.36 (empty model) to 2.19, hence explaining 35% of total variance.In the last model, we examined the extent to which the effects of the image predictors show systematic individual differences by adding subject random slopes for the image predictors to the model (Table 8, Model III, Figure 3).We retained subject random effects for the intercept as well as for arousal, familiarity and perceived threat slopes.Given significant subject variation in the effect of three image predictors, the next step was to test that variation with subject predictors.The strongest interrelation with liking was obtained for arousal, b = 0.60, t (35.84) = 16.89,p <.01, d = 2.82.The importance of familiarity and perceived threat were moderated by gender: For female participants familiarity and liking were positively associated, b = 0.13, t (33.40) = 2.90, p <.01, d = 0.50, while the negative correlation between perceived threat and liking was less pronounced for male participants, b = -0.16,t (21.81) = 2.74, p <.05, d = 0.59.Model III explained 39% of total variance.

Best predictors for kitschiness
Again empty models were calculated to split total variance of kitschiness.At both levels the diagonal matrix was defined.24.2% of total variation regarding kitschiness reflected differences between participants, while another 25.0% was due to variations between images leaving 50.8% of unexplained residual variance (Table 9, Model I).
Ratings of familiarity, determinacy, arousal, and perceived threat as well as their interactions were added to the next model.Arousal and determinacy had a significant direct effect.Perceived threat, however, revealed both a significant direct effect and interaction effects with familiarity and determinacy.For Model II predictors on the level of participants were added (Table 9).Since non-significant predictors and interactions were left out, only two predictors on the level of participants were retained: cultural background and IMA-parents.In Model III we examined the extent to which the effects of the image predictors show systematic individual differences (Table 9, Figure 4).We retained the subject random effects for the intercept and all four image predictor slopes-arousal, familiarity, determinacy, and perceived threat.Arousal was negatively linked to kitschiness, b = -0.29,t (35.89) = 8.36, p <.01, d = 1.39, while interrelations with perceived threat and determinacy differed with regard to cultural background, F (2, 28.6) = 5.93, p <.01; F (2, 26.64) = 3.68, p <.05: In the Bavarian sample, perceived threat and kitschiness were unrelated, whereas ratings of perceived threat were positively correlated with kitsch ratings in the Serbian and the Slovenian sample.A positive link between determinacy and kitschiness was only observed in the Bavarian sample.Model III accounted for 18% of total variance.

Liking and kitschiness
Finally, the relationship between the two dependent variables-liking and kitschiness-was explored using multilevel modelling.With liking as a predictor centred on the grand mean, a moderate negative relationship between the two variables was obtained, b = -0.41,t (7140) = 41.52,p <.01, d = 0.50.This interrelation was not affected by cultural background.

Discussion
The German word kitsch has been adopted by many modern languages, including Serbian and Slovenian, yet it remains unclear whether its colloquial use is based on the same notion of bad taste.The aim of the present cross-cultural pilot study was to fathom differences and commonalities between concepts of kitsch from Bavaria, Serbia and Slovenia.Thirty-six persons rated 200 images of kitsch objects in terms of liking, familiarity, determinacy, arousal, perceived threat, and kitschiness.Additionally, art expertise, ambiguity tolerance and value orientation were assessed and a multilevel regression analysis with crossed random effects was conducted to identify predictors for liking and kitschiness.In the following, we discuss results with regard to literature on kitsch and previous findings.Finally, the limitations of the present pilot study and the implications for further research are pointed out.
Descriptive analysis showed a consistent pattern: Most kitsch objects were disliked and a majority of them were characterized as perfectly harmless, highly determinate, but hardly arousing.From an item-based perspective, this accords with Kulka's (1996) assumption that non-threatening content and determinacy are necessary, but not sufficient preconditions for kitsch classification.Of course, these item-characteristics have to be taken into consideration when we interpret results from multilevel modelling.
Independently of cultural background, kitsch was used as a derogatory term: Apart from mean liking ratings, this is clearly indicated by a negative correlation between liking and kitschiness that did not vary significantly between cultures.A moderately high interrelation between the two dependent variables also suggests that the relationship between liking and kitsch is an ambivalent one: Although it is widely acknowledged that a proclivity to "[k]itsch is to be found in every human being" (Schmidt, 1994, p. 136), it is rejected as overly simplistic and consoling.
What aspects of kitsch are most predictive of liking?Interestingly, arousal ratings allowed for the best discrimination between likable kitsch objects and rejected ones: While arousal was positively associated with liking, it was inversely related to kitschiness.At first glance it may seem that these findings support Berlyne's (1971) basic assumption, that aesthetic appreciation results from an increase in arousal which is brought about by something surprisingly new, ambiguous or otherwise enigmatic.Upon closer examination, however, the range of arousal ratings and other item characteristics do not fit in: Mean arousal ratings were low-to-medium and a majority of kitsch objects was described as familiar, determinate and non-threatening.This pattern is perfectly in line with Kulka's (1996) criteria for kitsch: In order to be comforting, kitsch objects need to appear familiar, unambiguous and perfectly harmless.Possibly, responses to kitsch also follow Berlyne's (1971) positive correlation of arousal and liking, but only on a lower level.This hypothesis must be tested based on physiological measures of arousal.Due to our choice of method-self-reports instead of bio-feedback-our approach seems prone to emotional appraisal theories.In accordance with Schachter and Singer's (1962) two-factor-theory of emotion, for example, arousal may also be regarded as the unspecific component of an immediate emotional response to kitsch.From this point of view, the close interrelation between liking of kitsch objects and arousal seems concordant with the widely shared assumption that kitsch is liked for its capacity to "spontaneously trigger an unreflective emotional response" (Kulka, 1996, p. 26).
Finally, it cannot be excluded that operationalization might have contributed to a positive link between arousal and liking.As ratings were used, the concept of unspecific arousal had to be translated into everyday language.Unfortunately, near equivalents of arousal from common language are either limited to certain contexts of use, or they convey an evaluative surplus meaning.In German and Serbian, for example, the term "erregend/uzbuđen" implies sexual arousal, while the alternative translation "aufregend/uzbudljiv" (exciting) may also be used to express aesthetic appreciation.Thus, our choice of wording ("aufregend/ uzbudljiv") might partly explain why stimuli were rated concordantly in terms of liking and arousal.
The SSVS-dimension of Self-transcendence reflects sympathy, tolerance and well-being of one's social environment.Across all samples, commitment to these values was positively related to liking of kitsch objects.Since kitsch ratings were unrelated to SSVS-scores for self-transcendence, it appears that participants scoring high on self-transcendence did not judge stimuli less harshly.They rate these objects as likable in spite of the fact that they recognize them as kitsch.Interestingly, self-transcendence figures prominently in Dissanayake's (1990) study on the arts in pre-modern societies: From a cross-cultural perspective, she argues that-with the only exception being Western avant-garde art-art production was originally about sharing "valued states of mind and body such as self-transcendence [and] intimacy with our fellows" (Dissanayake, 1990, p.132).Possibly the familiar, unambiguous and consoling qualities of kitsch compensate for these needs in Western societies (Ortlieb & Carbon, 2017).
Ambiguity of the parental image was associated with higher kitsch ratings.High tolerance of ambiguity regarding the parental image indicates that close relationships are not idealized, but reflected critically.Since students were overrepresented in all of the three samples, high ambiguity tolerance towards the parental image and higher kitsch ratings could be interpreted as statements of independent judgement from the younger generation: Although the parents are honoured and respected, the offspring claim autonomy with regard to aesthetic taste.This interpretation remains highly speculative since we cannot infer from our study data whether kitsch objects were truly perceived as "old-fashioned".In any case, kitsch has been repeatedly associated with intergenerational competition (Avenarius, 1920;Stemmle, 1931).
Independent of cultural background, strengths of interrelations between familiarity, perceived threat and liking varied between male and female participants: A positive association between familiarity and liking was only significant for female participants (see Fig. 3).Besides, non-threatening content seemed more relevant for women than for men.These results are consistent with previous research from empirical aesthetics indicating that women tend to prefer stimuli with non-threatening content over stimuli with threatening content (Ortlieb, Fischer, & Carbon, 2016).However, there was no further indication for gender differences in terms of liking or kitsch ratings.
To what extent do kitsch and kič reflect culture-specific notions of bad taste?So far, we have focused on cross-cultural commonalities.Multilevel modelling also suggests some minor cross-cultural differences: For participants from Serbia and Slovenia, kitschiness was positively associated with perceived threat, whereas participants from Bavaria rated kitschy objects as more determinate.Since the entire stimulus material was provided by researchers from Bavaria, it might be argued that these culture-specific results are due to the fact that kitsch objects were simply more familiar to the German subsample.However, there is no indication that stimuli were particularly typical for Bavaria: Apart from a Christmas bauble portraying Ludwig II of Bavaria, stimuli reflected an international canon of kitsch (e.g.Japanese maneki-neko; miniature Eiffel tower).Most importantly, there was no difference between participants from Bavaria and Slovenia regarding familiarity ratings.In the Serbian sample, familiarity ratings were even significantly higher than in the other samples.Anyhow, our culture-specific findings require replication and closer examination.For example, it would be desirable to conduct two otherwise identical studies based on a selection of typical kitsch objects from Serbia, respectively Slovenia.
As a very broad aesthetic concept, kitsch is applicable to "virtually anything subject to judgements of taste" (Cǎlinescu, 1987, p. 235).Due to the stimulus material used-images of everyday objects-findings cannot be generalized to phenomena outside of the visual domain (e.g.kitsch in music or literature).Strictly speaking, the present study has focussed on kitsch in interior decoration (e.g.German Nippsachen) at the expense of other aspects in the visual domain (e.g.kitsch in architecture).Additional research is needed to clarify the consistency of the kitsch concept across various domains.
Finally, with kitsch being a derogatory term, the issue of social desirability has to be raised: Especially since there is some preliminary indication for discrepancies between implicit and explicit kitsch judgements (Reiter, Ortlieb, & Carbon, 2015), future studies should include implicit measures.Participants with an academic background may be particularly reluctant to admit that they have a 'heart for kitsch'.Since our pilot study was based on small samples which were also very homogenous with regard to age and education, results must be interpreted very cautiously.The three models presented in this paper have to stand further testing with larger, more diverse samples and stimuli before they can be generalized.

Conclusions
From the results of our pilot study, we conclude that there is a common understanding of kitsch which prevails over culture-specific aspects.Moreover, this shared notion of bad taste seems widely in agreement with prominent theoretical accounts of kitsch.So far, our findings are limited to the visual domain and a small homogenous sample.Further quantitative and qualitative research is needed to validate these findings and to shed light on the culturespecific facets of kitsch.

Figure 1 .
Figure 1.Examples of typical stimuli from the study.All objects were photographed and provided by the Department of General Psychology and Methodology (University of Bamberg).

Figure 2 .
Figure 2. Procedure of the study.

Figure 3 .
Figure 3. Illustration of final model for liking including strengths of relations based on regression coefficients.Legend: solid lines-positive regression coefficients; dashed lines-negative regression coefficients; numbers-unstandardized regression coefficients.

Table 1
Likert-scales and translations

Table 2
Self-devised questionnaire for art expertise

Table 3
Coding of art expertise questionnaire(Carbon, unpublished)

Table 5
SSVS-scales and translations a The original SSVS by Lindeman and Verkasalo (

Table 6
Descriptive statistics at the level of participants and Kruskal-Wallis H values for differences between the three groups(N = 36)

Table 7
Descriptive statistics of averaged ratings of images and Kruskal-Wallis H values for differences between the three groups(N = 36)

Table 8
Results for crossed subjects and images conditional models for liking with random intercepts (Models I and II) and random slopes(Model III).Bold values are p <.05.

Table 9
Results for crossed subjects and images conditional models for kitschiness with random intercepts (Models I and II) and random slopes (Model III).Bold values are p <.05.