Is craft beer consumption genderless? Exploratory evidence from Italy and Germany

Abstract

Purpose

This study aimed to investigate the gendered nature of craft beer (CB) consumption in Italy and Germany.

Design/methodology/approach

Data were collected through online surveys in Italy (N = 210) and Germany (N = 211). Based on an enhanced version of the theory of planned behaviour, mean value difference tests and moderated regression analyses with gender as a moderator were performed to test gender effects on CB consumption behaviour.

Findings

The study results provide evidence that the gap in CB consumption behaviour is not very pronounced. In the German sample, gender did not moderate the effects of the model components on behavioural intent. However, the study found significant mean differences in all model variables. In the Italian sample, gender moderated the effects of several components of the theory of planned behaviour on behavioural intention. Hence, CB consumption appears to represent an opportunity for Italian women to negotiate their womanhood in a historically masculine-dominated space.

Research limitations/implications

The limitations of these data are the focus on two specific countries, the use of small-sized samples and the prediction of behavioural intentions instead of actual behaviour.

Practical implications

The study may help marketing managers develop appropriate marketing strategies based on a better understanding of gender-specific needs in CB consumption.

Originality/value

This investigation provides the first comparative analysis of gender-specific behavioural patterns in CB consumption in two European countries characterised by notably different beer cultures.

Citation

Rivaroli, S., Lindenmeier, J. and Spadoni, R. (2020), "Is craft beer consumption genderless? Exploratory evidence from Italy and Germany", British Food Journal, Vol. 122 No. 3, pp. 929-943. https://doi.org/10.1108/BFJ-06-2019-0429

1. Introduction

There is a widespread belief that there are gender differences in the consumption of alcohol (Beardsworth et al., 2002). Despite there being no reason to doubt the validity of this statement, a clear understanding of the gender differences in beverage consumption is of paramount relevance. Much of the research on alcohol consumption has focused on the role of gender in craft beer (referred to here as CB) purchasing and consumption behaviour (Clarke, 2012; Snider, 2016; Darwin, 2017, 2018; Donadini and Porretta, 2017; Gómez-Corona et al., 2016b; Gómez-Corona et al., 2017a, 2017b; Muggah and McSweeney, 2017). This rising interest in uncovering the effect of gender on CB consumption seems to stem from three primary motivations. First, the stagnation of beer consumption has led large brewing companies to look at the sale of CB as an opportunity to improve their profit margins (Garavaglia and Swinnen, 2018). Second, despite men being heavier beer drinkers than women, approximately 50 per cent of the world's population is represented by females (The World Bank, 2017); this means that brewers can look at this part of the population as a major market opportunity. Finally, and most important for the present study, the commonly assumed reconfigured role of gender in beer consumption allows for hypothesising new consumption patterns that cross gender borders.

From an overview of the literature on the effect of gender on CB consumption, we can note that cross-national studies on the gendered nature of CB consumption are still missing. Among European countries, Germany and Italy represent two countries characterised by different beer cultures. Italy does not enjoy a long beer tradition; however, despite Italians being ‘light drinkers’ of beer, several studies have reported their rising interest in CB (Aquilani et al., 2015; Donadini and Porretta, 2017; Rivaroli et al., 2018). On the other hand, Germany has a long-standing tradition in beer production and consumption, and Germans are known for being ‘heavy beer drinkers’; however, almost 68 per cent of them are not inclined to drink CB (Statista GmbH, 2017). Accordingly, it makes sense to carry out comparative CB studies in countries with different beer cultures to increase the external validity of CB research; this rationale also holds for studies that assess gender effects.

2. Theoretical framework

In accordance with popular belief, the behaviour of men and women differs in different behavioural domains. Many researchers have explored this idea, and accordingly, it is not surprising that numerous studies have found gender differences in the consumption of food and beverages (Graham et al., 1998; Baraona et al., 2001; Liebman et al., 2003; Roberts, 2004; Kerr et al., 2004; Holmila and Raitasalo, 2005; Lieber, 2006; Mead, 2009; Wilsnack et al., 2009; Clarke, 2012; Snider, 2016; Chapman et al., 2018). Previous research considered gender differences that are caused by the biological sex. In this context, for example, the survey article by Holmila and Raitasalo (2005) highlights that the same amount of alcohol results in different alcohol blood levels among male and female consumers. This difference is due to the fact that the female body contains more lipids and less water. In addition, and among other factors, differences in metabolic processes may also explain that alcohol has a different effect on females. The present study does not explicitly focus on these biological explanatory factors of gender differences in alcohol consumption; rather, socialisation-related and cultural differences in alcohol consumption are considered. Due to the exploratory nature of our study, biological differences are not completely neglected (e.g., genetic explanatory factors).

Meyers-Levy and Loken (2015), in an insightful review article, show three complementary theories that help explain the gender differences in consumer behaviour. The first theory notes that differences in the levels of oestrogen and testosterone influence the brain activities of men and women differently. This theory helps explain differences in the expression of aggression, cognitive skills and the sensation of emotions, among others. The anthropological evolutionary theory represents a second approach that stems from biology, according to Meyers-Levy and Loken (2015). This theory claims that there are several programmes stored in human brains for problem-solving in different situations. In the process of evolution, specific programmes have prevailed, and because our male and female ancestors faced different tasks, it is believed that gender differences exist in these programmes stored in human brains. Gender differences are explained, in particular, by taking into account the problem of finding a partner (i.e., the mating problem). As a consequence of the rivalry for sex partners, it is believed that these evolutionary differences make men behave more aggressively and explain why men are more likely willing to take risks than women. Authors such as Fernie et al. (2010) or Gorka et al. (2015) show that a high risk propensity is related to positive attitudes towards alcohol and excessive alcohol consumption; therefore, one can assume that males positively assess CB. Furthermore, in accordance with this theory, men engage more often than women in pretentious and showy consumption behaviour to attract short-term mates (Meyers-Levy and Loken, 2015). Because CB consumption can be seen as hip and cool behaviour, one can assume that men in public places tend to buy products in this beverage category. The third explanatory approach named by Meyers-Levy and Loken (2015) is the sociocultural theory, which assumes, on the one hand, that gender differences in behaviour are due to physical differences and, on the other hand, that they are due to sociocultural influences. Within the scope of the sociocultural approach, gender socialisation theory assumes that gender roles and gender identity result from socialisation. According to Weeks et al. (1999) or Slentz and Krogh (2017), for instance, and based on Freud's (1925) research, early childhood development is assumed to be fundamentally different between boys and girls. More precisely, gender identity is built up to in preadolescence, the age of three years through the relationship with the mother (Chodorow, 1978; Cooper et al., 2013; Ruhl et al., 2015), and socialisation theory states that boys and girls experience a fundamentally different relationship with the mother. After the age of three years, these gender roles are further reinforced; for instance, boys and girls play different kinds of games (Roxas and Stoneback, 2004). In further personality development, social expectations affect individual behaviour by rewarding role-conforming behaviour and punishing non–role-conforming behaviour. The resulting socialisation disparities cause males and females to enter their adult lives with different ideas and goals (Betz et al., 1989). In accordance with former research studies (Kurt and Frimer, 2015, or Newman and Trump, 2019), men strive for promotion (i.e., agency orientation), whereas women strive for the successful completion of the tasks typically assigned to them (i.e., household activities and childcare) and for harmonious relationships (i.e., communal orientation). According to Meyers-Levy and Loken (2015), individuals are attracted to brands whose brand personality matches their own gender identity. Because advertising and other media, as well as the behaviour of male role models, often link an image of masculinity with beer consumption and drinking is regarded as a masculine behaviour (Medina-mora, 1994; de Visser and McDonnell, 2012; Darwin, 2018; Fugitt and Ham, 2018), it can be assumed that the perceived personality of CB is more firmly in line with a typical male gender identity. Hence, males should be more inclined towards CB than females because the consumption of CB helps men satisfy their desire to demonstrate their masculinity.

The approaches presented thus far help to predict gender differences in CB consumption and to assume that CB is more prevalent among male consumers. Gender-related differences in the sensory preference for beer and other alcoholic beverages (Mejlholm and Martens (2006) or Mora et al. (2018) for accordant findings) may be a cause for different consumption levels among male and female consumers. However, due to the multitude of CBs available for consumers, there should be CB available that satisfies the female preference (i.e., sweet and fruity flavours); hence, no general distaste among women can be assumed. In addition, there are further lines of thought that may help explain the missing or unclear gender differences in CB consumption. The aforementioned sociocultural theory, which focuses more on culture-based explanations, predicts varying effects across different cultures. Thus, gender differences in CB consumption should be lower in countries, such as Sweden or Germany, where gender equality is more pronounced (Meyers-Levy and Loken, 2015).

Furthermore, the structural theory of sex differences highlights the relevance of situational determinants and explains weak gender effects. Betz et al. (1989) and Dawson (1997) showed that women adapt their behaviour in male-dominated working environments to confirm the male gender role. If one assumes that the situations in which CB is bought or consumed are dominated by male role models (i.e., beer bars, pubs, brewery restaurants or CB conventions), the structural theory would predict that women adapt their behaviour to these situational circumstances and, consequently, prefer CB to the same extent as their male companions. Finally, multidimensional theories of gender identity do not conceptualise gender as a binary construct (i.e., males and females). Instead, and in line with authors such as Spence (1993) or Menon (2017), gender is assumed to consist of biological gender, psychological gender and individual attitudes towards socialised gender roles. The psychological gender comprises instrumental or agency-oriented (i.e., male gender roles) and expressive or communal (i.e., female gender roles) components. Accordingly, the gender identity of men may also include female role models and vice versa. In line with this notion, CB consumption represents an opportunity to negotiate female role models in a historically masculine-dominated space.

Moreover, men adapt to female consumption practices in the form of so-called gender journeys, according to Thompson and Holt (2004). An unambiguous classification of individuals into a binary gender scheme is therefore undoubtedly not possible. Following the gender identity theory on this point, the effects of a binary gender variable on individual CB consumption, for instance, can only be regarded as a proxy for a multifaceted gender identity. In a study on undergraduate students' alcohol consumption, Dempster (2011) reveals an accordant complex process of gender identity construction. Moreover, a negative attitude of women towards stereotypical female gender roles can encourage women to practise allegedly typical male behaviour such as drinking beer. In line with this notion, gender differences in CB consumption should not be pronounced. Confirming to this rationale, de Visser and McDonnell (2012) reveal nondifferences in actual alcohol consumption even though typical gender stereotypes appeared in their survey study.

3. Empirical study

3.1 Conceptual framework

This research takes a more in-depth look at the subject than a previous study conducted by Rivaroli et al. (2019). The conceptual framework of the present study is based on the theory of planned behaviour (TPB; Ajzen, 1991) and considers consumer self-identity (Sparks and Shepherd, 1992) and consumer desire for unique consumer products (Lynn and Harris, 1997) as additional psychosocial motivations that affect the individual's intention to consume CB.

3.2 Participants

An online survey was conducted in Italy (N = 210) and Germany (N = 211) between November and December 2017. A market research company provided access to representative online consumer panels and handled the process of recruiting respondents. Quotas ensured that the gender and age distribution of the sample matched the distributions in the basic population. In addition to a similar age distribution, the Italian consumer sample has a similar profile in terms of education level and willingness to pay for 0.4 lt of CB across the two genders, whereas a gender-specific difference occurred in terms of employment situation (Table I). However, this difference is typical for the basic population, with women occupying more part-time jobs than men. The respondents belonging to the German sample have a similar profile in terms of age and education level, whereas gender differences occur in terms of employment situation and willingness to pay for 0.4 lt of CB, with females having a higher willingness to pay for CB than men. When comparing the willingness to pay, a distinct difference between Italian and German customers appears. The lower willingness to pay for CB in Germany indicates a sceptical attitude towards CB that might be a result of the long-standing traditional German beer culture. In particular, the German purity law of 1519 (‘Reinheitsgebot’), which prohibits additives in beer, could have contributed to this cultural difference.

3.3 Questionnaire and measures

The questionnaire was developed in English and was then translated into the Italian and German language by native speakers. In the opening instructions of the survey, a detailed description of the term ‘craft beer’ was provided to the participants (i.e., ‘craft beer is a beer produced by small independent breweries, and it does not follow the pasteurisation and microfiltration processes during the production phase’). The construct measures adopted in this research were based on the existing measures from previous research (Ajzen, 1991; Lynn and Harris, 1997; Sparks and Shepherd, 1992) and were measured using seven-point Likert scales. All measures are listed in the Appendix.

3.4 Data analysis

A principal component analysis (PCA) with an oblique factor rotation (oblimin with Kaiser's normalisation) was performed to examine the unidimensionality of each construct. The sampling adequacy of each domain was tested using the Kaiser–Meyer–Olkin (KMO) measure and Bartlett's test of sphericity. The PCA showed an acceptable unidimensional structure for all the constructs. The KMO values ranged from 0.500 to 0.873 for the Italian sample and from 0.500 to 0.880 for the German sample (Table II), with a significant Bartlett's test (p < 0.05) for both samples (Hair et al., 1998). The rotated solution provided the component loadings for each construct. Each loading was higher than the commonly used threshold value of 0.500 and significant. In sum, the results of the PCA indicate the convergent and discriminant validity of the measurement scales. The reliability of each construct was tested using Cronbach's α coefficient, and values below the limit of 0.700 were considered unacceptable (Hair et al., 1998). The Cronbach's α values ranged from 0.728 to 0.968 for the Italian sample and from 0.715 to 0.959 for the German sample (Table II). Thus, the internal consistency of measurements was confirmed.

In a further step, an ordinary least squares (OLS) regression analysis was carried out, with the intention to consume CB (behavioural intention (BI)) as the dependent variable. The regression analysis considered attitude towards CB (attitudes (ATT)), social pressure towards CB consumption (subjective norm (SN)), individual perception of being able to control the consumption of CB (perceived behavioural control (PBC)), self-identity (SI) as a “real” craft consumer, desire for unique consumer products (DI), and CB-drinking experiences (past experience (PE)) as the independent variables.

Gender effects were tested using a moderation approach, with simultaneous interaction terms between gender (male as the reference category) and each independent variable. Gender was also considered as an independent variable. Multicollinearity was investigated by inspecting the variance inflation factors (VIFs), which examine the possibility that one construct is a linear function of the other (Kutner et al., 2004). The VIFs for the predictors in the Italian sample ranged from 1.77 to 3.38 with a mean of 2.24 (Table III), whereas for the German sample, the VIFs ranged from 1.46 to 3.32 with a mean of 2.68; therefore, the values of the sample are below the commonly considered cut-off value that signals potential multicollinearity issues (Hair et al., 1998).

4. Results

4.1 Descriptive statistics

Table II shows the descriptive statistics of the samples. Italian males showed a significantly stronger belief of being able to control their CB consumption (mean value = 5.68; standard deviation [SD] = 1.09; p = 0.052) than females (mean value = 5.34; SD = 1.36). In contrast, the mean values of all the other constructs were not significantly different between Italian male and female participants. However, considerable gender differences were found in the German sample for all constructs, with higher mean values for male respondents.

4.2 Moderated regression analysis

Table III depicts the results of the moderated regression analysis. For the Italian sample, the F-test (F(13, 186)=33.01, p < 0.000) indicated a good model fit of the regression model. In addition, the validated model explained well more than 67 per cent of the variance in the respondents' intention to consume CB. ATT towards CB predicted the intention to consume CB for both men (ATT: β = 0.190, p = 0.007) and women (ATT: β = 0.205, p = 0.000), without a significant gender difference. Social pressure and the desire for unique consumer products were not significantly predictive for either men (SN: β = 0.106, p = 0.115; DI: β = 0.012, p = 0.849) or women (SN: β = 0.072, p = 0.248; DI: β = 0.001, p = 0.981) and were also without significant gender differences.

Significant moderation effects of gender in the Italian sample were found for three drivers of CB consumption: PBC × Gender (β = −0.194, p = 0.083), SI × Gender (β = 0.455, p = 0.002) and PE × Gender (β = −0.222, p = 0.061). These moderation effects revealed that perceived behavioural control and consumption experience (social influence) have (has) a stronger effect on males' (females') inclination towards CB consumption. In line with the uncovered moderation effects, the belief of men of being able to control their consumption of CB (i.e., PBC: β=0.264, p = 0.004) and their self-perception as experienced CB drinkers (i.e., PE: β = 0.219, p = 0.017) were two relevant drivers of the intention to consume CBs among males. However, self-identification as CB consumers did not affect men's intention to purchase and consume CB (SI: β=0.012, p = 0.900) but was a significant behavioural driver for women (SI: β = 0.468, p = 0.000). In addition, a supplementary mediation analysis with multiple independent variables and attitude as a mediator, which was conducted with Hayes' (2012) MEDIATE macro, revealed a significant indirect effect of SI on behavioural intention (β = 0.130, lower limit confidence interval (LLCI) (95 per cent) = 0.057, upper limit confidence interval (ULCI) (95 per cent) = 284) for all Italian respondents.

The regression results for the German sample (Table III) revealed the predictive power of the validated model (F(13, 186) = 34.7, p < 0.000), which accounted for more than 68 per cent of BI's observed variance. Both ATTs towards CB (ATT: β = 0.299, p = 0.000) and social norms (SN: β = 0.140, p = 0.063) were significantly associated with the German male respondents' intention to consume CB (BI). Again, a mediation analysis with multiple independent variables and attitude as a mediator with Hayes' (2012) MEDIATE macro was conducted. The mediation analysis revealed significant and positive indirect effects of PE (β = 0.040, LLCI (95 per cent) = 0.002, ULCI (95 per cent) = 0.091) and DI (β = 0.062, LLCI (95 per cent) = 0.028, ULCI (95 per cent) = 0.106). As shown in Table III, the regression approach revealed no significant moderation effects of gender; hence, the drivers of the inclination of CB consumption do not differ among German males and females.

5. Discussion

Prior investigations have documented the reconfigured role of gender in beer consumption, highlighting how consumption patterns are changing; Darwin, for example, reports that beer has become a totem of femininity through the consumption of CB (Darwin, 2017). The findings of our investigation provide an initial picture of gender-specific CB consumption patterns in Italy and Germany, two European countries characterised by different beer cultures. To assess whether CB consumption behaviour represents an example of gender convergence (or difference) in beer consumption, we conducted a moderated regression analysis with gender as a moderator.

In line with the findings of Gómez-Corona et al. (2016a), we identified gender differences in attitudes towards CB consumption both in the Italian and German samples. This finding is in line with the predictions of the gender socialisation theory, among others. However, these gender differences are only moderate. On the other hand, and in particular in the Italian sample, the findings of the present study point in the direction of the findings of Darwin's study that shows that women do not exhibit any frustration with being considered CB drinkers and suggests that CB consumption is genderless. In addition to other theoretical explanations, this finding may be the result of greater gender equality in the countries considered.

This study's finding reveals that the individual self-identification as a CB drinker has no strong direct effect on the intention to consume CB and that no gender differences exist in this regard. In particular, Italian women are as inclined as Italian men to consider themselves as typical CB drinkers (Table II). In line with the theory of multidimensional gender identities, these findings support those of Chapman et al., confirming that ‘to be a real craft beer drinker and a woman are not two mutually exclusive categories’ (Chapman et al., 2018: 306). Gender-specific differences in consumer identities have disappeared, or multidimensional identities have blurred the effect of the binary gender variable in the CB domain in this regard.

In the Italian sample, we found that men's intention to consume CB was more strongly associated with PBC and with PE than that of women. These findings support Darwin's considerations, according to which men tend to legitimate the consumption of any beer by invoking their ability to take a risk in alcohol consumption and their experience as beer drinkers (Darwin, 2018). This finding reflects a typical male consumer identity, which is probably not limited to the consumption of CB but is valid for all alcoholic beverages. For Italian males and conforming to the anthropological evolutionary theory, drinking CB appears to be a showy consumer behaviour through which self-images can be enhanced. Furthermore, this result is in agreement with Chapman's notion, which states that gender is actively renegotiated and accomplished through CB consumption (Chapman et al., 2018); in particular, Italian men seem to renegotiate their role as ‘real’ men who have everything under control in the realm of CB consumption, which is a consumption arena in which the gender hierarchy is progressively losing its power. The significant and positive gender × SI moderation effect in the Italian subsample can be regarded as an indication of the ongoing process of gender convergence in the CB domain. In sum, the contrary moderation effects of gender may be the reason for the consistent behavioural intentions in Italian men and women.

When looking at the German sample and in contrast to the Italian sample, no significant moderation effects of gender were found. This result could be understood as the absence of gender barriers within the realm of CB consumption in Germany; that is to say, CB consumption is genderless. These insights also echo the argument of Holmila and Ratislao, who have raised the question of the convergence of male and female drinking (Holmila and Raitasalo, 2005). However, the descriptive statistics show distinct gender-specific differences in the mean value of all considered constructs. Here, the lower mean values of BI and ATT indicate that German females are less inclined towards CB than are German males. This result is in line with the anthropological evolutionary theory, which predicts that males are more inclined towards risky or adventurous behaviour such as alcohol consumption. Interestingly, women's willingness to pay for CB contradicts these findings (Table II). Women may be willing to pay more for CB either because they assume a positive price–product quality relationship when making their purchasing decisions or they plan to consume less CB and thus believe in having a higher per-item budget. Furthermore, the significant mean value difference that appeared for the PBC construct is in line with the moderated regression findings in the Italian sample. In addition, the significant mean difference in SI indicates that there is still a gender-specific identity difference in the CB domain. In summary, one can state that there are, to a certain extent, gender differences in CB consumption in Germany. The mean differences in gender indicate a steady and still masculine value system regarding beer consumption. The absent moderation effect of gender shows that gender-specific attitudes and beliefs are stable and rather invariable in Germany. Against the background of the centuries-old beer culture in Germany, these findings have face validity.

6. Conclusions

To the best of our knowledge, this research is the first comparative study that investigates the gender differences in CB consumption in Europe. This research fills a gap in the literature by exploring whether CB consumption in Italy and Germany is genderless. The findings indicate that in a consumption sphere strongly associated with masculinity, such as beer consumption, new consumption patterns in which femininity is not undermined per se are possible.

These findings can help industry managers to create appropriate marketing strategies to better capture gender-specific preferences in CB consumption. In particular, the CB segment could be an entry point for women into the typically male-dominated beer market. Marketing managers could try to specifically target women who already consume CB and get them excited about other products from their product portfolio. This strategy could alleviate the decreases in overall beer sales. In Italy, marketing managers could promote a feminine product image of CB by different means of communication, such as print advertisements or social media campaigns. The resulting higher SI should boost females' intention to buy CB. Advertisements that are seen by Italian men should encourage the men in their CB expertise and experience, and this should get more men to buy CB. Because the drivers of CB consumption in the validated model did not differ across the two genders for the German respondents, a gender-specific segmentation strategy does not appear effective in Germany. Instead, marketing managers could use celebrity and noncelebrity endorsements to increase the social pressure to consume CB that appears to be a distinct driver of CB consumption. In addition, a similar effect could be achieved using social marketing that aims to boost electronic word of mouth (eWOM) behaviour.

The study results should be interpreted against the background of the existing study limitations. One of the limitations of our data is the small sample sizes that might result in the limited validity of the study findings. Thus, further large-scale quantitative replication studies could undoubtedly strengthen the findings. Furthermore, the present study solely focuses on two European countries, and thus, the external validity might be limited. However, the research design explicitly considers two countries with distinctly different beer cultures, and this approach should have increased the generalisability of our findings. Finally, the validated model considers behavioural intentions as the dependent variable and intentions never correctly translate into actual behaviour. Hence, and based on market data from CB manufacturers, future research should try to explain actual CB purchasing behaviour based on the considered behavioural model.

Moreover, the present study findings provide avenues for future research; the findings of this study expand the knowledge on the gendered nature of CB consumption and bring forward valuable insights into the consumption of crafted food products. Our results are encouraging and should be validated in specific age cohorts that are relevant for the CB sector (e.g., millennials), as well as in other countries, to shed light on the gender differences in CB consumption. In our study, we found substantial differences between the CB consumption patterns of Italians and Germans. Such findings support those that have demonstrated that drinking behaviour is also related to the culture of different societies (Baraona et al., 2001; Gómez-Corona et al., 2017a). For example, even though Italians and Germans both express positive attitudes towards CB, the perceived social pressure (i.e., social norms) has a significant positive effect on the intention to consume CB for only the German sample. The long-term tradition of beer production and consumption has probably made this aspect relevant to the decision-making process of German consumers. Instead, previous CB drinking experiences and the conviction of being able to control the consumption of CB appeared as two fundamental motives that are relevant for the explanation of Italians' intention to consume CB. Thus, future studies can address these cultural differences in CB consumption more deeply. For example, the culturally diverse United States, with their long history of migration from different parts of the world, could be a fruitful research object. Moreover, studies that elaborate on different consumption patterns across various subcultures (e.g., urban vs rural beer cultures) could result in fruitful insights. Finally, as the study is based on questionnaire data, it must be noted that the study results are, to some extent, descriptive and one cannot make statements regarding causal dependencies. Accordingly, future study on CB consumption should increasingly consider experimental research design.

Table I

Sample description

	Italy			Germany
	Male (N = 100)	Female (N = 100)	p valuea	Male (N = 100)	Female (N = 100)	p valuea
Age
Mean (SD)	40.2 (11.2)	38.1 (11.3)	0.188	51.4 (14.9)	48.6 (13.2)	0.159
Min	19	18	–	21	20	–
Max	68	63	–	87	72	–
Education level	–	–	0.670	–	–	0.154
With an academic degree	44	47	–	49	39	–
Without academic degree	56	53		51	61	–
Employment situation	–	–	0.075	–	–	0.000
Full-time	72	57	–	56	28	–
Part-time	10	18	–	10	32	–
Other b	18	25	–	34	40	–
Willingness to pay for 0.4lt of craft beer	–	–	0.470	–	–	0.008
< 3.00€	11	20	–	42	61	–
3.00€–4.00€	32	26	–	29	26	–
4.00€–5.00€	29	27	–	19	11	–
5.00€–6.00€	18	19	–	8	0	–
> 6.00€	10	8	–	2	2	–

Notes: ^a Significance level of chi-square test for categorical data, or 2-sample t test for continuous data; ^b Significance level of chi-square test

SD = standard deviation.

Table II

Constructs and descriptive statistics (Mean, standard deviation in parenthesis)

Constructs	Number of items	Italy					Germany
		KMO	Cronbach's α	Male	Female	p valuea	KMO	Cronbach's α	Male	Female	p valuea
BI	2	0.500**	0.864	5.87 (1.43)	5.59 (1.78)	0.229	0.500**	0.843	4.10 (2.14)	3.32 (2.00)	0.008
ATT	6	0.826**	0.905	5.91 (1.00)	5.66 (1.46)	0.156	0.867**	0.942	4.65 (1.50)	3.89 (1.60)	0.007
SN	5	0.676**	0.728	5.37 (0.95)	5.39 (1.02)	0.920	0.831**	0.848	3.91 (1.45)	3.57 (1.35)	0.090
PBC	4	0.673**	0.735	5.68 (1.09)	5.34 (1.36)	0.052	0.650**	0.715	4.23 (0.93)	3.79 (0.91)	0.008
SIb	2	0.500**	0.968	4.92 (1.63)	4.83 (1.83)	0.714	0.500**	0.938	3.57 (1.91)	2.73 (1.76	0.002
DIc	7	0.873**	0.861	2.99 (1.19)	2.93 (1.22)	0.744	0.880**	0.862	4.15 (1.34)	3.71 (1.46)	0.029
PE	2	0.500**	0.941	4.59 (1.67	4.64 (1.84)	0.856	0.500**	0.959	3.15 (1.94)	2.01 (1.59)	0.000

Notes: Bartlett's test for sphericity: *, ** significant at p < 0.10; p < 0.05, respectively.

BI = behavioural intention; ATT = attitudes; SN = subjective norm; PBC = perceived behavioural control; SI = self-identity; DI = desire for unique consumer products; PE = past experience.

^aSignificance level of 2-sample t tests; ^bAdapted by Sparks and Shepherd's (1992) scale; ^cAdapted by Lynn and Harris's (1997) scale

Table III

Ordinary least squares regression analysis of the gender convergence of craft beer's drinking intention in Italy and Germany

Regression analysis	Italy				Germany
	Coef.	S.E.	p>|t|	VIFa	Coef.	S.E.	p>|t|	VIFa
ATT male	0.190	0.068	0.007**	2.23	0.299	0.072	0.000**	3.23
ATT female × male	0.015	0.086	0.864	–	−0.141	0.090	0.119	–
SN male	0.106	0.067	0.115	1.77	0.140	0.078	0.075*	2.98
SN female × male	−0.034	0.092	0.714	–	0.002	0.105	0.987	–
PBC male	0.264	0.089	0.004**	2.13	0.087	0.065	0.183	1.46
PBC female × male	−0.194	0.111	0.083*	–	−0.085	0.086	0.323	–
SI male	0.012	0.099	0.900	3.38	0.076	0.110	0.490	3.32
SI female × male	0.455	0.142	0.002**	–	0.220	0.143	0.125	–
DI male	0.012	0.060	0.849	1.85	0.040	0.068	0.556	2.24
DI female × male	−0.010	0.078	0.895	–	−0.047	0.083	0.576
PE male	0.219	0.090	0.017**	2.08	0.066	0.091	0.472	2.84
PE female × male	−0.222	0.118	0.061*	–	0.136	0.134	0.310	–
Constant male	0.032	0.076	0.672	–	−0.021	0.082	0.795	–
Constant female × male	−0.075	0.109	0.491	–	0.062	0.112	0.578	–

Notes: *, **significant at p < 0.10; p < 0.05.

Dependent variable: BI = behavioural intention; ATT = attitudes; SN = subjective norm; PBC = perceived behavioural control; SI = self-identity; DI = desire for unique consumer products; PE = past experience.

^aVIF: variable inflation factor referred to the whole sample.

The difference of female's coefficients compared to male's coefficients is in italics. Gender (male as reference)

Appendix Measures

Behavioural intention (BI)

1. * Will you consume craft beer soon? (1 = Certainly / 71 = Definitely not);

2. Will you consume craft beer soon? (1 = I do not want to consume craft beer / 7 = I want to consume craft beer).

Attitude (ATT)

1. * For me to drink craft beer is 1 = Tasty / 7 = Disgusting;

2. * For me to drink craft beer is 1 = Useful / 7 = Useless;

3. For me to drink craft beer is 1 = Negative / 7 = Positive;

4. * For me to drink craft beer is 1 = Healthy / 7 = Unhealthy;

5. * For me to drink craft beer is 1 = Intelligent / 7 = Foolish;

6. For me to drink craft beer is 1 = Unpleasant / 7 = Pleasant.

Subjective norm (SN)

1. Think about people that are important for you: How many of them drink craft beer? (1 = No one / 7 = All);

2. * How probable is that they would drink craft beer? (1 = Highly probable / 7 = Improbable);

3. How many of them could drink craft beer? (1 = 0 per cent / 7 = 100 per cent);

4. * Do they approve or disapprove the consumption of craft beer? (1 = Approve / 7 = Disapprove);

5. How many think that drinking craft beer is a good alternative to the consumption of other beers? (1 = No one / 7 = All).

Perceived behavioural control (PBC)

1. * It is simple/easy for me to consume craft beer. (1 = Agree / 7 = Disagree);

2. * I can completely control my consumption of craft beer. (1 = Agree / 7 = Disagree);

3. I feel able to control the consumption of craft beer. (1 = Totally false, 7 = Totally true);

4. * In the next future, consuming craft beer will be for me: 1 = Likely / 7 = Impossible.

Self-identity (SI)

1. I consider myself as a typical craft beer consumer (1 = Strongly agree, 7 = Strongly disagree);

2. In the future, I could consider myself as a typical craft beer consumer (1 = Agree very strongly, 7 = Disagree very strongly).

Desire for unique consumer products (DI)

1. I am very attracted by special beers (1 = Completely agree, 7 = Completely disagree);

2. I have the tendency to be nonconformist rather than following the trends of the moment (1 = Completely agree, 7 = Completely disagree);

3. * I have the tendency to buy products that are hard to be found (1 = Completely agree, 7 = Completely disagree);

4. * I prefer to use personalised products rather than standard ones (1 = Completely agree, 7 = Completely disagree);

5. * I rarely miss the opportunity to personalise things that I'm buying (1 = Completely agree, 7 = Completely disagree);

6. * I like to try new products and services before others (1 = Completely agree, 7 = Completely disagree);

7. * I like to buy in shops that sell different and uncommon products (1 = Completely agree, 7 = Completely disagree).

Past experience (PE)

1. How often have you consumed craft beer within the last six months (1 = Never, 7 = Usually);

2. How often have you consumed craft beer within the last year (1 = Never, 7 = Usually).

Notes: * Reverse-scaled