Show Me the Money: On Predicting Actual Purchases in Cross-National Sponsorship

The improvements in new media technologies in conjunction with the expansion of innovative opportunities for marketing and consuming sport have played direct roles in the globalization of sport. However, those in the Sport Management academic field are still trying to understand the effect of culture on sport consumer behavior. Guided by Hofstede’s cultural dimensions theory, the purpose of this study was to examine the sponsorship and cross-national relationships among the short-term/long-term orientation and individualism/collectivism cultural dimensions, attitude toward a sponsor, gratitude, purchase intentions, and actual purchases. Data were collected via longitudinal web surveys conducted with soccer fans from the United States, the United Kingdom, and India. The results from a structural equation model provided evidence that the individualism/collectivism cultural dimension had a significant effect on gratitude but not on actual purchases, and that the purchase intentions variable was a predictor of actual purchases.


Introduction
The magnitude of sport sponsorship has risen steadily in recent decades, with companies now allocating in excess of $60 billion globally (IEG, 2016).Rapid and ongoing development of new media technologies (e.g., broadband and mobile platforms) has reduced distances between international markets, and organizations-including sports teams-have become truly global brands (Amis & Cornwell, 2005;Santomier, self-centered or integrated into groups; Hofstede et al., 2010) from Hofstede's cultural dimensions theory.Previous research on cross-national consumer behavior has utilized Hofstede's cultural dimensions theory (De Mooij & Hofstede, 2010;Singh, 2006), which reflects aspects of a culture that can be measured relative to other cultures (e.g., scales from 0 for the most collectivist and short-term orientation country to 100 for the most individualist and long-term orientation one; Hofstede et al., 2010).To the researchers' knowledge, there have been no academic studies that have included Hofstede's cultural dimensions variables in a sport context.Moreover, despite an increasing number of studies measuring sponsorship outcomes such as attitude toward the sponsor and purchase intentions (e.g., Alexandris, Tsiotsou, & James, 2012;Biscaia, Correia, Rosado, Ross, & Maroco, 2013), there is a need for established theoretical frameworks explaining consumer decision-making that include variables such as gratitude and actual purchase behaviors (e.g., Kim, Lee, Magnusen, & Kim, 2015;Kim, Smith, & James, 2010;Mazodier & Merunka, 2012).Gratitude is believed to play an important role in understanding how marketing investments increase purchase intentions and sales growth (Palmatier, Jarvis, Bechkoff, & Kardes, 2009).In addition, although intent to purchase is commonly used in the sponsorship academic literature, a more accurate assessment of consumer behavior would include measuring actual purchase data (Gwinner & Bennett, 2008;Kim et al., 2015;Mazodier & Merunka, 2012).However, longitudinal research is needed if, for example, scholars are to understand the extent to which attitude toward a sponsor, and purchase intentions influence actual purchase behavior (Funk, Mahony, & Havitz, 2003;Yoshida & Heere, 2015;Yoshida, Heere, & Gordon, 2015).In this study the researchers also employed two demographics information (i.e., annual household income and household's decision maker) as control variables because scholars have found that they may predict actual purchase behavior (Armstrong, 1985;Sun & Morwitz, 2010).
The purpose of this study was to examine the sponsorship and cross-national relationships among the short-term/long-term orientation and individualism/collectivism cultural dimensions, attitude toward a sponsor, gratitude, purchase intentions, and actual purchases, controlling for the annual household income and household's decision maker.In doing so, the researchers believe there are three important contributions to the Sport Management literature: (1) extension of cross-national research on sponsorship effectiveness, (2) examination of jersey sponsorship specifically, due to the prevalence of this type of sponsorship at a global level, and (3) evaluation of sponsorship effectiveness using a potentially more robust outcome, actual purchase behavior.This research initiative was addressed by analyzing responses from soccer fans from the United States, the United Kingdom (U.K.), and India in the area of a sport sponsorship through a jersey sponsorship.

Cultural Dimensions
Hofstede's cultural dimensions theory was the theoretical framework utilized in this study, as much of the research on cross-national consumer behavior has utilized these dimensions, which reflect aspects of a culture that can be measured relative to other cultures (i.e., individualism/collectivism, short/long-term orientation, power distance, masculinity/femininity, uncertain avoidance, and indulgence/restraint; Hofstede et al., 2010).Hofstede defined culture as "the collective programming of the mind that distinguishes the members of one group or category of people from another" (2001, p. 9).
Although countries' cultural scores originally were produced in the early 1970s, many replications of Hofstede's study of different samples have provided evidence that the country ranking in his data is still valid (Hofstede et al., 2010).A benefit of Hofstede's cultural dimensions rests in being able to explain and compare aggregate national behavior (Magnusson, Wilson, Zdravkovic, Zhou, & Westjohn, 2008;Singh, 2006).Moreover, Hofstede has shown more than 400 significant correlations between his index scores and data from other sources that validate them (Hofstede, 2001).Hofstede initially created four dimensions (there are now six dimensions), assigned indexes on each to all nations, and linked the dimensions with demographic, geographic, economic, and political aspects of a society (Kale & Barnes, 1992), a feature unmatched by other frameworks.In addition, this framework is useful in formulating hypotheses for comparative cross-cultural studies.Consequently, Hofstede's operationalization of cultures (1980) is the norm used in international marketing studies (Dawar, Parker, & Price, 1996;Sivakumar & Nakata, 2001).
However, academic consideration dedicated to developing a better understanding of sport sponsorship, as part of international marketing, has yet to be conducted (Amis & Cornwell, 2005;Santomier, 2008;Yoshida & Heere, 2015).Furthermore, sponsorship can vary across geographic regions in that there are the "moderating effects of country" (Wang, Cheng, Purwanto, & Erimutri, 2011, p. 5), which convey that a sports team's fans are not all similar in the way they view sport sponsorships.Also, for firms to be successful in the global marketplace they need to grasp the distinction between countries and cultures in order to construct appropriate marketing strategies based on these distinctions (Craig & Douglas, 2001;Geng, Burton, & Blakemore, 2002).

Attitude toward the Sponsor
Attitude is defined as "a learned predisposition to respond in a consistently favorable manner with respect to a given object" (Fishbein & Ajzen, 1975, p. 6).The development of favorable attitudes toward a sponsor is a pivotal factor for sponsorship effectiveness (e.g., Alexandris, Tsaousi, & James, 2007), and, as such, are expected to point to positive behavioral intentions (Laczniak, DeCarlo, & Ramaswami, 2001).Thus, a sponsorship can change consumers' responses towards a specific sponsor, resulting in the development of positive attitudes towards the sponsor, which can then lead to increased consumer willingness to buy the sponsor's products (Swanson, Gwinner, Larson, & Janda, 2003).
Nevertheless, little is known about how sponsorship outcomes, and implicitly attitudes toward sponsors, work in a cross-national setting (Yoshida & Heere, 2015).The attitudes toward various sponsor categories may vary across countries, and in particular, as they pertain to jersey sponsorship.Although jersey sponsorship has been supported by many professional team sports and fans in Europe and internationally, this practice is relatively new in the U.S. and has been met by public opposition (Jensen, Bowman, Larson, & Wang, 2013).In North America, some fans consider jersey sponsorship on game uniforms an "untouchable territory" that should remain free from financial exploitation (Lukas, 2009).Therefore, fans from the U.S. might have a negative attitude toward a jersey sponsor compared to Indian and British fans, as the U.S. has a short-term orientation cultural dimension, which means that although people from the U.S. have a respect for [sports] traditions (Hofstede et al., 2010), they prefer the status quo, which is to not have a sponsor on a team's jersey (Lukas, 2009).Following the previous propositions from past research, it is hypothesized that: H 1 : Attitude toward the sponsor will have a direct positive effect on purchase intentions.
H 2 : Attitude toward the sponsor will have a direct positive effect on actual purchases.
H 3 : The short-term orientation cultural dimension will have a direct positive effect on attitude toward the sponsor.

Gratitude
Algoe, Gable, & Maisel (2010) defined gratitude as an emotional appreciation response for costly, yet intentionally provided, benefits from another individual.There is a limited amount of research in which scholars have examined the nature and role of gratitude within a sponsorship context (Kim, Kwak, & Bunds, 2012); however, the significant role of gratitude in a consumer behavior context has been emphasized in previous research (Palmatier et al., 2009).Further, while there appears to be general accord among sponsorship researchers that a distinctive aspect of sponsorship that differentiates it from traditional advertising is consumer gratitude (Meenaghan, 2001), few researchers have empirically analyzed the impact of gratitude on sponsorship effectiveness (Kim et al., 2015).Kim and colleagues (2010) found that feelings of gratitude significantly predicted consumers' intent to purchase the sport benefactor's products in the U.S. Considering that purchase intentions are, in effect, favorable intentions toward actually purchasing a product or service (e.g., Dees, Bennett, & Ferreira, 2010), then gratitude can be considered predictive of consumers' actual behavior.However, it is likely that the results of Kim and colleagues' (2010) study were impacted by the study's setting-the U.S. as trust, which is linked to gratitude (Palmatier et al., 2009), has been found to be influenced by national culture (e.g., Schumann et al., 2010).In addition, an implication derived from the academic literature is that there is a strong positive relationship between trust and collectivism, and a negative relationship between trust and individualism (e.g., Huff & Kelley, 2005).Therefore, gratitude toward a sponsor may be enhanced among citizens in collectivist countries (i.e., India) compared to citizens in individualist countries (i.e., the U.S., the U.K.).It would seem that collectivists, who appear to place more importance on relationships and nurture them with more care than individualists, would have higher levels of trust than individualists (Huff & Kelley, 2005), and hence, higher levels of gratitude.Furthermore, in collectivist societies, it is reasonable to expect that people think of themselves as members of a larger group (which could include sponsors), and expect members of the group to look after one another based on shared trust (Gwinner, 2005).In exchange for this support, one can feel absolute gratitude is owed to another (Hofstede et al., 2010).Thus, the researchers posit: H 4 : Gratitude will have a direct positive effect on purchase intentions.H 5 : Gratitude will have a direct positive effect on actual purchases.
H 6 : The individualism cultural dimension will have a direct negative effect on gratitude.The threat of non-response bias (Dillman, Smyth, & Christian, 2014;Jordan, Walker, Kent, & Inoue, 2011) was addressed by comparing demographic information (i.e., age, education, income, and gender) of the Phase 2 non-respondents with the demographic characteristics of the respondents that completed the questionnaire in Phase 2. Based on the results of these comparisons, the authors concluded there were no major differences between the Phase 1 sample and the Phase 2 sample.A comparison on variables that were measured was also made between early and late respondents in Phase 2 (i.e., first thirty respondents and last thirty respondents), as late respondents have been shown to be an appropriate proxy for non-respondents (Dooley & Lindner, 2003).No significant differences between early and late survey respondents were found.

Measures
The online survey included items adapted from previously validated instruments to measure three constructs: attitude toward the sponsor (Gwinner & Bennett, 2008), gratitude (Palmatier et al., 2009), and purchase intentions (Gwinner & Bennett, 2008;Hong, 2011).Slight modifications were made to suit the specific needs of this study.The items were presented in the same order for all three countries and contained identical designs.Actual purchase behaviors were collected using continuous numeric data (e.g., How many <sponsor name> products did you buy during the period between the first survey and today?), and 40.26% of the respondents acknowledged that they bought at least one sponsor's product.The researchers also employed two demographic items as control variables: annual household income (1 = less than $20,000; 2 = $20,000 -$59,999; 3 = $60,000 -$89,999; 4 = $90,000 or more) and household's decision maker (0 = no, I am not the primary decision maker in my family with regard to making purchases; 1 = yes, I am the primary decision maker in my family with regard to making purchases).Descriptive statistics for these demographics appear in Table 1 and Table 3.
Each of the three nations was assigned a national culture index score according to Hofstede's typology of cultural dimensions used in this study (i.e., scales from 0 to 100 for individualism/collectivism and for short-term/long-term orientation; Hofstede et al., 2010).Also, Hui & Triandis (1989), and Clarke III ( 2001) recommended scales with more response categories (i.e., scale points) as appropriate for cross-national research; thus, for this study the researchers used a 10-point Likert-type scale, anchored by "Strongly Disagree" (1) and "Strongly Agree" (10).Another reason for implementing a Likert-type scale with 10 points is that previous researchers have found strong differences in response styles between countries (e.g., Harzing, 2006).In particular, East Asian respondents have been shown to display a higher proportion of middle responses in comparison to U.S. and Canadian respondents who displayed more extreme response styles (e.g., Takahashi, Ohara, Antonucci, & Aakiyama, 2002).Dolnicar & Grün (2007) confirmed that this difference was also apparent between Australian and Asian respondents when scales with explicit mid-points are used.Similarly, Johnson, Kulesa, Cho, & Shavitt (2005), and Harzing (2006) suggested that national-level collectivism might be related to middle responses styles.This tendency is reinforced by the fact that most collectivist countries are characterized by an indirect communication style, where the expression of strong opinions is avoided (Hall, 1976).Therefore, researchers need to pay more attention to response styles in their data collection procedures as research clearly shows that there is stability in response style differences between countries (Harzing, 2006).
Finally, the attentiveness of survey participants was tested by inserting the statement "On this question please click on "Strongly Agree" so we can ensure you are paying attention" among the sponsorship outcomes' items to account for measurement error, which is a possible survey error that needs to be minimized to improve survey estimates (Dillman et al., 2014).

Data Analysis
Data were analyzed using SPSS 21 and AMOS 21.Before any analyses were conducted, the normality of the data was assessed.Then, to assess the measurement model, a confirmatory factor analysis (CFA) was conducted.Internal consistency of the constructs was measured through composite reliability (CR; Hair, Black, Babin, & Anderson, 2009).Evidence of convergent validity was evaluated through assessment of the average variance extracted (AVE) scores, while evidence of discriminant validity was established when the AVE score for each construct exceeded the squared correlations between that and any other construct (Fornell & Larcker, 1981).
The researchers utilized structural equation modeling (SEM; Byrne, 2010) to test the hypothesized relationships.Goodness of fit for the measurement and structural models was assessed with the ratio of chi-square (χ²) to its degrees of freedom, Tucker-Lewis Index (TLI), comparative-of-fit-index (CFI), goodness-of-fit index (GFI), root mean square error of approximation (RMSEA), and standardized root mean residual (SRMR).Sample sizes of 200 or more have been considered acceptable for use of SEM (Garver & Mentzer, 1999;Hoelter, 1983), thus the current sample was deemed suitable.

Results
The

Discussion and Implications
There is a growing interest in determining how sponsorship outcomes such as attitude toward a sponsor, gratitude, and purchase intentions function in relation to actual purchase behaviors at a cross-national level (Amis & Cornwell, 2005;Yoshida & Heere, 2015); yet few researchers have pursued such an investigation, highlighting the significance of this inquiry.Scholars have previously theorized upon the likelihood that consumer behavior might be culture-bound (e.g., Hofstede et al., 2010;Yoshida & Heere, 2015); in the present study the researchers have empirically-tested these assumptions in the context of sport sponsorship.The results suggest that gratitude in sponsorship may be responsive to the individualism/collectivism cultural dimension.Furthermore, a positive link between intentions and actual purchase behaviors has often been suggested (e.g., Dees et al., 2010), but there has been a lack of empirical work providing analysis of actual behaviors to support such an assertion.In the present study, the researchers added actual purchase behavior to assess whether there was a significant correlation between intentions and actual purchase behaviors at a cross-national level.
The link between attitude toward the sponsor and purchase intentions was significant, consistent with findings reported in prior sponsorship research (e.g., Alexandris et al., 2012;Biscaia et al., 2013); scholars have proposed that the development of positive attitudes towards a sponsor leads to increased consumer willingness to buy the sponsor's products.The attitude toward the sponsor variable, however, was not a significant predictor of actual purchases for the jersey sponsor's products in this study.There was no consistency between attitude and behavior in this study, which is contrary to the Theory of Planned Behavior, which includes the premise that actual behavior is influenced by an individual's attitude (Ajzen, 1985).This may indicate that, although international fans do have the intent to purchase the sponsor's products, they did not acquire the sponsor's products based on their positive and strong attitudes toward the sponsor (see Table 3 and Figure 2).It could be the case that, at a global level, positive attitudes toward the sponsor will normally lead to intentions to purchase, but there can be a complex transition from attitudes to action which will necessitate additional variables and increasing specificity in the measures of attitudes and behaviors (Funk, Haugtvedt, & Howard, 2000).Also, as Christensen (2006), andTsiotsou &Alexandris (2009) noted, sponsorship evaluation research is still at early stages, due to lack of established theoretical frameworks that can explain a consumer's decision making process.This aspect certainly warrants further investigation.
In addition, the short-term/long-term orientation cultural dimension was not a significant predictor of the attitude toward the sponsor, contrary to what was proposed with H 3 .When it comes to possibly not respecting [sports] traditions, it can be argued that the surveyed fans are not concerned if their favorite team has a jersey sponsor.This can be particularly true in the U.S. for jersey sponsorship, where some professional sport leagues (i.e., Major League Baseball, the National Basketball Association, the National Football League, and the National Hockey League) have yet to implement game-day jersey sponsorships.Therefore, as levels of the orientation cultural dimension decrease, attitude toward the sponsor's levels may not necessarily decrease.Moreover, making a case for a potentially successful implementation of jersey sponsorship in the U.S., the favorable opinions/attitudes fans have for a global company may be further enhanced if the affiliations between jersey sponsors and sport teams result in lower ticket prices, reduce team expenses, or assist in attracting/retaining star players (Jensen, Bowman, Wang, & Larson, 2012).Thus, there can be unique circumstances related to sport that international corporations should be aware of when they attempt to build more effective cross-national sponsorship initiatives.
In line with this study's findings on attitude toward the sponsor, gratitude was a significant predictor of purchase intentions, but not of actual purchase behaviors.This finding is contrary to prior speculation, in which scholars have stated that gratitude may lead to actual behavior (e.g., Kim et al., 2010;Palmatier et al., 2009).Expecting that feelings of gratitude generate an ingrained sense of psychological pressure to return the benefit received (Dahl, Honea, & Manchanda, 2005), it appears that fans' gratitude did not reciprocate through actual purchases of the team sponsor's products.Additionally, considering that the individualism/collectivism cultural dimension in this study had a significant negative effect on gratitude, it can be acknowledged for the first time that as there is movement to the individualistic end of the cultural dimension, gratitude levels will decrease.In this study the findings provide evidence that the collectivist fans (i.e., Indian fans) have more appreciation for the team's sponsor than the individualist fans (i.e., the British and American fans).
One possible way to increase gratitude levels of fans from individualistic countries could be through corporate social responsibility (CSR) strategies that include social objectives within sponsorship initiatives (Alexandris et al., 2012;Cunningham, Cornwell, & Coote, 2009), as the more consumers perceive an organization as socially responsible, the more they will trust products sold by that organization (Ko, Rhee, Kim, & Kim, 2014).CSR activities could inspire consumers' trust and, thus, higher levels of gratitude.Also, the importance of trust as a building block of relationships is even greater in the sport context, since support from fans for players, coaches, and teams is basically based on the created relationships (Lee, Bang, & Lee, 2013).
The Theory of Planned Behavior (Ajzen, 1985), one of the most widely applied theories in consumer research (e.g., Manning, 2009), includes the premise that a link can be present between purchase intentions and behavior, as "at its core, the Theory of Planned Behavior is concerned with the prediction of intentions" (Ajzen, 2011(Ajzen, , p. 1115)).The results from the current study provide evidence that purchase intentions is a predictor of actual purchases for the jersey sponsor's products, in contrast to Hickman's (2015) and Yoshida and colleagues' (2015) research on intentions and actual sport behaviors in the U.S., and, respectively, Japan.Furthermore, sponsorship should account for behavioral change in order to be proven effective (Amis & Cornwell, 2005), and the most desirable behavioral change from a sponsor's perspective is the influence on sales (Choi et al., 2011;Kim et al., 2015); thus, this research fills a gap regarding the examination of actual behaviors.However, the observed sponsorship outcomes accounted for only a small percent of variance in predicting behavior (i.e., 7 %); therefore, future research should take into account more (sponsorship) variables when studying behavior.Also taking into consideration past results, it appears that the link between intentions and actual purchases is a complex one that will require further research to more fully understand.
The link between the individualism/collectivism cultural dimension and actual purchase behaviors was not significant in this study, dissimilar to what was assumed as a greater pressure to conform with group norms when it comes to consumer behavior in collectivist cultures (e.g., Foscht et al., 2008).Sport has become a global phenomenon because of its capacity to attract people of different nationalities (Ratten, 2011).To exemplify, the declining birth rate and the aging U.S. population, and the large increase in middle class households in India together with its large population, have enticed more professional sport teams to India (Ratten & Ratten, 2011).Lately, the use across the globe of multiple new media/digital platforms (i.e., the convergence of telecommunications, computing and traditional media, such as social media, video and audio streaming, internet protocol television, online video gaming, etc.) in sports sponsorship communications, enables brands to employ multiple media channels and publicity methods in order to sell products (McAllister & Turow, 2002;Santomier, 2008).Consequently, sport consumers have a strong desire to be global citizens and this desire is manifested by their purchases of global brands' products (Kim & Heere, 2012).
Finally, the household production theory of Becker (1965) includes the premise that expenditure rises with income.Therefore, household income can have a positive influence on expenditures (Thibaut, Vos, & Scheerder, 2014).Moreover, people belonging to low-income groups are more inclined to be influenced by family members because, for example, they are financially dependent on the household's decision maker (Yousaf & Huaibin, 2013).For instance, a respondent will provide his or her own intention to purchase the product, but the decision maker in the respondent's household may play a role in the final purchase decision (Morwitz, Steckel, & Gupta, 2007).However, this study's control variables (i.e., annual household income and household's decision maker) were not found to be predictors of actual behavior.These findings can be possibly explained by the fact that consumers may be better able to predict their own future behavior about buying durable goods (e.g., electronics which this study's jersey sponsor is selling), as purchase decisions for durable goods are seen as more important to the consumer than purchase decisions for non-durable goods (e.g., food; Morwitz et al., 2007).Therefore, consumers may deliberate considerably about the purchase decision for durable goods as respondents might have reflected on all of the aspects of the purchase situation, not just their household income and decision maker.Moreover, these two respondents' demographics that might have influenced the analyzed sponsorship outcomes were controlled in the current analysis and were found not significant, further strengthening the validity of the outcomes' correlations.
To summarize, research analyzing multi-country effects of sponsorship is almost non-existent (Amis & Cornwell, 2005;Yoshida & Heere, 2015); therefore, highlighting the importance of this paper for the field, which thus far has included few studies which have advanced our understanding of how cultural differences among continents might affect the sport marketing strategies of international firms (Yoshida & Heere, 2015).It seems that some sponsorship variables (e.g., gratitude) can be influenced by Hofstede's cultural dimensions; therefore, sport research developed in one country should be assessed in other countries as well.Given these results and their broad implications, further investigation on the way countries influence the relationships among key variables in sponsorship contexts is warranted.

Limitations and Future Research
The authors of this study tested the application of sponsorship outcomes using just one team and sponsor.To test the validity and generalization of the research findings, future research will require a greater variety of sponsorship contexts, such as different sports, teams, cultural dimensions, product categories, purchase cycles (i.e., durable/non-durable goods), and sponsor levels.Also, the variation of approximately four to ten months in the number of days between collecting purchase intentions and actual purchases for the jersey sponsor's products (i.e., electronics) may not have been sufficient to be certain the actual purchase behaviors are accurate.However, literature from other academic fields suggest that it is not clear whether the strength of the intent-behavior relationship should increase or decrease with the length of time between the intent and behavior measurement (e.g., Morwitz et al., 2007).Second, while fans engaged in membership programs are often suggested to be highly identified with the team, the actual level of team identification for the CFC fans was not controlled in these research analyses; previous studies suggest that fans' link with the team tend to have a role on sponsorship outcomes (e.g., Alexandris et al., 2012).However, there were some cases where team identification didn't have a significant connection with actual (e.g., Yoshida et al., 2015).Future studies could compare fans with different levels of team identification to better understand its role on sponsorship outcomes.
While this research was developed within three local contexts (i.e., U.S., U.K., and India), it might not be applicable to consumers in other countries.Thus, researchers should test these findings with more countries where sponsorship has experienced growth, such as China and Brazil.Fourth, the current study considered only five variables, and other variables may help to further explain sponsorship effectiveness.In future studies, researchers ought to test cross-national differences with other sponsorship effects, such as awareness, fit, word of mouth, goodwill, and image transfer.Furthermore, future attempts should be made to also include additional variables in global sponsorship models, as the results could have been significantly influenced, for example, by sponsorship activation in each market/country.However, many sports organizations market themselves globally, instead of focusing on one geographic region, as they can save time and money (e.g., Ratten, 2011).Fifth, the data for this research was collected with the use of a purposive sampling method, which can make research methods susceptible to bias.However, the sampling judgments made by the authors were based on clear and analytical criteria in an effort to reduce bias.