5. Propiedades de los cardinales de la forma κ λ
5.6. La hipótesis de los cardinales singulares
As discussed in Section 4.6 and identified in Stage Two of Figure 4.1, upon determining the appropriate sampling plan, the next step involves identifying the method by which data are collected. Within the marketing literature, three common approaches to gathering primary data are identified – survey, observation and experiment (Malhotra, 2015). The survey approach was adopted in this study for the following reasons. First, as discussed in Section 4.5, given the positivistic and descriptive nature of this study, the use of primary data sourced from surveys is appropriate due to the specific information required to test the research questions and hypotheses identified in Chapters One and Three respectively. Second, implementing surveys permits access to large sample sizes at a relatively low cost and facilitates the administration of questions and answers to tap into factors and relationships that are not directly observable (Cavana et al., 2001; Groves, Fowler, Couper, Lepkowski, Singer, & Tourangeau, 2011). Third, surveys are widely used as a means of data collection in similar areas of study (e.g., Vorhies et al., 2011; Huang & Tsai, 2013; Hirvonen & Laukkanen, 2014).
The implementation of the survey approach can be undertaken in three ways – person-administered, computer-administered and self-administered (Burns & Bush, 2006; Groves et al., 2011). The presence of the researcher and respondent is required when data is collected through the person-administered approach, while computer- and self-administered approaches can be implemented without the presence of the researcher (Kaplan, Sieber, & Ganiats, 1997; Groves et al., 2011). While each approach has its own advantages, they are also not without their challenges. For example, while the person-administered approach allows the researcher to obtain high response rate and rich feedback, it may
however introduce interviewer bias and involve high costs in monetary and temporal terms (Kaplan et al., 1997; Hair et al., 2002). Similarly, while the computer-administered approach increases the speed of administration and reduces interviewer bias, it however incurs high set-up costs and engenders confidentiality issues (Moutinho & Chien, 2007). Likewise, while the self- administered approach is cost-effective and allows the researcher to obtain large amounts of data without introducing interviewer bias, it can however lead to lower response rate and respondent fatigue, especially when the survey is exceptionally long (Kaplan et al., 1997; Malhotra, 2015). In light of these benefits and challenges associated with each survey method, the self-administered approach was adopted, and the rationale for this choice is detailed next.
First, because the presence of the researcher is not required in self- administered surveys, it therefore eliminates the threat of interviewer bias. Second, because it incurs less start-up costs and enables the researcher to gain access to large samples, it therefore represents a cost effective approach that suits the requirements of this study. Further, the self-administered approach has been widely adopted by scholars in areas of research similar to this study (e.g., Vorhies et al., 2011; Huang & Tsai, 2013; Hirvonen & Laukkanen, 2014).
Self-administered surveys can be distributed through a variety of means, including mail and drop-and-collect which are methods commonly adopted by scholars in similar areas of study (e.g., Ngo & O’Cass, 2009; Vorhies et al., 2011; Stock, 2014). The mail survey approach administers surveys through the employment of postal systems (e.g., letter, fax, email) which permits a wider reach at a relatively low cost (Ibeh, Brock, & Zhou, 2004; Malhotra, 2015). The drop-and-collect approach, on other hand, is argued to yield higher response rates
due to the interpersonal interaction that occurs during the distribution and collection of surveys personally by the researcher (Ngo & O’Cass, 2009; Sok & O’Cass, 2011a, b). Thus, in light of their respective benefits and following previous research (e.g., Lee et al., 2008; Stock, 2014), both mail and drop-and- collect approaches were adopted. Consistent with Stock (2014), initial contact during pre-screening further revealed that some respondents preferred to have the surveys mailed to them while some dropped to them.
Following Lee et al. (2008), this study is operationalised at the brand level, each brand was treated as an independent unit and respondents were recruited from each of these independent brand units. A customised brand database was developed by a professional database supplier, consisting of the targeted brands embedded within the product categories identified in Section 4.6.1.1 as well as the contact details of the key management personnel who were associated with the identified brand. This procedure permitted the establishment of a direct contact with informants and as discussed in Section 4.6.1.2, emphasis was given to recruiting from each brand unit: (A) the most senior managerial person responsible for and actively involved in overseeing the day-to-day management of the brand (e.g., Senior Brand Manager); and (B) the manager in the brand unit who reported directly to the first informant (e.g., Sales Manager).
Respondents were contacted by telephone and briefed about the study. Consistent with previous studies (e.g., Auh & Menguc, 2007), respondents were promised a summary of key findings of the study as an incentive to encourage participation. As discussed previously, upon the provision of consent to participate, surveys were mailed to respondents or delivered via the arrangement of an appointment on the basis of their indicated preference during pre-screening.
Following previous studies (e.g., O’Cass & Weerawardena, 2010; Murray et al., 2011), in order to minimise variability in responses, specific instructions were given to informants to ensure they responded to each question and/or statement in reference to the brand they managed which was specified in the survey.