Conducting Survey in Non-English Speaking Country: The Methodology Concerns

This paper demonstrates how to conduct survey in non-English Speaking country from the methodology perspective. Specifically, this paper discusses the research design and justifies why fixed design and quantitative methods, in terms of survey have been used to collect data. Moreover, the process of the translation-backtranslation of the survey is presented. The finally part concerns the issues about instrument refinement.


Research Design
The research design is "concerned with turning research questions into projects" (Robson, 2002, p.80), focusing on the relationship between research questions, strategy and methods. The relationship between these three issues needs to be handled in a consistent and systematic manner. According to Robson (2002), the general principle is that the research method and strategy must be appropriate for the research questions to be answered. It is appropriate to follow Robson's (2002) principle to design the research in a way that fits the research method and strategy to the research questions.

Fixed Design Strategy
According to Robson (2002, p.4), a fixed design means that "a very substantial amount of pre-specification about what you are going to do, and how you are going to do it, should take place before you get into the main part of the research study." He also classifies fixed and flexible designs as alternative classification criteria of research strategies to quantitative and qualitative research. A fixed design is theory-driven, which means that a conceptual model should be developed in advance of the research process. Accordingly, researchers should have a deep conceptual understanding before undertaking the fieldwork. By contrast, flexible design involves less prespecification, in that it evolves, develops and unfolds as the research proceeds Therefore, in order to test the conceptual model developed based upon an extensive review of the literature, a fixed design should be employed. The conceptual model can be used to guide the data collection, whereby the data is used in turn to examine the relationships presented in the theoretical model.

Quantitative Method
According to Neuman (2006, p.63), a quantitative method can be described as "an organized method for combining deductive logic with precise empirical observations of individual behavior in order to discover and confirm a set of probabilistic causal laws that can be used to predict general patterns of human activity". Applying a quantitative method can provide statistical evidence concerning the strengths and directions of relationships among constructs (Amaratunga et al., 2002). Hence, the measurement of constructs is a significant and integral aspect of quantitative research (Cavana et al., 2001).
Moreover, following the principle from Robson (2002) and Punch (2013), the research method used to conduct the research needs to be in line with research questions that are linked to the research strategy and suitable for data collection. As demonstrated previously, the research questions in this study explore the relationship between social capital and innovation with the mediating role of social decision-making constraints and the moderating effects of culture. A quantitative method should be used in the research to empirically investigate causal relationships in the conceptual model.  Figure 1. Framework for research design The framework for the research design is presented in Figure 4.1. The conceptual model can be viewed as an intermediary, linking the research question, fixed design (research strategy) and quantitative method (research method). In order to address the research question, this study adopts a fixed design as the research strategy to develop a conceptual model as the pre-specification before the fieldwork, and then employs a quantitative method to collect data to empirically test the conceptual model. Van Maanen et al. (2007) demonstrate that theory and method are interrelated. They also note that, although this view is widely held, the relationship is complicated and occasionally controversial across various communities of organizational researchers. Edmondson and McManus (2007) label the relationship between theory and method as fit. They provide recommendations for the appropriate methods when the theoretical development of a research area is in the mature stage. Research in the mature stage "leads to further refinements within a growing body of interrelated theories" (Edmondson & McManus, 2007, p.1159. As Chapters 2 and 3 have demonstrated, the impact of social capital on innovation management is in the mature stage. According to Edmondson and McManus (2007), the appropriate methods of data collection for the mature stage are survey, systematically coded and quantified interviews/ observation and secondary data. Secondary data is information or data that has already been collected and recorded by someone else for a different purpose (Blumberg et al., 2008). For example, the annual reports of public companies and information pertaining to financial markets are widely available and often used as secondary data. However, the main problem with using secondary data is that the data may not fit the requirement of the research questions.

Survey as a data collection method
As indicated previously, when the study adopts fixed design as the research strategy, surveys are one of the most commonly used methods of fixed design research (Robson, 2002). In particular, a self-administered survey can be defined as "a data collection technique in which the respondent reads the survey questions and records his or her own responses without the presence of a trained interviewer" (Hair et al., 2003, P.265). The selfadministered survey has been used as the data collection method owing to the following advantages. Firstly, selfadministered surveys are relatively cheap to produce and require less time than other data collection methods such as personal interviews. When the population requires a large number of respondents; therefore, a selfadministered survey can be used to collect data more quickly and economically. Secondly, using selfadministered surveys can reach participants who might otherwise be inaccessible. Thirdly, a self-administered survey can provide greater anonymity to the individual respondents.
Another alternative is the use of systematically coded and quantified interviews. The interview approach potentially could be used. It is, however, prohibitively time consuming and unnecessary, given that selfadministered surveys can be used to collect the same data more quickly and with less expense.

Snowball sampling and key informant as sampling strategy
Snowball sampling is a non-probability sampling technique whereby the researcher starts by collecting data from one or more firms, and using these initial respondents as key informants to identify further respondents (Robson, 2002). Snowball sampling is useful when it is difficult to identify members of the population from which respondents are to be drawn. It is for this reason that snowball sampling has been suggested as one of the most effective methods to access hard to reach populations (Valdez & Kaplan, 1998). In particular, respondents' social networks can be used to identify potential contacts and greatly expand the sample size (Cohen & Arieli, 2011). Moreover, snowball sampling is more likely to ensure a better response rate when gaining the initial collaboration from respondents (Lindlof & Taylor, 2010). In this sense, snowball sampling can be used to reach