4.5 At the Method and Technique Levels
4.5.5 Instrument Design and Constructs and Items Developments
Survey questionnaires are techniques of collecting data where individuals are required to respond to a similar set of questions formally designed in predetermined order as interview schedule or questionnaire involving only proportion and sample of the population (Ticehurts & Veal 2000; De Vaus 2002). Ticehurts and Veal (2000) stress that it is useful when the research questions indicate the condition for relatively structured data and when data are needed from samples representative of a defined population. This is why the design of the questionnaire is very important in business research and influences the structure and content of the questionnaire survey affecting its accuracy and relevancy (Sekaran 2003; Zikmund 2003; Burns & Bush 2006). Ticehurts and Veal (2000) emphasise the concepts and constructs involved and the relationships being investigated should be clear and guide the questionnaire design process. Tanur (1992) further identifies several factors that an investigator needs to consider when designing a survey questionnaire, this includes: decide on what to look for and find, keep questions simple and clear, decide on choice of closed-ended and open-ended questions, avoid leading questions, consider the arrangement of questions, and pre-test questions. Other factors are: avoid leading questions implying certain answers, loading questions slanted with social desirability and biased with emotional accuse, double barrelled questions addressing a number of issues, and burdensome questions demanding the respondent’s memory (Sekaran 2003). In the survey questionnaire, the questions and items were worded both positively and negatively to minimise the tendency of mechanically circling the points at the end of the scale (Sekaran 2003; Zikmund 2003). These questions usually act as cognitive speed bumps that require respondents to engage in controlled, as opposed to automatic, cognitive processing
The process of the initial development of the survey questionnaire has two stages. First, the survey questionnaire is developed during a period of six months where the items used to measure the constructs in the initial survey questionnaire are based on extensive reviews of related literature,
two pilot group discussions with 20 individuals (i.e. including SMEs owners/managers, academic researchers, and industry and market experts), and attendance of key seminars and conferences in Australia, Germany, Oman, Switzerland, United Kingdom, United Arab Emirates, and United States to collect additional qualitative information. Second, the drafted survey questionnaire is pre-tested for refinement, deletion, and addition of questions and items and adjustment is made to its formatting. The survey questionnaire was printed in a colour booklet format and was divided into sections and subsections with separate topics (Dillman 2007). In contrast to Dillman’s (2007) recommendation, it was not feasible to have the back cover of the survey questionnaire contain an invitation for respondents to provide further comments and suggestions. Instead, questions (i.e. closed-ended and open-ended) were continued on the back page and some space left at the end of the page. An expression of appreciation and a reminder of the purpose of the research study to respondents for participation were included at the end of the survey questionnaire. Altogether, the entire design process included: specific constructs development, operational definitions, draft questionnaire preparation, questionnaire pre-testing and modification, survey type specification, questionnaire pre-assessments, and final questionnaire administration (Churchill 1979; Zikmund 2003; Malhotra 2004).
4.5.5.1 Measurement Scales Determination
The designed questions and items capture quantifiable data allowing the investigator to conduct statistical analysis by using different mathematical software programs. This is to determine the correlations between the survey questionnaire respondents and to use the emerging patterns as a foundation for deriving conclusions and formulating recommendations. The questionnaire uses multiple-item measurement scales to ensure that the overall observed score is a reliable reflection of the underlying true score and the improvement on the confidence level of measure by proving greater exploratory power (Peat et al. 2002). Further, the measurement of constructs requires a system for organising information into a different level of measurements, which include nominal, ordinal, interval, and ratio scales (Cooper & Schindler 2003; Zikmund 2003; Davis 2005; Burns & Bush 2006; Hair et al. 2006; Manning & Munro 2006; Neuman 2006). First, the nominal-level measurement identifies only differences in types among the categories of constructs and classifies objects, individuals, and groups. Second, the ordinal-level measurement identifies differences among the categories of constructs allowing the categories to be ordered or ranked and provides information about relative amount of traits possessed by objects, individuals, and groups. Third, the interval-level measurement identifies differences among the attributes of constructs, ranks of
categories, and measures of distances; however, there is no true zero. Finally, the ratio-level measurement identifies differences among the attributes of constructs, ranks of categories, and measures of distances; however, there is a true zero making it possible to state the relations in terms of proportion and ratio. The lowest and least precise level of measurement is nominal and the highest and most precise level of measurement is ratio (Neuman 2006). Neuman (2006) argues that the quantitative method can use different measurement scales in a survey questionnaire to capture intensity, direction, level, and potency of constructs along the continuum. In designing the survey questionnaire, nominal, ordinal, and interval scale-level measurements were used in order to measure the objective and subjective characteristics of the respondents and their firms.
Zikmund (2003) and Davis (2005) further identify two groups of scales that are rating and attitude with each one having advantages and disadvantages. The rating scale is to evaluate a phenomenon at a period along a continuum or in a category. It includes graphics, itemised, and comparative. The attitude scale is to provide the respondent’s predisposition toward phenomena and is easy to respond it but it cannot provide the distinction in the respondents’ attitudes. It includes Likert and Sematic differential. In survey questionnaire research, Likert scale is used in which respondents express their attitudes and responses to propositions and the importance they attach to constructs in terms of ordinal-level categories that are ranked along a continuum (Ticehurts & Veal 2000). Likert scale can be ordinal and/or interval (Neuman 2006) and many investigators consider it as being an ordinary interval in character (Hair, Bush & Ortinau 2003; Aaker, Kumar & Day 2004). The advantages of using Likert scale are the variability of scores increasing the spread of variance of responses in providing a stronger measure of relationship, the favourable responses to attitude in exploratory research, and the ease of construction and administration (Malhotra 2004; Burns & Bush 2006). However, the disadvantages of using Likert scale are the tendency for the aggregate total score for respondents to be identical and the length of time to complete the question is longer than the itemised rating scales (Malhotra 2004).
This research study adopted the seven-point Likert scale (as being interval-level measurement) as the measured scale in the survey questionnaire because it is simple to administer and code, offers more options for respondents (with less skewed distribution), and is adaptable to varied statistical analyses (Burns & Bush 2006; Manning & Munro 2006). The constructs in the hypothesised conceptual model were measured with a multiple-item scale due to common practice and usage by other researchers in the innovation management literature (Mole & Worrall 2001; Calantone,
Cavusgil & Zhao 2002; Mahemba & De Bruijn 2003; Hult, Hurley & Knight 2004; Salavou, Baltas & Lioukas 2004; Aragon-Sanchez & Sanchez-Marin 2005; Scozzi, Garavelli & Crowston 2005; Allocca & Kessler 2006; Blumentritt & Danis 2006; Kenny & Reedy 2006; Laforet & Tann 2006; Martensen et al. 2007). Therefore, well-validated measures reported in previous research studies were used when questions and items had to be developed and/or modified, multiple-step and multi-validation methods were to be followed (Churchill 1979).
4.5.5.2 Constructs and Items Determination
The term construct is used by psychologists and the term latent variable is used by social scientists to carry out the connotation of more than abstract ideas and they are specifically defined terms (Creswell 2003). Neuman (2006, p.161) argues that “variables are classified depending on their location in a causal relationship”. There are three latent variables: independent, intervening, and dependent constructs (Cooper & Schindler 2003; Creswell 2003; Sekaran 2003). For example, the independent latent variable, of prior causes, has a cause effect on the dependent variable in a causal hypothesis and is called predictor variable, the intervening latent variable stands between the independent and dependent variables through which their causal relationship operates, and the dependent latent variable is the outcome of the influence of the independent variable and is called predicted variable. It is useful to relate alternatively to constructs as independent, intervening, and dependent latent variables and manifest variables as questions, items, and indicators to the specific hypothesis and on the survey instrument (Creswell 2003).
The exploratory approach is used here to provide further input into the identification of items and latent variables, including the literature review in Sections 2.2, 2.3, 2.4, 3.2, and 3.3; the pilot group discussions comprising 20 individuals of small and medium firms owners/managers, academic researchers, and industry and market experts; and the attendance of key seminars and conferences in the Dubai market and other markets. In the two-pilot group discussions, dialogue regarding the initial survey questionnaire was primarily dominated by the participants, with the researcher contributing only at times when paraphrasing, probing, and promoting were necessary (Blaikie 2000). The process seeks to verify the latent variables and generate more questions and items for the draft survey questionnaire design and layout. Individuals were asked more questions in order to express their opinions and experiences concerning the Dubai SMEs’ innovation practices in terms of macro and micro environmental factors and business growth performance (see Appendix D). Although these viewpoints are confined in Dubai, they are nevertheless useful
in understanding the conditions prevailing and the factors that shape SMEs’ behaviour in similar emerging markets and economies. The outcomes support the latent variables identified in the literature and depicted in the conceptual innovation-based model. These inputs resulted in the assembly of the 11 latent variables and the initial pools of 66 questions and items of the initial survey questionnaire of which 54 were from the literature reviews and 12 were from the pilot group discussions and the selection and formatting of scales that the outcomes of which are the production of the initial draft survey questionnaire. Table 4.3 illustrates the cross-reference of the independent, intervening, and dependent latent variables, items, and the research questions and hypotheses.
Table 4.3: Latent variables, items, and research questions and hypotheses.
Latent Variables and Items Research Questions & Hypotheses
Independent variable 1: Government Supported Developments (5 Items) Descriptive research question 1 & Hypothesis 1A Independent variable 2: Financial Resources (5 Items) Descriptive research question 1 & Hypothesis 1B Independent variable 3: Academia-Industry Collaborations (5 Items) Descriptive research question 1 & Hypothesis 1C Independent variable 4: Market Dynamics (5 Items) Descriptive research question 1 & Hypothesis 1D Independent variable 5: Management Orientation (5 Items) Descriptive research question 2 & Hypothesis 2A Independent variable 6: Organisational Culture (5 Items) Descriptive research question 2 & Hypothesis 2B Independent variable 7: Technology Orientation (5 Items) Descriptive research question 2 & Hypothesis 2C Independent variable 8: Alliance and Cooperation (5 Items) Descriptive research question 2 & Hypothesis 2D Independent variable 9: Market Orientation (5 Items) Descriptive research question 2 & Hypothesis 2E Intervening variable: Innovation Practices (10 Items)
(Independent and/or dependent latent variable)
Descriptive research questions 1, 2, and 3 and Hypotheses 1A, 1B, 1C, 1D, 2A, 2B, 2C, 2D, 2E, and 3 Dependent variable: Business Growth Performance (11 Items) Descriptive research question 3 & Hypothesis 3
Source: Developed for this research with parts adopted from Creswell (2003).
4.5.5.3 Conceptualisation and Operationalisation of Latent Variables
The definition of latent variables in the conceptual model is developed from the literature review of Sections 2.2, 2.3, 2.4, 3.2, and 3.3 along with the pilot group discussions. The conceptual and operational definitions of latent variables are needed before the data is collected and the precise delineating of how latent variables are to be measured and analysed (David & Cosenza 1993; Manning & Munro 2006). It is important to provide clear, specific, and unambiguous definitions
of latent variables for observations and manipulations, which are linked to the proposed research hypotheses and conceptual innovation-based model to avoid any misunderstanding and improve understanding and generalisation in this research study (Gill & Johnson 1991; Cooper & Schindler 2003; Allocca & Kessler 2006; Neuman 2006).
The term conceptualisation demonstrates the process of applying theoretical and abstract sets of meanings to the latent variables in order to specifically explain and define the constructs (Cooper & Emory 1995; Cooper & Schindler 2003). For example, the concept “innovation practices” is conceptually defined as “the determination in terms of the ability of the firm to seek new and better management and administrative systems, internal cultures, processes, products, services, distributing channels, and marketing methods-segments”. Further, the term operationalisation exemplifies the process of defining unobservable latent variables to be measurable in a series of scale questions and items in order to describe the observable characteristics in terms of specific testing criteria (Cooper & Schindler 2003; Sekaran 2003; Hair et al. 2006). It can connect the conceptual definitions to the latent variables measurements and improve the internal validation of the correlations and relations between the independent and dependent latent variables within the proposed conceptual model (Schwab 1999). For example, the concept of “innovation practices” is operationally defined as “the level of agreement with statements in an interval scale about how the firm performs related to management and administrative systems, internal cultures, processes, products, services, distributing channels, and marketing methods-segments innovations”. In this research study, the latent variables and item measurements were selected due to their alignment with the conceptual definitions (Neuman 2006).
4.5.5.3.1 Macro and Micro Innovation Indicators Measurements
The independent latent variables are divided into two main categories, including macro and micro environmental determinants, which are sometimes referred to as environmental and organisational factors, affecting the firm’s innovation practices and business growth performance (Avlonitis & Gounaris 1999). Accordingly, the macro-environmental determinants (external-driven) are factors that can directly affect the firm’s attitude toward innovation, either by stimulating or inhibiting its innovative activities such as government supported developments, financial resources, academia- enterprise collaborations, and market dynamics. Further, the micro-environmental determinants (internal-driven) are factors that can facilitate the firm’s ability to innovate, either by enhancing or inhibiting its innovative behaviours such as management orientation, organisational culture,
technology orientation, alliance and cooperation, and market orientation. These independent latent variables are divided into nine independent latent variables to be further defined conceptually and operationally that include four external-driven and five internal-driven factors of the firm’s ability to innovate and perform in the emerging Dubai market. The conceptual and operational definitions and measurement scales of each construct included in the survey questionnaire are illustrated in Table 4.4. Each construct was evaluated using a five-item scale and was measured with end points of “1-strongly disagree” and “7-strongly agree”.
Table 4.4: Innovation indicators definitions and scales.
Latent Variables Conceptual Definitions Operational Definitions Scales
Government Supported Developments
(gov_sdev)
The determination in terms of the ability of government to establish policies, infrastructure, and institutional support and the degree of importance to the firm’s innovation activities and growth (Teece 1986; Gregersen 1992; Smith 1997; Cooke, Uranga & Etxebarria 1997, 1998b; CEC 2000; Gibbs 2000; Kuhlmann 2003; Haour 2004; Veuglers 2005; Carayannis et al. 2006; Rahl 2008; Sullivan 2008; Lee et al. 2010; Teece 2010; Mani 2011).
The level of agreement with statements in an interval scale about what the firm needs related to policy, infrastructure, and institutional support to encourage innovation activities. It is subdivided into the effectiveness of state government macro and micro policies, quality of overall infrastructure, presence of supportive institutions and development agencies, and existence of commercialisation mechanisms. Arithmetic mean of responses to questions and items 21-25 of the government policy, infrastructure, and institutional support scale. (Question and item 25 is negativity-worded)
Interval
Financial Resources (fin_resrcs)
The determination in terms of the availability of capital and funding support and the degree of importance to the firm’s innovation activities and growth (Teece 1986; Gregersen 1992; Greene & Brown 1997; Cooke, Uranga & Etxebarria 1997; Mishkin 2001; Veuglers 2005; Siems & Ratner 2006; SMEs Conference 2006; Szadkowska 2007; Teece 2010; WEF 2010-2011; Mani 2011).
The level of agreement with statements in an interval scale about what the firm’s needs related to capital resources and funding to encourage innovation activities. It is subdivided into the access to funding schemes and programs, existence of venture capital and funding mechanisms to raise funds and turn commercialise viable ideas into products and services, presence of customised SMEs, financial and technical support to stimulate research and development investment, ability of listing in the local capital market to raise equity capital, and protection of investments and innovators in the local market via financial transparencies and accountability standards. Arithmetic mean of responses to questions and items 26-30 of the financial resources capital and funding scale.
(Question and item 27 is negativity-worded)
Interval
Academia-Industry Collaborations (acdindstr_collbs)
The determination in terms of the partnership between academia and industry and the degree of importance to the firm’s innovation activities and growth (Parker 1992; Keizer, Johannes & Halman 2002; Haour 2004; McAdam, Reid & Gibson 2004; Veuglers 2005; Siems & Ratner 2006; Walters, Kadragic & Walters 2006; Segarra-Blasco & Arauzo- Carod 2008; Wright 2008; Haour & Mieville
The level of agreement with statements in an interval scale about what the firm needs related to talent, transfer technologies, and sourcing ideas to encourage innovation activities. It is subdivided into the access of talent and competencies, outputs of educational institutions related to needed industrial skills, access to local academic institutions and technology centres research capabilities, collaborative research between
2010; Teece 2010). academic institutions and industries to provide technological information and new ideas, and development of entrepreneurial skills and attitudes. Arithmetic mean of responses to questions and items 31-35 of the talents and technology-transfers scale. (Question and item 35 is negativity-worded) Market Dynamics
(mrk_dynmcs)
The determination in terms the interaction and competition of market and the degree of importance to the firm’s innovation activities and growth (Porter 1990; Raider 1998; Mahemba & De Bruijn 2003; Broome 2007; Gao, Zhou & Yim 2007; Shediac et al. 2008; Teece 2010; WEF 2010-2011; Bao, Chen & Zhou 2011; Martinez-Roman, Gamero & Tamayo 2011; Zhu, Wittmann & Peng 2011).
The level of agreement with statements in an interval scale about what the firm needs related to market interaction and competition to encourage innovation activities. It is subdivided into the presence of suppliers and supportive industries to access raw materials and components, efficiency of the local market by understanding market demand conditions and consumer orientations, effectiveness of anti-monopoly policy and healthy market competition, presence of risk-taking business climate and culture, and presence of an exit mechanism in the local market. Arithmetic mean of responses to questions and items 36-40 of the interactions and competitions scale.
(Question and item 39 is negativity-worded)
Interval
Management Orientation (mgmt_orint)
The determination in terms of the extent of management characteristic and strategic directions and the degree of agreement to the firm’s innovation activities and growth (Kets de Vries 1977; Miles & Snow 1978; Snow & Hrebiniak 1980; Smith, Guthrie & Chen 1986; Lampikoski & Emden 1996; Hoffman et al. 1998; Motwani et al. 1999; Aragon-Sanchez & Sanchez-Marin 2005; Blumentritt & Danis 2006; Wheeler, McFarland & Kleiner 2007).
The level of agreement with statements in an interval scale about how the firm directs related to management orientation to encourage innovation activities. It is subdivided into the importance of innovation to achieve strategic goals and ambitions, focus on long-term goals and objectives, favour of high-risk projects with aggressive posture to explore new potentials, commitment and involvement in development of new initiatives and programs, and allocation of resources to support and sustain innovation programs. Arithmetic mean of responses to questions and items 41-45 of the characteristic and strategic orientation scale.
Interval
Organisational Culture (org_cltr)
The determination in terms of the extent of learning processes, designs, and flexible practices culture and the degree of importance