Introduction

In an increasing number of industries and business activities, technological innovation is viewed as one of the cornerstones of firms' competitive advantage and profitability (Subramaniam & Youndt, 2005; Galende, 2006). Technological innovation leads to product and process improvements that increase firms' sales, and/or reduce firm costs, helping firm survival and ultimately making them more profitable than non-innovators Thus, new product-development strategy is a primary determinant of the firm's performance (de Bentrani et al, 2010).

Management academics, and especially fruitfully those from the Resource-Based View (RBV) (Barney, 1991; Barney et al, 2011), as well as other derived frameworks such as the Intellectual Capital-Based View (ICBV) (Bontis, 1996, Reed et al, 2006) or Knowledge-Based View (Grant, 1996; Johnson et al, 2002; Díaz-Díaz et al, 2008; Bueno et al, 2010), highlight the strategic role of different types of intangible resources, capabilities, intellectual and knowledge assets on firms' technological innovation, and competitive advantage.

From an ICBV (Reed et al, 2006; Martín-de Castro et al, 2011), the work of Subramaniam & Youndt (2005) remarked that an organization's capability to innovate is closely tied to its intellectual capital endowment, and although the basic link between knowledge assets and innovation is on the whole persuasive these authors argue for further research, because it is necessary to develop investigations in order to understand about its precise and complex nature.

To empirically test the basic link of firms' intellectual/knowledge endowments and technological innovation, two main perspectives have been majority adopted. The first one tries to determine the role of a resource/capability, or a set of them, in order to achieve and sustain competitive advantages, superior performance, or, in our case, technological innovation. From the RBV, the theoretical proposal was made by Barney (1991) with his well-known VRIO model, which is the origin of a multitude of empirical studies (Kraaijenbrink et al, 2010). The research aim inside this perspective is to determine the better resources/capabilities and so on that lead firms to success, independently of any context and/or circumstance.

The second research trend adopts a deeper and contingent approach, admitting that the value and strategic role of a specific resource/capability depends on the context (internal and/or external to the firm), or the existence of other variables affecting and explaining the basic link between resource/capability and technological innovation. In this sense, from an ICBV, the study of Subramaniam & Youndt (2005) highlights the moderating effects of complementary intellectual assets on the relationship between intellectual capital and technological innovation. In the same way, during the last decade, numerous studies have been developed in order to explore those indirect effects: mediating and moderating effects (Díaz-Díaz et al, 2008), proposing that the real sources of firm superior performance – financial, market, and technological – as well as sustained competitive advantage are complex phenomena that need the exploration and analysis of business context.

Nevertheless, the relationships between resources, capabilities, knowledge and intellectual assets, and technological innovation are complex in nature (Galende, 2006), which require a new approach in order to consider them as firms' complex configurations that produce several technological paths and firm success. This argument has been developed by Ragin (2000, 2008) and more recently by Schneider et al (2011), Crilly (2011), and Fiss (2011) in which it can be labelled as a ‘configurational approach’, which represents the new framework of the present study.

A configurational approach assumes the premise that causation is not easily unravelled because the outcomes of interest rarely have any single cause, causes rarely operate in isolation from each other, and a specific causal attribute may have different and even opposite effects depending on context (Greckhamer et al, 2008). This approach suggests that equifinality may exist, that is, different configurations would lead to the same outputs.

Configurational approach and its technique – fs/QCA – has been successfully used in political and sociological studies, but its application in Management and Strategy issues is in an emerging state (Fiss, 2011; Schneider et al, 2011).

Nowadays, very little empirically driven research exists about the configuration of bundles of resources and capabilities (Gruber et al, 2010) and intellectual capital/knowledge assets in achieving different types of firm technological innovations.

In that sense – both theoretically and empirically – this research adopts a new configurational perspective (Ragin, 2000; Fiss, 2011), where the integrity of firms' technological innovations as complex configurations of causal factors is preserved. This way, using a new statistical technique – Qualitative Comparative Analysis (QCA) – and primary data of 251 technology-based firms based in Spain, this paper explores the firms' intellectual capital configurations and technological innovation. Results included in the ‘truth matrix’ show how different intellectual capital configurations (human capital, structural capital, relational capital) lead to high- and low-average technological innovation.

Theory development

Human, technological, and relational assets: Exploring firm's intellectual capital

Intellectual capital is gaining acceptance as a key concept in the management agenda (CIC, 2003; Reed et al, 2006; Martín-de Castro et al, 2011), and although there is no consensus in its definition and main elements some of the proposals converge at the same point, defining intellectual capital as ‘the combination of intangible assets that allows the company to operate’ (Brooking, 1996, p. 25), or as the sum of all knowledge possessed by the employees of an organization that confer it with a competitive advantage. In general terms, and following Bontis (1996), CIC (2003), and Subramaniam & Youndt's (2005) proposals, intellectual capital refers to the sum of all knowledge stocks that firms utilize for competitive advantage-creating value, which represents distinctive knowledge stocks accumulated and distributed through individuals, relationships among individuals, and the organization itself.

In general terms, it has been recognized that economic wealth comes from knowledge assets – intellectual capital – and their use in application, replacing, or perhaps supplementing land, labour, and capital (Dean & Kretschmer, 2007). The term ‘intellectual capital’ could be used as a synonym for knowledge assets (Stewart, 1991). On the basis of the proposal by Edvinsson & Malone (1997), intellectual capital is a two-level construct: human capital (the knowledge created by and stored in a firm's employees) and structural capital (the embodiment, empowerment, and supportive infrastructure of human capital).

Although there is no agreement among scholars about the concept, components, and variables of intellectual capital (CIC, 2003; Dean & Kretschmer, 2007), as a basic level, the conceptual separation of these IC aspects is related to how each one accumulates and distributes knowledge and information differently, either through (i) individuals; (ii) organizational structures, technological knowledge assets, and IT systems; and (iii) relationships and networks. In a wider sense, these issues represent different expressions of intangible resources and knowledge within a firm.

Following the study by Subramaniam & Youndt (2005), human capital refers to the knowledge, skills, and abilities residing within and utilized by individuals. Structural capital (Kang & Snell, 2009) describes the institutionalized knowledge acquired and captured inside the firm and its organizational members and stored in organizational processes, systems, patents, and R&D efforts. Inside structural capital, CIC (2003) and Martín-de Castro et al (2011) split the institutional knowledge into organizational and technological capital. Owing to the objective of our research, this investigation focuses its interest on technological capital or technological knowledge assets, which include the company's R&D efforts as well as knowledge storage and utilization. Relational capital (CIC, 2003; Martínez-Torres, 2006) refers to the useful knowledge and information gathered by the firm in its relationships with its customers, suppliers, allies, competitors, public administrations, society in general and so on. Again, and owing to the importance given to customer relationships in the development of product innovations (Cabrita & Bontis, 2008; Sánchez-González et al, 2009), this investigation focuses its interest on the ‘customer relational capital’.

Firms' product innovation

Firm innovation is a primary instrument of competition for many firms, especially within technology and knowledge-based industries. Generally speaking, the innovation process can be understood as a complex activity in which new knowledge is applied for commercial ends (Nieto, 2001; Galende, 2006; Escribano et al, 2009), creating new products, services, or processes.

Generally speaking, the innovation process can be understood as a complex activity in which new knowledge is applied for commercial ends (Galende, 2006; Escribano et al, 2009). Innovation is viewed as one of the most important sources of sustainable competitive advantage because it leads to product improvements that increase the value of the product portfolio (Coombs & Bierly, 2006); helps firms survive; makes continuous advances (Liu et al, 2005); and allows innovators to grow faster, being more (dynamically) efficient, and ultimately more profitable than non-innovators (Mansury & Love, 2008).

Literature (Tushman & Nadler, 1986; Hill & Rothaermel, 2003; Stieglitz & Heine, 2007, among others) recognizes a wide range of innovation types within the firm (product/process, radical/incremental, technological/managerial, market pull/technology push, or competence-enhancing/competence-destroying). This research will use this typology that classifies innovation types according to the results or outputs of the innovation process, focusing specifically on product innovation, since it is one of the most promising areas in the field of Knowledge Management (Corso et al, 2003). This way, according to Nieto (2001), when new technological knowledge is materialized in the development of new products or in the improvement of those already existing, we are talking about a product innovation.

Linking firms' intellectual capital and product innovation

Once key variables of the research have been presented, from a configurational approach, we develop theoretical arguments linking firms' intellectual capital and technological innovation in a general way, exploring different possible intellectual capital configurations – human, technological, and relational assets – and their technological innovation and permitting the existence of equifinality. Following this logic, no explicit hypotheses are presented.

Thus, having brilliant, motivated, and experienced human capital should be the base for all innovation processes in a firm. This kind of intellectual capital provides the main source for developing new ideas and knowledge (Snell & Dean, 1992). Individuals and their associated human capital assets are crucial for exposing an organization to technology boundaries that increase its capacity to absorb and deploy new different knowledge domains (Subramaniam & Youndt, 2005), being linked to product innovations. Highly motivated and trained employees may question the established organizational routines; hence, this kind of human capital becomes critical to push the firm to its technological borders, constituting the best incentive towards obtaining new knowledge and innovation (Nonaka & Takeuchi, 1995; Hill & Rothaermel, 2003; Zhou & Li, 2012).

Furthermore, the organization can accumulate, codify, and store individual and collective technological knowledge in databases, proceedings, and organizational structures. Taking into account the nature of this collective and structured knowledge, and focusing on technological innovation we could talk about technological capital or technological knowledge assets (Díaz-Díaz et al, 2008). Both types of intellectual capital – human and technological – could be an effective base for firms' product innovations (Galende, 2006). In that sense, Zahra & George (2002) affirm that knowledge depth – based on technological capital or knowledge assets – is key to develop new ideas and technological product innovations.

Finally, from a relational view or a collaborative or open innovation perspective (Chesbrough, 2003), the external relationships maintained by a firm with the different agents of its competitive environment (mainly customers) constitute a good source of information and knowledge gathering for the firm's product innovation (Tseng & Goo, 2005). In fact, some researchers claim that most advances and new sources of innovation are created within networks, with them being the locus of it in inter-organizational networks rather than in isolated firms (Wincent et al, 2010). Gaining innovation outputs from external sources has been recognized as a key mechanism to enhance radical innovations (Chesbrough, 2003). For the case of new product innovations, relationships, with users involved at the prototype stage to gain market first evaluations, lead to firm potential benefits in the development of new products and services (Lettl, 2007; Sánchez-González et al, 2009).

Nevertheless, as we argue in the main thesis of this work, product innovation is a very complex organizational phenomenon (Galende, 2006), which needs additional efforts to understand its precise nature and sources. Beyond basic direct links between intellectual assets and product innovation, or taking into account possible contingencies, as mediating or moderating effects of other organizational variables or among intellectual capital assets, as we have previously remarked, more efforts are needed in order to analyse the links between intellectual capital and technological innovation from a configurational approach, such as the one by Morris & Snell (2011), exploring different intellectual capital configurations and their radical/incremental technological innovation links. In the following sections, we develop a field research adopting the latter framework to analyse the real complexities inherent in the innovation phenomenon.

Method

Research design, population, and sample

Following methodological recommendations of Rouse & Daellenbach (1999), King & Zeithaml (2003), and Newbert (2007, 2008), RBV as well as its start up, ICBV, requires the careful selection of a homogeneous and adequate population and sample, using in the majority of cases primary data sourcing because of the non-existence of databases about internal organizational factors. According to Leitner (2005), high and medium-high technology industries are strongly focused on intangible factors and knowledge, being an appropriate setting for developing the field research. In addition, we selected firms belonging to a homogeneous industry in order to avoid different effects derived from environmental factors. Finally, we focused on firms with 50 or more employees as they have an appropriate size to investigate interesting external – supplier and customer relationships – and internal ties that lead to achieving innovations.

Thus, in order to collect data, we designed a questionnaire with items for each of the IC variables – human, technological, and customer relational – and the product innovation variable used in the study (7-point Likert scale, see Appendix), as well as firm age and size as control variables. The questionnaire was administered to senior managers during the first 3 months of 2009. It was designed to investigate valuable and unique competences of the firm – intellectual capital – as they can lead to a firm's success (King & Zeithaml, 2003): secondary sources do not provide enough information about those aspects. At this point, it is important to highlight that the respondents to the questionnaire were senior managers who have more information around relationships at the institutional level. The questionnaire was piloted with 10 companies that met the selection criteria, which enabled checks to be made that questions were clear and respondents were able to complete sections on diverse activities across their organization. This questionnaire was administered during the first quarter of 2009 as a telephone survey, which also allowed any clarification to be offered by the interviewer.

From a population of 1270 firms collected in the SABI database, the sample consisted of 251 high and medium-high technology Spanish manufacturers with 50 or more employees, which represents a response rate of 19.8%. Although the survey's response rate is relatively low, as Fiss (2011) remarks, a 10–12% response rate is typical for surveys mailed to CEOs in countries such as the United States, and better than his response rate (14%) reached in high-technology manufacturing firms located in the United Kingdom.

Measures: Constructing the property space

Regarding three intellectual capital assets (see Appendix), human capital makes reference to the knowledge – explicit and tacit – that people possess, as well as their ability to generate it, which is useful for the mission of the organization (CIC, 2003). Edvinsson & Malone (1997) remark that human capital includes knowledge, skills, innovativeness, and the ability to meet the task at hand. This capital may be taken away by employees, and includes employees' and managers’ competence, experience, knowledge, skills (Martínez-Torres, 2006), attitude, commitment, and wisdom (Hsu & Fang, 2009). Cabrita & Bontis (2008) mention individuals' education, skills, and experiences.

In that sense, human capital (hc) was assessed by averaging three items referred to as education and training (based on Snell & Dean, 1992; Zárraga & Bonache, 2005; Wu et al, 2008; Cabrita & Bontis, 2008), experience and abilities, and their creative and innovative character (based on CIC, 2003; Subramaniam & Youndt, 2005; Reed et al, 2006, among others).

Technological capital, as CIC (2003) points out, refers to the combination of organizational knowledge directly linked to the development of the activities and functions of the operations technical system, responsible for obtaining new products and services. It includes efforts in research and development, technological infrastructure, and intellectual and industrial property. On the basis of a literature review, technological capital (tc) was measured by averaging three items related to R&D efforts and the use of the knowledge base (based on Cohen & Levinthal, 1990; Tsai, 2001; Greve, 2003; Nieto & Quevedo, 2005; Subramaniam & Youndt, 2005).

Customer relational capital or customer capital (Kaplan & Norton, 1992) includes knowledge and networks developed in the interaction between organizational employees and its customers, and represents the potential an organization has because of ex-firm intangibles (Wu et al, 2008). On the basis of previous literature, customer relational capital (crc) was assessed by averaging three items related to external ties with customers: customer relationships as key informants for product innovation, joint work between the company and its customers to develop solutions, and the quality of the firm's customer base (Chen et al, 2004; Subramaniam & Youndt, 2005; Reed et al, 2006;Cabrita & Bontis, 2008; Hsu & Fang, 2009).

Product innovation (PI) was assessed by three items considering the total number of product innovations developed by the firm, the percentage of sales derived from new products, and the number of new products with respect to the product portfolio (Souitaris, 2002; Akgün et al, 2007; or Wu et al, 2008).

For the three IC variables and product innovation variable, an Exploratory Factor Analysis was developed including all items and showing good fit (KMO>0.8 and the resulting four factors corresponding with each of the variables: hc, tc, crc, PI, and an explained variance of 73.5% and factor loadings over 0.7), their reliability was considered as acceptable and good (with Cronbach's α=0.68 for human capital, and over 0.75 for technological and customer relational capital), and given the relatively normal distribution of these variables we used the average of them as the breakpoint to determine a set membership in each of the variables (human capital=1 for firms with human capital above average (5.53), 0 for firms with human capital below average; technological capital=1 for firms with technological capital above average (4.18), and 0 for firms with technological capital below average; and customer relational capital=1 for firms with customer relational capital above average (5.55), and 0 for firms with customer relational capital below average).

With respect to the dependent variable (product innovation), we calculated the mean of the three items and we also used the mean of the distribution as the breakpoint to assess the set membership: those corporations above the mean (4.64) were coded as being in the set of corporations with high product innovation (c.slack =1), whereas those below the mean were coded as not being members in this set (c.slack=0).

Finally, owing to the fact that a firm's size and age may influence the innovation development carried out by a firm, they are considered as control variables (based on Tsai, 2001; Reed et al, 2006, among others). Firm size (fsize) was measured by the mean number of employees who belong to the firm, and because of the highly skewed nature of the distribution we used the median of the distribution as the breakpoint to assess the set membership: those corporations above the median (80 employees) were coded as being in the set of large corporations (c.slack =1), whereas those below the median were coded as not being members in this set (c.slack=0). Firm age (fage) was measured by the number of years since its constitution, and given the relatively normal distribution of this variable we used the average of it as the breakpoint to determine set membership in the set of variables (firm age=1 for firms with an age above average (25–58 years), 0 for firms with an age below average.

Fuzzy set/QCA

QCA is gaining acceptance as a rigorous statistical technique especially well suited to investigate configurations (Fiss, 2011; Ragin, 2000). This technique is used because of several reasons: (i) it enables us to test the propositions regarding the influence of intellectual capital configurations as complex sets of firms' attributes on technological innovation outputs (Ragin, 2000; Greckhamer et al, 2008), that is, the idea that a combination of various intellectual capital assets as causal conditions, rather than one variable alone, are linked to the firm performance; in our case, innovation outputs; (ii) QCA has the advantage of being suited to small sample sizes – with less than 300 cases – and to limited diversity and (iii) as Fiss (2011) highlights, some promising applications of QCA to management and strategy research include the RBV framework, and parallel developments as an ICBV.

In fsQCA (Schneider et al, 2011), a case is described by the combination of ‘causal conditions’ and the ‘outcome’. All data are calibrated into set membership values ranging from 0 to 1, and these boundaries constitute the ideal type values of a variable. The values of most cases will not meet the ideal types, and thus a third boundary, the crossover point (0.5), defines the anchor for a qualitative distinction between being ‘in’ or ‘out’ of a set. In essence, fsQCA explores how the membership of cases in causal conditions is linked to membership of the outcome.

Results

The property space constituted by the research attributes can be developed as Table 1 shows. It reports all the logically possible combinations of causal attributes. This Boolean space is constituted by 2n configurations, where n is the number of causal attributes included. In our case, 25=32 configurations.

Table 1 Truth table and assignment of firms to ideal type of intellectual capital and product innovation configurations
Full size table

Following the study by Schneider et al (2011), we can continue by testing whether any of the causal conditions can be considered a necessary condition for the outcome. A causal condition is called ‘necessary’ if the instances of the outcome constitute a subset of the instances of the causal condition (Ragin, 2008). A consistency score of 1 indicates that the combination of causal conditions meets the rule across all cases. The more cases fail to meet the consistency criterion and the larger the distance from meeting the criterion, the further the consistency score will fall below 1. Conventionally, a condition, or a combination of conditions, is called ‘necessary’ or ‘almost always necessary’ if the consistency score exceeds the threshold of 0.9. We analysed whether any of the five causal conditions and their negations are necessary to account for product innovations.

In the main thesis proposed in this research, we argue that intellectual capital endowments lead to product technological innovations.

The first step in analysing sufficient conditions is the creation of ideal types by converting the set membership values for the causal conditions into crisp-set values. A case is assigned a value of 1 if the fuzzy-set membership value exceeds the threshold value of 0.5, and a value of 0 in all other cases. The resulting ideal types are reported in the Table 1.

Causal combinations of conditions exceeding an appropriate cut-off consistency score are categorized as sufficient, and therefore the outcome is assigned a value of 1 in the truth table. Conversely, causal combinations with a consistency level below or at the cut-off value are not considered sufficient, and the outcome is assigned a value of 0. Applying a cut-off value of 0.75 as the minimum recommended threshold (Fiss, 2011) for our data yields the combinations of causal conditions and outcome reported in Table 1. We observe all 32 logically possible causal combinations for each type of innovation.

In the case of product innovation, 86 cases in configurations with seven or more cases in the configuration exceeded the minimum consistency threshold for high product innovation outputs.

In order to examine the causal conditions that are sufficient for the outcome of interest (high product innovation in that case), the QCA approach uses Boolean algebra to logically reduce the truth table rows to simplified combinations (Fiss, 2011). As Greckhamer et al (2008) remark, although necessity and sufficiency should be investigated, it is recommended to focus on the sufficiency of combinations of causal conditions. In assessing causal sufficiency, QCA uses the concept of quasi-sufficiency, based on several benchmarks (>0.80 always sufficient; >0.65 usually sufficient) in causing the desired output.

Table 2 shows the following analysis by utilizing a Boolean algorithm in order to assess whether any causal conditions – human capital, technological capital, customer relational capital, firm size, or firm age – are sufficient to cause the high product innovation. Basic Boolean operators introduced are: logical and (represented by *) represents the intersection of sets. Logical or (represented by +) represents the union of two sets. Finally, logical not (represented by ∼). For example, configuration ID 1 in Table 1 can be written in equation form as:

Table 2 Parsimonious solution for high product innovation performance
Full size table

Using Boolean algebra, for example, configurations ID1 and ID 5 can be found to be sufficient for achieving high product innovation performance, and we can represent them as follows (where → denotes Boolean implication):

Although the number of logically possible groupings of all Boolean attributes included in the study follows the formula 3k−1, where k is the number of attributes (in our case 35−1=242), Table 2 reports the final Boolean expressions obtained from the truth table analysis as the parsimonious solution.

The general interpretation of data reported in Table 2 represents the causal conditions that are usually sufficient for high product innovations performance. As a first important result, we can see that there are three paths to a firm's product innovation success, permitting equifinality, that is, there is not a unique and better way to obtain firm success.

Second, and using Boolean equations, we can appreciate in Table 2 the most raw coverage (54% of cases analysed) of one of the three different configurations of intellectual capital assets, firm size and age, and high product innovation performance denotes high endowments of human capital, and customer relational capital. This key result shows that highly creative, experienced, and skilled employees jointly with well structured networks with company's customers are the key sources in obtaining high degrees of product innovations in technology-based industries.

Third, in older technology-based firms, technological assets could substitute high endowments of human capital, because of the institutionalized and documented knowledge useful for the organization and accumulated through a long period of time. This way, as Paths 2 and 3 show, technological capital acts as a remedy for non-existence or low endowments of the other intellectual capital assets.

Nevertheless, in order to obtain better product innovation performance, firms need to closely interact with their customers, corroborating a ‘relational view’, a ‘collaborative approach’, or ‘an open view’ of firm's product innovation (Chesbrough, 2003).

Results show the important interdependences among intellectual capital components, as well as firm age, as the conditions are usually sufficient for superior product innovations in technology-based manufacturing industries. No empirical evidence has been obtained for the role of firm size.

Conclusions and limitations

Our paper explores product innovation phenomena in the firm from both internal and external points of view. In this vein, empirical research links human, technological, and customer assets to product innovation.

This paper offers value added to literature debate in the intellectual capital and innovation literatures (Subramaniam & Youndt, 2005) because of its configurational approach. In that sense, this proposal advances from a contextual approach towards one that allows taking into account complexity and equifinality.

To explain firms' product innovation phenomenon in a complex way (Galende, 2006) both internally and externally, treating the independent variables (human, technological, relational customer) jointly with control variables (firm age and size) as complex and interdependent configurations of variables responsible for whole firm product innovation. This way, results show configurations of organizational attributes that jointly explain product innovation in the case of high-tech manufacturing firms.

Equifinality means that there could be more than one configuration explaining better product innovation results. In this investigation, the main path developing high product innovation outputs, with a unique coverage of 0.38, implies the joint development of internal and external, individual, and collective knowledge assets, represented by high endowments of human capital and customer relational capital. This path is coherent with previous studies, by Snell & Dean (1992), Subramaniam & Youndt (2005), or Zhou & Li (2012), who stated the key role of human capital assets in order to expose the organization to diverse knowledge, develop ideas, and, consequently, to develop product innovations. Customer relational capital creates opportunities and increases the motivation of organizational employees to seek and develop better product innovations jointly with customers, as Adler & Kwon (2002) and Zahra & George (2002) pointed out. The other two successful product innovation paths have certain substitutive effects depending on the organizational age. Thus, young firms need the development of high endowments of technological knowledge assets in order to effectively develop high product innovation outputs.

Finally, considering the discussion of research results, the three paths show the necessity to develop customer relational capital in order to reach better product innovation results, with our results being in concordance with an extensive and recent body of innovation literature considering it as ‘open innovation’ or ‘collaborative innovation with customers’ (Chesbrough, 2003; Sánchez-González et al, 2009).

Among the research limitations, this study is a qualitative–quantitative one, being in an early stage of development in the field of strategy and innovation studies. Furthermore, the study is cross-sectional, which impedes searching for better causal relationships. Finally, because of the use of a questionnaire for gathering data, this study suffers from a well-known series of issues. The results and conclusions should be read taking into account these research limitations.