Various studies seek to characterize which factors have the greatest impact on the treatment of innovation. And among those found in the literature, the environment stands out. However, despite numerous discussions about the influence of the environment on innovation (DAMANPOUR, 1996; TIDD, 2001; MANUAL de OSLO, 2005; ZHANG; MAJID; FOO, 2011; TSUJA; MARIÑO, 2013), very little about this influence has been effectively portrayed in scientific studies (FAGERBERG, 2004; RIBEIRO; CHEROBIM, 2017a).

The restrictive and deterministic character of the environment is evidenced from the conception of the contingency theory, which has a strong influence on organizations and their strategies for achieving survival. However, we observe in the literature that the influence of the environmental context is generally treated in a generic way, without characterizing its specificities. The various characteristics attributed to the environment that make it difficult to address are commonly observed. These include complexity, dynamics, uncertainty, unpredictability and volatility (RIBEIRO; CHEROBIM, 2017a). Although these adjectives are very pertinent, we note that they are comprehensive, generalist and linked to any environment.

Regardless of the type of environment, organizations are faced with the challenge of knowing and understanding environmental contingencies so that they can create and leverage innovation (LI; ATUAHENE-GIMA, 2001; TIDD, 2001), understood as the successful application and exploitation of an idea, even if the novelty is only at the local level (MANUAL DE OSLO, 2005). However, it is difficult to have an effective conception of the environment without knowing its parts, characterized by environmental variables and configurations. Thus, in this article we seek to answer the following research question: do environmental variables assume different values over time and conform differently, providing different environmental configurations?

By characterizing the dynamicity of variables and the possibilities of environmental configurations, understanding and managing the interaction of the environment and innovation becomes less complex. This paper aims to demonstrate that environmental variables assume differentiated values over a given period and correlate in a way to configure different contexts. Knowing the possible configurations of the environment is important in order to exploit innovation in competitive organizational strategy.

Thus, we justify this study by seeking to present empirical elements to reduce some of the gaps with practical application results, these include: 1) the absence of studies effectively considering the relationship of the environment to the innovation process; and 2) the lack of evidence that environmental variables or factors can correlate in order to shape certain configurations that impact innovation and require more adequate indicators for their measurement. We use data from environmental variables related to Brazil and the states of São Paulo, Paraná and Sergipe, which have differentiated degrees of innovation.

In an objective way, environment is all that we find outside a system, which is conceived as a set of interacting and interdependent parts that relate to a common goal (BERTALANFFY, 1968). The environment concerns the various social and physical factors that influence the organizational decision-making process and that are beyond the limits of the organization (McGEE; SAWYERR, 2003).

Tsuja and Mariño (2013) define environment as a set of external factors that interact causing reflections in the organization. These factors are characterized by uncertainty and complexity. For these authors uncertain environment is where frequent changes occur in the external variables that impact the organization. However, the complex environment is characterized by covering a large number of external variables that influence the organization.

Another characteristic of the environment, according to Duncan (1972), is the dynamicity, described as the speed of changes in environmental variables in a given period of time. As environmental conditions change rapidly and constantly, another typical feature of the environment emerges: volatility. Therefore, uncertainty, complexity, dynamicity and volatility are the main characteristics of the organizational environment.

The environment impacts the organizational dynamics and strategy (DAMANPOUR, 1996; TIDD, 2001), and understanding it with its attributes is fundamental for the organization to remain in the market (MOYSÉS FILHO et al., 2010). Therefore, their strategies must be thought and conducted considering the necessity of the homeostasis, so that the organizational survival is guaranteed. And one of the strategies in this sense is the environmental mapping (HAMBRICK, 1982; ZHANG; MAJID; FOO, 2011), which provides the information necessary for the actions of organizations.

In addition to environmental mapping, we need to know and understand the dimensions of the environment: 1) microenvironment: task or domain environment; and 2) macroenvironment: remote environment. The first type is significant and has a direct impact on the organization's tasks and results and includes consumers, suppliers, competitors and other stakeholders. The other type, which is also important, has an indirect and long-term impact, including economic, educational, social, cultural, technological and legal variables (MANUAL DE OSLO, 2005; MOYSÉS FILHO et al., 2010; MYBURGH, 2004).

Despite the various models of innovation, its concept has always revolved around the application of knowledge that adds value to something. For Schumpeter (1939) innovation is a function based on creative thinking and action, where products and consumption habits are replaced by new ones; innovation is everything that differentiates and gives value to a business. With the Oslo Manual (2005), the concept of innovation was expanded, characterized as the implementation of a new or significantly improved product, a process, a new marketing method, or the implementation of a new organizational method in business practices, in the workplace organization or in external relations.

It is true that the creative action of innovation gives the organization the ability to produce new products and services. In meeting the needs of the market, innovation expands organizational competitiveness. This is the reason for the increase in the number of studies on innovation and its impact on productivity and competitiveness (SANTOS; VASCONCELOS; DE LUCA, 2013).

To better understand innovation, it is important to situate it in the context where it occurs and its scope. In this sense, innovation presents differentiated characteristics if analyzed at company, regional or national level (MANKIW, 2003). This segmentation has a consistent impact on internal and external factors (environmental context) related to the creation, application and diffusion of innovation, such as human resources (CASSIOLATO; LASTRES, 2000), markets (BARNEY, 1991; PORTER, 2008), institutional conditions (SCHUMPETER, 1939), political and economic aspects (SILVA; DA SILVA; MOTTA, 2012).

The literature presents several types of innovation (RIBEIRO; CHEROBIM, 2017b), but the radical and incremental types proposed by Schumpeter (1939), still predominate on a consolidated basis. Regardless of type, it is paramount to identify secure ways of measuring innovation. Although there is no universal tool for measuring innovation and competitiveness (FREEMAN; SOETE, 2007), a positive way of reducing uncertainties as measurement goes through the understanding of how the environment influences them. In this aspect, the understanding of the behavior of the environmental variables and the way of conforming are presented as basilar for the management of innovation.

We use the exploratory and descriptive research strategy in this study. To make them operational, we conduct a bibliographical research, through the EBSCO host platform database and the journal portal of the Coordination for the Improvement of Higher Education Personnel - CAPES, and bibliometric analyzes, using the public databases described in Table 1.

We conduct an empirical research through the secondary data analysis, characterizing as the analysis of data previously collected and tabulated by other sources (BHATTACHERJEE, 2012). We collected the data from the research's focus areas: Brazil, São Paulo, Paraná and Sergipe. The choice of the Brazil unit is due to the globalized approach of the variables. We selected the states based on their general characteristics of innovation, reflected in the competitiveness (EXAME, 2015), and the accessibility to the statistical data of the environmental variables.

Regarding the diversity of environmental variables (MANUAL DE OSLO, 2005; MOYSÉS FILHO et al., 2010; MYBURGH, 2004), these can be synthesized in: economic, educational (including technological), social (encompassing cultural) and political (which encompass legal variables). In this research, we disregard the political variables for two reasons: 1) in the period of the study there was no alternation of the political group that commanded the country; and 2) the political variables do not show constancy of publication and reliability. Therefore, the environmental variables we considered are: economic, educational and social.

In the literature review we verify the absence of studies dealing with the relation between the environmental configuration and innovation and its indicators regarding the reflexes in the measurement process. In order to achieve the objective of this study, we establish two research hypotheses:

H1 - The environmental variables - economic, educational and social - can assume different values over a given period ranging from extremely positive positions to extremely negative positions and can correlate with each other; and

H2 - The correlation of the economic, educational and social variables, in their different forms (ranging from positive to negative), results in eight distinct environmental configurations.

We selected the indicators used to measure the value of each of the environmental variables from public databases. The database were built with annualized and available data. The result was comprised of 15 indicators available for each of the studied variables.

For the analysis of the H1 hypothesis, data were collected on the indicators referring to the variables in the Brazilian scope, allowing for a broad and generic analysis. For the H2 hypothesis, the data were collected deal with the indicators of the states selected for the research. The variables and indicators are summarized in Table 2.

Considering that this research deals with innovation and that the first research on the innovative process in Brazil occurred in the triennium 1998-2000, and the fact that this period was greatly influenced by the crisis of the Asian tigers (1997) an the Russian crisis (1998), with negative impact in Brazil, we excluded data from this period, as possible biases resulting from these crises could hardly be neutralized. Thus, we adjusted the period of data collection between 2001 and 2013. We established the upper limit based on the complete availability of data.

To facilitate the identification of the variable, we created a structure to comprise the data. Each variable received the designation of construct, represented by the letter "C", and named each indicator as variable, represented by the letter "V". Thus, the economic variable termed as C1 and its first indicator of V1, so that the control code is V1C1 (variable 1 of construct 1). For the educational variable we assigned the designation C2 and for the social C3. We adopted the same procedure for these variables regarding their indicators (V1C2; V1C3). Due to the fact that there are several variables/indicators over the years, we used the technique known as "time series analysis" characterized by the set of observations over a period (STEVENSON, 1981).

In order to perform the calculations demanded in this research, we use the SPSS software (IBM/SPSS, 2012) and the GRETL (GRETL, 2013). Once the environmental variables were defined, with the respective measurement indicators, we identified some the database some missing values. We use an SPSS procedure that allows us to enter missing values.

Because the database is presented as a panel, the literature points out that the main problems are related to normality (cross section) and whether or not the time series is stationary (stationarity). Regarding normality, we use secondary data only, they were already treated (normalized) at the time of the original research. In addition, the normality test is not recommended for small samples, which is the case of this research (2001-2013), depending on the loss of potency. Therefore, we calculated the Spearman correlation coefficient. As for stationarity, to avoid working with spurious correlations, we test the series using KPSS and stationarity by applying the first difference method, both in the GRETL application.

We verified that the use of correlation was responsible for standardizing the data under analysis, which is the process of transforming the standard score into Z-score (LEVINE et al., 2015). Therefore, the database was ready for the statistical calculations and analyzes resulting from the adaptation of the data contained in the preliminary spreadsheet migrated to the final spreadsheet, completing the missing value and carrying out the stationarity tests of the time series.

We used secondary data for each indicator per variable, expressed in Table 2, and we elaborated specific tables for each research site - Brazil, São Paulo, Paraná and Sergipe. Based on these tables, we ran the correlations through the SPSS application (IBM/SPSS, 2012), resulting in four data files with correlations for the sites surveyed considering the economic, educational and social environmental variables.

Because the series of the research is short, the literature recommends the use of Spearman's correlation coefficient, whose reading of the correlation hypothesis is related to the following conditions of comparison between null hypothesis (H0) and alternative hypothesis (H1): a) H0: There is no correlation between the pair of variables; the correlation between them is null; and b) H1: There is correlation. Thus, if the p-value is lower than 0.05, the H0 hypothesis must be rejected.

The correlation coefficient can range from -1 to +1 (minus one to one), indicating whether it is directly or indirectly related. It should also be pointed out that the SPSS application (IBM/SPSS, 2012) highlights the significant "correlations" (marked with an asterisk (*)) and the "very significant", (where two asterisks appear (**)).

Effectively the environmental characteristics present reflexes in the level of innovation and development in relation to the place in which it is analyzed, according to Damanpour (1996); Tidd (2001); Oslo Manual (2005); Zhang, Majid and Foo (2011) and Tsuja and Mariño (2013).

The study sought to demonstrate that environmental variables can assume differentiated values over a given period and associate themselves to form different environmental configurations. Thus, if the possible configurations of the environment are characterized, it is feasible to identify which indicators are most adequate to measure innovation, thus meeting the basic condition for managing innovation: measure to manage with competitiveness in sight.

After we established the two hypotheses of research, we verified that the environmental variables - economic, educational and social - assume different values over a certain period of time, ranging from positive to negative, as well as correlate with one another. It is also found that the correlations of these variables, in their different forms (positive or negative), eight different environmental configurations may result.

Therefore, it is evidenced that a new need arises in the study of innovation measurement, which is to characterize and identify in what environmental context the innovation process occurs so that it is possible to select which indicators are the most adequate for achieving this purpose. However, in order to make this attempt feasible, studies are needed that relate the indicators of innovation measurement to the respective environmental configurations.

Therefore, it is mandatory to conduct new studies to expand the scope of approach to the theme, since in this research the environmental policy variable was not considered due to the absence of valid and reliable indicators for its treatment.

Thus, failure to consider the political variable is a limitation of this study, because, in theory, there is the possibility of conformation of 16 environment configurations (CA => 24 = 16), and this study considers only the economic, educational and social variables, totaling eight possible environmental configurations.

Finally, the knowledge of the possible environmental configurations allows understanding the interaction between the environment and innovation. It allows greater rationality to the innovative process, because the activity of measuring innovation according to the type of context where it occurs makes management more effective with the ensuing desired results, such as competitiveness.