Introduction

Aguascalientes has historically been characterized by producing grapes (Vitis vinifera L.) and is currently the fourth largest producer in Mexico. In 2013 the area planted to vineyards in the area was 827 ha, and generated a production of 10,500 tons and a value of production of 39.8 million pesos (^{SIACON-SIAP, 2012}). The grapes are used in the production of juices, concentrates, table wines and fresh consumption. Although there is diversity in the grape varieties that are produced in Aguascalientes (32), in 2013 the 75% of the area planted with vines corresponded to the Salvador variety whose destination is the industry of juices and concentrates.

Council data Aguascalientes Wine Growers indicate that in 2013 there were 234 producers dedicated to wine activity, which is important for job creation in the phases of production, processing and marketing. Despite the importance of the vine in the state, there is currently little information regarding the characteristics of producers to support the design of policies for crop management, technical assistance and training.

One way that has been very useful for the design of public policies to support field is through performing types of producers (^{Reyes et al., 2009}). According ^{Vilaboa and Diaz (2009)}, the characterization of producers and production systems is crucial for policy development because it allows to know the conformation of production systems, its technological components, potential and limitations of other systems. ^{Betancourt et al. (2005)}, based on management variables, productive and social characterization allows to know the level of use of technology and the process of decision-making at the farm level, allowing differentiated development and production system policies. Additionally, ^{Colonel and Ortuño (2005)} point out that the proper classification of production systems helps to understand the dynamics of development of a region or the design and management of development projects. In 1979 the Center for Agricultural Research (^{CDIA, 1979}) presented one of the first classifications of agricultural producers in Mexico; Similar work was done by the United Nations Economic Commission for Latin America and the Caribbean (^{CEPAL, 1982}) and ^{De Janvry et al. (1995)}.

Among the techniques used to perform typologies multivariate producers are allowing the simultaneous analysis of multiple measurements of the individuals studied (^{Nsahlai and Sedumedi, 2000}; ^{Forclaz et al., 2007} and ^{Carrillo et al., 2011}). Among the most used multivariate techniques are principal component analysis (ACP) that can be used to identify trends in a dataset and eliminate redundancy in the univariate analysis; Also, this analysis allows restructure a dataset containing many variables correlated in other groups of smaller components of the original variables data (^{Lezzoni and Pritts 1991}).

Cluster analysis (AC) is another technique used to identify groups of similar items together and aims to sort people into groups so that the degree of association or similarity between them is strong. Each agglomerate describes a set of members with similar features and is what is known as types (^{Hair et al., 1998} y ^{Betancourt y Villanueva, 2005}).

Establish productive and socioeconomic way grape growers in the state of Aguascalientes is important for the following reasons: a) allows the design of state or federal policies based on the socioeconomic characteristics of the different groups; b) focuses government support for those producer groups most in need, or to those where it is expected to achieve faster political impact; and c) identifies the most competitive leading producers (those who earn more profit) that can serve as an example to improve the standard of living of the producers.

Based on the importance of vinifera production in the state, the aim of this study was to characterize and group to grape growers Aguascalientes from production and socioeconomic variables through multivariate analysis, in order to obtain information on support in the design of public policies to specific strata of winemakers.

Materials and methods

The study was conducted in the vineyard area of the state of Aguascalientes which includes seven municipalities: El Llano, Cosio, Rincon de Romos, Seats, Tepezalá, Arteaga Pavilion and San Francisco de los Romo. The wine-growing area of the state has altitudes ranging from 1 700 to 2 400 meters above sea level (m); the climate is semi-dry with historical average annual rainfall of 400 to 450 mm in the towns of Cosio and Tepezalá and 450-500 mm in El Llano, Rincon de Romos and Pavilion Arteaga. The annual average temperature is 16-17 °C (^{INIFAP, 2013}). Rincon de Romos, San Francisco de los Romo, Tepezalá, Seats, Pavilion Arteaga and Cosío predominate xerosols and El Llano floors planosol type soil (^{INEGI, 2012}).

To gather the information a survey of 50 producers of grapes during the months of August and September 2013. The survey questions were developed based on the following applied: 1) characteristics of the producers; 2) management of the vineyard; 3) production costs; and 4) marketing.

To get a sample state representation for finite populations was determined. Based ^{Sánchez et al. (2012)}, the sample size was calculated as follows:

Where: n= the sample size; N= the population; z= the level of confidence; e= the error; p= the probability that the sample is representative, and q= the probability that the sample is not representative.

Considering the population (N) consisting of 234 vine growers in Aguascalientes, a confidence level of 90%, an error of 10% and a probability that the sample represents 50%, the sample size was n estimated 46 polls; however, they applied 50 represent 21% of the total population of producers and 227 hectares of vineyards. After determining the sample size, stratification of producers considering plantings took place. Finally, producers were selected by systematic random sampling survey, for which a constant interval selection (N/n) was determined.

The sampling used the population allowed to select group of individuals completely random (^{Scheaffer et al., 2006} y ^{Sánchez et al., 2012}). The amount of wine growers surveyed in each municipality is presented in Table 1.

The field data were systematized in spreadsheets in Excel 2013. Descriptive statistics were used for data analysis of the general characteristics of producers, productive aspects of the vineyards, production costs and income. Based ^{Vasquez and Cabas (2003)} and ^{Vilaboa and Diaz (2009)} principal component analysis (ACP) to the results was performed to compact the data and identify the interdependence between variables.

For the ACP PRINCOMP procedure of SAS version 9.0 was used. The correlation matrix between variables, eigenvalues and the proportion of variance explained by each of them, the eigenvectors and the main components are generated. Kaiser's criterion used to determine the number of components, which includes only those eigenvalues greater than 1 (^{Demey et al., 1994}). For the ACP 18 original variables that measured productive and socioeconomic aspects of the producers and vineyards (Table 2) were considered.

To estimate the technology index the following variables with arbitrary weighting factor used: 1) variety: Salvador (0.5) and Red Globe (0.5); 2) Site preparation: ground (0.2), fallow (0.1), multiarado (0.3), dredge (0.2) and ridged (0.2); 3) propagation: branch (0.0), bearded (0.2) and graft (0.8); 4) plant density: between 1 500-2 500 (0.0) and higher than 2500 (0.5); 5) System: bilateral to a bank (0.25) and bilateral two banks (0.25); 6) type of poles: concrete poles (0.5) and studs (0.0); 7) irrigation system: street gravity (0.5), gravity grooves (0.5) and dripping (1.0); 8) fertilization: chemical down (0.3), foliar chemistry (0.4), fertirrigación (0.4) and organic (0.4); 9) weeding: manual (0.25), chemicals (0.25) and mechanical (0.5); 10) technical assistance (1.0).

Based on ^{Vilaboa and Diaz (2009)} AC (Analysis and Cluster) we were performed to identify the producer groups; This analysis was based on the compacted and concentrated information from the ACP. The variables used in the cluster analysis were: total area under vines, technology index, the price of wages, jobs created, production costs, product pricing, income vineyard, age of the producer, producer education, experience producer, surface Salvador planted grape production value grape Salvador, value of production of Red Globe grapes, old grape vineyard Salvador, age vineyard Red Globe grapes, for the analysis of the hierarchical cluster analysis method was used Ward and CLUSTER procedure of SAS version 9.0 was used.

Results and discussion

The results of the socioeconomic aspects indicated that 98% of respondents were men aged between 23 and 81 years, averaging 59 years old. With regard to schooling, 54% of producers had only primary education, 22% secondary school 10% 12% Professional, and 2% say they have no education. The producers surveyed had 7.8 years of schooling on average. Age and education are factors that directly influence the way to grow and adopt new technologies available so the analysis of these variables turns out to be important (^{Damián et al., 2007}; ^{Velasco et al., 2009}; ^{Vélez, 2012}). Producers have indicated an average of 20 years of experience in the production of grapes, of which 34% were located in a range of 1 to 10 years, 18% between 11 and 20 years; and 48% between 21-40 years. The average plantings of producers surveyed was 4.8 ha; however, the most common area of 2 ha. In addition to growing grapes, 58.7% of farmers supplement their income with other crops such as corn and beans, 19.6% of farming activities, 10.9% of employment and 10.8% of others as their own business.

The results obtained in the ACP are shown in Table 3 where the eigenvalues of the principal components (CP). According to the data, the top five were considered CP as they are those with a value greater than 1 and accounted for 72.3% of the variation among winemakers. As it moves away from the main component CP1, the proportion of explained variance is reduced; opting for the convenience of more than five main components will increase the complexity of the model and the increased variability is insignificant (^{Demey et al., 1994}).

In the Table 4 shows the eigenvectors of the correlation matrix of the five major components are located. CP1 explained 28.5% of the total variance and is the most influential in the principal component analysis and, consequently, best explains the differences between growers and vine production systems (^{Coronel and Ortuño, 2005}). The outstanding variables were: production value of Red Globe grapes (X₁₆), income per hectare (X₇), selling price of the product (X₆), age grape vineyard Red Globe (X₁₈) performance Red Globe grapes (X₁₄) (Table 4). Within this component, pre-eminence of variables that express the economic performance of the vineyards on which reflect the technical side, as seen ^{Berdegué et al. (1990)} explain these results because economic variables express results of the production system and are the synthesis of other variables and partial processes.

Based on the variables that make up the CP1, this was called "production of table grapes" and reflects one of the main features of viticulture in Aguascalientes, as currently growth in plantings and grape production area for table seen in the northern part of the state; this is how the CP1 defines the difference between the production of growers who are table grapes in his vineyard and those who do not.

The second component was observed that the highest values were given for the variables planted grape Salvador (X₁₁), value of production of grapes Salvador (X₁₅), the total area under vines (X₁) and accounted for 19.1% of variance (Table 4). At CP2 it was called "grape production for the industry,"the relationship between the variables that make up the main component represents another feature of viticulture state, since according to the Wine Growers Council Aguascalientes AC, 75% plantings vine is grape varieties for the industry of juices and concentrates, 14% table grapes and 11% wine grapes, so the productive trend is reflected in this component.

The CP3 variables joined by experience, age and education of the producer (X₁₀, X₈ and X₉) and explained 9.6% of the variance. This component is related to the social reality of the producers and according ^{Galindo et al. (2000)} and ^{Damian et al. (2007)}, the variables that comprise influence the use and appropriation of technological innovations by winemakers.

The CP4 presented high correlation with technological index variables (X₂), the price of wages (X₃) and generated jobs (X₄), these variables were associated with the vineyard management component and explained 8.8% of the variance. The vineyard management component is important because it represents the agronomic and cultural practices expressed in the technology index, which producers make his garden and have their impact on production and fruit quality. On the other hand, the variables of wage employment and price are important in crop management because, as reported by ^{Gonzalez and Fuentes (2013)}, grape production is highly demanding job, especially in cutting fruit, watering, weeding and pruning.

Finally, the CP5 "vineyard productivity" was called and explained 6.4% of the variance; It consists of the variables age grape vineyard Salvador (X₁₇) and production costs (X₅). This component includes agronomic and economic culture of grapevine, first aspects because the age of the vines is related to performance, as during the first three years of the plantation production is not obtained; however, costs are incurred. For ^{Ojeda et al. (2012)} the physiological age of the vines is an element that should be considered in the proper handling of the vineyard where production costs are derived.

Through cluster analysis three groups of winemakers formed in Aguascalientes, which are displayed in Table 5. The Group I was formed by 78% of the sample are small producers. The average age of this group was 60 years, schooling ranging from third grade to sophomore year and have an average of 22 years to engage in the production of grapes, placing them as producers with extensive experience. The land ownership is 100% ejido and its vineyards have an average vineyard area of 2 ha and a yield of 13 t ha^-1; annually generate 24 jobs per hectare and is mainly engaged in the cultivation of grapes Salvador for the manufacture of juices and concentrates. The technological index of small producers was 3.90 ± 1.01. The largest proportion of small producers were located in El Llano, Cosio and Rincon de Romos.

Within Group I a subset of producers who have to produce vineyards Salvador Red Globe grapes and grape detected. This subgroup represented 25.6% of small producers and are located in the municipality of Cosio. You have between 0.5 and 1.5 ha with young plantations of Red Globe grapes; ie vineyards aged between 1 and 2 years that have not yet produced, only three vineyards are in production. In this subgroup diversification Vineyard it observed and incorporate alternative activities that broaden sources of income and producing grape varieties for different markets and consumer types (^{Viladomiu et al., 2002}); In addition, this subgroup was shown to have a higher technological index (4.38) compared to other small producers. It is considered that this subgroup is composed of winemakers with entrepreneurial vision where the strategy is to generate higher revenues and reduce the risk inherent in agricultural production caused by various factors such as falling prices, rising input prices, adverse weather conditions, complications in market access, the presence of pests and diseases to produce and government policies and regulations, including (^{Ostertag, 1999} and ^{Toledo et al., 2011}).

The group II (18% of the sample) consists of medium-sized producers with an average age of 56 years schooling from primary to university level and have 15 years devoted to producing grapes. Compared to Group I, the medium producers have a higher educational level and less time experience in viticulture. ^{Vilaboa and Diaz (2009)} argue that the producers of older, low education and greater experience, such as Group I, have rooted, about how to produce knowledge, which are considered reluctant to technological change; while producers with experience in the business and education, it is possible that they are in a process of transition to greater openness to technological change, a traditional production to a more entrepreneurial vision, as in the case of Group II. In addition to these features, winemakers Group II observed an average planted area of 4.5 ha and obtained a yield of 16 t ha^-1.

Unlike the Group I, this group is dedicated to producing grapes for table and has vineyards with an average age of eight years. Some of these winemakers combine production Salvador and Red Globe grapes; however, more than 90% of its revenues come from the sale of Red Globe. The medium producers have increased the Group I and higher costs of production technology index, which is explained by agricultural management practices associated with varieties for table as higher demand in irrigation and use of fertilizers and fungicides in production table grapes (^{Borja et al., 2014}). Median is located in Cosio and Aguascalientes.

The group III (4% of the sample) consists of producers with an average age of 50 years and 10 years devoted to grape production. This group is similar to Group II school. The size of land in this group is 16 ha, specialize in producing grapes Salvador, and obtained an average yield of 24 t ha^-1. This group observed the minor technological index of 3.70 and lower costs of production, this behavior is because they are producers with the largest area of vineyards, thus minimizing costs from make fewer cultural practices that involve less use of hand work. The Group III does not show a tendency location, specific cases of producers is in Rincon de Romos, seats and based on information provided by the Council Winegrowers Aguascalientes AC there are producers with these features in the municipalities of Jesus Maria and Cosio.

Conclusions

Based on production characteristics and socioeconomic as the area under vines, yields, cultivated varieties, the destination of production (industry and grapes), the cost of production, selling price and the income from unit surface, the winemakers of Aguascalientes can be classified into three groups: small, medium and large producers. Most small producers are located in the municipalities of El Llano and Cosio, medium in the town of Cosio and large producers are scattered in several municipalities in the state. Of the three groups, small farmers receive the lowest income per hectare; however, within this group there are farmers who have diversified their production by producing grapes for table grapes industry and as a mechanism to reduce climate risks and market.

In the design of public policies, decision makers should consider the three groups of producers and the characteristics that distinguish them; therefore, training programs, government support and funding should include the expertise, the productive capacity of the vineyards and the socioeconomic aspects of the producers of each municipality.