Assessing the sustainability index of part-time and full-time hazelnut farms in Giresun and Ordu Province, Turkey

The study’s primary purposes were to assess the sustainability of hazelnut farms and explore the effects of part-time and full-time farming types on sustainability in hazelnut production in the Giresun and Ordu Province of Turkey. One hundred fifty-two hazelnut farms were selected using the stratified sampling method, and data were collected by using face-to-face questionnaires. Several steps were taken, including using factor analysis after standardizing the variables to determine their weights to calculate the composite hazelnut farms’ sustainability index. The research findings showed that overall hazelnut sustainability scores of farms varied from 0.28 to 0.59, and the average score was 0.44 at sampled farms. The composite hazelnut sustainability index was at an unsatisfactory level. The social and economic sustainability index values of farms were equal, and they were higher than the environmental index value. The values were 0.50 and 0.30, respectively. The economic sustainability index score of full-time farms was higher than that of part-time farms, and part-time farms had higher environmental sustainability index scores than that of full-time farms. Social sustainability scores were not different in terms of farm type. It was recommended that when designing and regulating support policies, policy-makers should differentiate part-time and full-time farming. Training and extension programs must be planned to increase the level of knowledge of every willing farmer. To increase sustainability, specific policies are developed according to the farming type.


Introduction
Sustainable development is an essential component of every government and research institution's vision, mission, and strategies (Roy and Chan 2012). Similarly, the sustainable agriculture concept is also commonly used, and it is strongly emphasized in the transition towards sustainable development (UN 2002). Various problems limit the empirical application of the sustainable agricultural concept in the real world due to the temporal nature of sustainability. First, it has little practical value due to the limited feasibility of performing long-term experiments. Second, there is a need to identify the demand that is required to be satisfied by the farming system to achieve sustainability (Gómez-Limón and Riesgo 2009).
Consequently, the concept of sustainability can be regarded as a social concept that can be reformed in response to society's requirements. Moreover, it can be stated that sustainability is the time and place-specific concept which describes sustainability as the social structure in geographical and temporary contexts (Gómez-Limón and Sanchez-Fernandez 2010). Three sustainability dimensions (economic, social, and environmental) have gained importance to resolve these problems and deal with sustainability assessment. This approach of assessing agricultural sustainability in operational terms can be performed by developing indicators of fewer than three dimensions of the sustainability mentioned above (Bell and Morse 2012).
Although there are many different interpretations of sustainable agriculture in the review of literature, the commonality among those definitions describes the farming should be economically sound, socially viable, and environmentally friendly (Edwards et al. 1990;Hansen 1996;Singh 2013). However, agricultural sustainability assessment based on the indicators still has some deficiencies. The difficulty of defining the entire set of indications is the fundamental issue with the indicator evaluation approach. Agricultural sustainability has been proposed to be measured by combining a multidimensional set of indicators into a single index (Gómez-Limón and Riesgo 2009). This approach is commonly used (Andreoli and Tellarini 2000;Pirazzoli and Castellini 2000;QIU et al. 2007;Rigby et al. 2001;Sands and Podmore 2000;ul Haq and Boz 2020;van Calker et al. 2006). Though the theoretical principles, dimensions, and goals of sustainability in agriculture worldwide are adaptable, the applicability of the indicator is minimal due to the different social norms and geographical and climatic differences among the areas, regions, and countries. For this reason, the sustainability assessment requires particular attention, and it needs sufficient knowledge and expertise throughout the developing goals process, selection of indicators, validation of indicators, and evaluation of agricultural sustainability (ul Haq and Boz 2020).
Commonly, there are three different spatial levels for assessing agricultural sustainability. Therefore, a minor spatial level is parcel/field level, followed by the farm level, and higher spatial level such as landscape, region, or state (van Cauwenbergh et al. 2007). Numerous researches have been carried out at various spatial levels all around the world, for example, farm-level (Eckert et al. 2000;López-Ridaura et al. 2002;Meyer-Aurich 2005;van der Werf and Petit 2002), field-level (Bockstaller et al. 1997;Mitei 2011;Terano et al. 2015), and regional-level (Payraudeau and van der Werf 2005;Zhen et al. 2005). With the many sides that must be satisfied at each geographic level, these investigations differ in their indication adoption (field, farm, and regional). Although many studies have successfully adopted the indicators in the same region to assess agricultural sustainability, they have limited applicability in other areas due to the climatical and geographical differences among the countries and regions (Hatai and Sen 2008;Sharma and Shardendu 2011;Tellarini and Caporali 2000). Consequently, the indicators became the essential tools for assessing agricultural sustainability in economic, social, and environmental dimensions over the years.
In Turkey, many studies focused on farm sustainability; for example, Füsun Tatlidil et al. (2009) studied farmers' perception with farm sustainability and its determinants; Gündüz et al. (2011) assessed the apricot farm sustainability and analyzed its determinants model; Saysel et al. (2002) described the environmental sustainability in the project of agricultural development; and Ceyhan (2010) assessed the economic, social, and environmental sustainabilities of traditional farming systems in Samsun province of Turkey. ul Haq and Boz (2018) have proposed the selection of indicators for the tea crop, which is friendly in use for the other crops. Moreover, ul Haq and Boz (2020) also applied the selected indicators to measure the tea farms' sustainability. Specifically, Demiryürek et al. (2018) focused on the sustainability of conventional and organic hazelnut farms. Although these studies explain farm sustainability from different aspects, literature related to the sustainability of hazelnut farms regarding full-time and part-time farming is not available.
The full-time and part-time farming have multidimensional impacts on farm management and on the use of farm resources; for example, in the effect of full-time and parttime on resource use efficiency, part-time farmers use low labor inputs, and full-time farmers invest less in capital and materials (Amodu et al. 2011;Haiguang et al. 2013). Giourga and Loumou (2006) describe that part-time farming can lower the demand for natural resources and make farming in mountainous and semi-mountainous areas sustainable. Therefore, the prevalence of full-time and part-time farming in hazelnut production may be different in their farm practices, impacting sustainable hazelnut farming. Moreover, too close planting, bad soil conditions, soil erosion, the prevalence of old and underproductive shrubs, limited availability of necessary inputs, and care were those problems which limited the yield of hazelnut in Turkey as compared to the other hazelnut producer of the world (Hütz-Adams 2012; Lundell et al. 2004). The sustainability of the hazelnut is necessary to be evaluated regarding the full-time and parttime farming for a progressive rise in crop production.
The main purpose of the study was to assess the sustainability index of hazelnut farms and explore the effects of part-time and full-time farming types on sustainability index in hazelnut production in the Giresun and Ordu provinces of Turkey. The current study has the following study objectives. First of all, a set of indicators was explicitly developed relevant to the assessment of hazelnut farming in the study area of Turkey. Based on these indicators, the sustainability of full-time and part-time hazelnut farms was assessed.

Research area
Turkey is the largest hazelnut producer, having 62% share in the total world production. The whole hazelnut-growing area was 0.735 million hectares, and 0.665 million tons of hazelnut was produced in 2020 in the country (Table 1).
The Eastern Black Sea Region makes one-sixth of Turkey's geographical area and plays a vital role in the economy as this area is famous for tea and hazelnut production. Especially, Ordu and Giresun have significant contributions to the country's total hazelnut production. These provinces have 43% of the total hazelnut production area of the country.
Moreover, these provinces contribute substantially (42%) to the country's total hazelnut production. They were selected as a research area by considering the importance of these provinces in Turkey's hazelnut production.

Climatic conditions of the research area
Both selected provinces have a humid subtropical climate with hot and humid summer seasons and a cool winter season. They have a high and evenly distributed precipitation throughout the year. Hazelnut needs a mild climate with temperatures varying between − 80 and 360 °C throughout the year and rainfall more than 700 mm for bountiful yield. The mean minimum temperature went between 3.90 and 20.550 °C, and the mean maximum temperature remained between 10.350 and 27.450 °C from 1981 to 2010 in the study area. The precipitation varied between 63 and 129 mm throughout the year during the same period (Table 2). Hazelnut grows well in loamy vegetal and deep soils rich in nutrition as they allow hazelnut plants to intake a more significant amount of soil nutrients. Hazelnut crops cannot bear windy and stormy conditions with high summer temperatures and low dampness. Thus, the climate and soil of the selected provinces are well suited for hazelnut production.

Sampling procedure
The target population of this study was hazelnut farmers residing in the Ordu and Giresun provinces of the Black Sea Region. The list of farmers involved in hazelnut farming was obtained from both provinces' Ministry of Agriculture and Forestry. The first challenge in the sampling procedure was to determine a sample size to represent all the hazelnut farmers residing in the study area. Therefore, the stratified sampling formula proposed by Yamane 2001 was used to estimate the sample size for each province separately. The formula used in the study is shown below (Yamane 2001): n = sample size. N = total hazelnut farmers in main layer. Nℎ = number of hazelnut farmers in each layer. Sℎ = standard deviation in each layer. D 2 = anticipated variance. e = accepted error from mean (10%). t = confidence interval (95%). The used sampling technique ensures that different groups in the mainframe are represented adequately in the study. Moreover, this method also reduces variance by separating homogeneous groups from the mainframe population. The accessible population was arranged in ascending order according to their land under hazelnut farming. After that, hazelnut farmers in the study area of both provinces were divided into three strata according to their land size. The first stratum contained hazelnut farmers having a land less than 1 hectare. The second stratum consisted of hazelnut farmers having land between 1 and 1.99 hectares, and the rest of the hazelnut farmers were in the third stratum. Samples were made separately for each province. In this way, 152 (Ordu 75 and Giresun 77) hazelnut farmers were selected as the total size of this study (Table 3).

Identification and classification of sample farms
In previous studies, classifications were made by income, farm size, labor force, farmers' residence status, farming income, and capital elements (Greeley 1942;Lien et al. 2010;Mittenzwei and Mann 2017;Pfeiffer et al. 2009;Schmitt 1989). However, considering the socio-economic characteristics in the study area, we preferred to use the site-specific classification method. Due to the unique structure of agriculture, it would be more accurate to evaluate the working hours during a production period instead of weekly or monthly working periods such as the service and industry sectors. In addition, the need for labor in fruit production made from perennial plants such as hazelnuts is in certain periods. For these reasons, in this study, the phrase "working less than 2/3 of the normal working time" indicated in the Labor Law of the Republic of Turkey and the relevant Cabinet Decision has been taken into account. The study used the percentage of the payment for family members in total labor cost in hazelnut production as a classification criterion. When classifying the farmers, the labor cost coefficient was also used to reflect the risk and workload of each production activity, such as fertilizing and harvesting. If the total labor cost percentage of family work payment were more extensive than 67%, the farms would be defined as full-time farms. Otherwise, farms were classified as parttime farms. The classification results showed that 53% of the total sample farms were full-time hazelnut farms, and the rest was part-time hazelnut farms. The percentages of part-time farms in Ordu and Giresun were 54% and 46%, respectively.

Indicator selection
As possible confusing and varied aspects of an indicator limit its applicability to a precise location and time; selecting acceptable indicators to assess the sustainability of a farm for a region is a very complicated and thorough task. Because the indicator's applicability is limited due to varying climatic and geographic conditions, ul Haq and Boz (2018) have developed a comprehensive and userfriendly process for selecting indicators. They extensively focused on the site-specific characteristics of the region to develop the "basic factors." The site-specific characteristics include (a) climate and land requisites of the tea plant, (b) farming community in locality, and (c) socioeconomic characteristics of the study area. Based on these site-specific characteristics and reviewing the literature, the basic factors were developed. The purpose of developing the basic factors was to obtain the basic information on the indicators representing the characteristics of the study area with concerned farming activity, for example, hazelnut farms in the study area. However, an indicator has limited applicability over the different areas, regions, and countries. The farm-level indicator-based sustainability studies can be consulted to determine the possible adaptable indicators in the study. Similarly, the comprehensive list of indicators was prepared considering the different worldwide conducted farmlevel sustainability assessment studies, for example (Dillon et al. 2009;Gafsi and Favreau 2010;Rigby et al. 2001;Sajjad and Nasreen 2016;ul Haq and Boz 2018). Subsequently, if the indicator is adoptable directly in the current study, then it is included in the final set of indicators after confirming the selection criteria are met. For example, education and age of the farm manager were the adoptable indicators in the current study.
The selection criteria were defined to select indicators to confirm their applicability in the study area. It included the ability of the indicator to clarify the complex phenomenon, measurability, user friendly, understandability, socially and economically viabilities, and fulfillment of defined objectives of the study (Bossel 1999;Nambiar et al. 2001;Pannell and Glenn 2000;Reed et al. 2006). The current study followed the selection criteria, which are explained below: 1. The indicator should be scientifically valid, having the ability to clarify the phenomenon clearly (scientific validity). 2. For calculating the true value of the indicators, the data should be available (data availability). 3. The method to measure the true values of the indicator should be available even though the data is available (measurability). 4. The indicator should be easily interpretable for any new researchers, and readers (easily interpretable). 5. The selected indicator should be easily understandable and easily usable by the end-user (understandability). 6. The indicator should be sensitive to the three dimensions of sustainability status changes. It means the indicator should represent the true changes whenever used in the future (sensitivity).
The indicator passed out the selection criteria and was adaptable, and then it passes through the validation process, which is necessary to confirm its creditability and correct performance (Cloquell-Ballester et al. 2006). While the development of the indicators for assessing farm sustainability has been happening for many years, very little has been described on the verification of the indicators (Rigby et al. 2001). It was defined that the indicator is valid when it is well-grounded, achieves the intended effects, and fulfills the desired objective (Bockstaller and Girardin 2003). To confirm the validity of the indicators, the 3S methodology such as self, scientific, and social authorization (Cloquell-Ballester et al. 2006) was used in this study.
The study's research team confirmed the self-validation considering the selection criteria defined herein. The research team included 1 professor and 1 research assistant who are experts in the field of agricultural policy, and 1 research assistant who is an expert in the field of agricultural management. In addition, opinions were taken from experts working in the region. They tried to confirm the correct performance of the selected indicators to prevent theoretical contradictions and operational errors. They also ratified that the other researchers' indicators are easily understandable, interpretable, and usable. The experts' opinions were obtained to ensure the indicator's objectivity, accuracy, and fairness to check its scientific validity. The hazelnut stakeholders confirmed the social validity. This activity ensured the social soundness of the indicator.
If an indicator depicts its adaptability directly in the study, its validity was confirmed. If adaptability was poor, then the indicator was replaced with a new one, and a similar process was repeated for the newly added indicator to check its validity. Consequently, the final set of indicators was developed and is illustrated in Table 4, Table 5, and Table 6 and used for the data collection and scoring to assess the hazelnut farms' sustainability regarding full-and part-time farming. This selection of indicators was performed according to the selection procedure extensively proposed by ul Haq and Boz (2018), and the basic factors were considered the same as in the current study. Similar indicators were used in this study   Erosion risk of soil 0 no, 1 yes Stable terracing 0 no, 1 yes Tree planting at landside and erosion-prone land 0 no, 1 yes due to their adaptability being confirmed after passing the indicators from the validation process.

Data collection and scoring
Face-to-face interviews with hazelnut growers were performed to collect data using a well-defined and well-structured questionnaire. A diversified set of indicators requires different types of data (quantitative and qualitative) to calculate the final or true values of the indicator. For this, quantitative data was gathered related to the quantity and prices of inputs used in the crops cultivated at the farm, amount and fees of output produced at the farm, farmers' action, and their vision to make hazelnut farming economically viable, socially acceptable, and environmentally friendly. Moreover, information about the farm structure, production technology, and characteristics of the farmers and farms was obtained. For example, the qualitative data and farmers' responses were recorded in yes/no form or on a Likert scale (5-point). Calculating the true values of the indicators was not a difficult task. The only efficiency score of hazelnut farmers, including production, economic, and eco-efficiency, was estimated using the data envelopment analysis (DEA) program.
The indicators given in Table 4, Table 5, and Table 6 were different; some were continuous, and some were in the response form. Furthermore, scoring was carried out to overcome the variety among these indicators to make them useable by converting them to a unit-free format. We used scientific information, a curve of production possibility, questionnaire responses, and expert judgment to score the indicators. For example, chemical fertilizer was scored based on scientific knowledge. It means the recommended quantity for one unit of land or one hazelnut tree was used as a benchmark value. Moreover, chemical fertilizer is supposed to contribute to the adverse environmental effects, although their share in production is not negligible; in such a case, some farmers may apply fertilizer less than the recommended value. Therefore, the actual minimum amount of fertilizer applied was used as a benchmark.
The production possibility curve described the maximum attainable output of each hazelnut farm upon using similar level inputs (van Passel et al. 2009). Since the efficiency scores will be the ratio of the actual productivity achieved by the hazelnut farmer to the maximum achievable productivity of the farm (Meul et al. 2008). Consequently, every efficiency score was based on this concept. The maximum value of the efficiency score was used as the benchmark value in the current study. It describes that maximum efficiency contributes to farm sustainability positively.
The data gathered from questionnaires analyze subjectively various indicators in each dimension (Meul et al. 2008). Some indicators were generated using information from the questionnaire regarding the price and quantity of inputs and outputs used to grow the farm's crops. For example, the gross margin of the hazelnut farm was calculated based on the actual information of the market value of the output of crops produced at far less than the cost incurred to make them at the farm.
The experts' judgment helped score the response variable when none of those abovementioned scoring methods was suitable. In this approach, many indicators were responded to by the hazelnut farmers in yes and no forms. If a yes type of response positively helps sustainability, it was given a score of 1; otherwise, it was given a 0. Similarly, if no response contributes to higher sustainability, it was scored 1; otherwise, 0. For example, indicators such as buying the new land for extending the hazelnut orchard are responsive indicators. Their contribution toward the farm sustainability is based on the yes response of the farmer: 1 was given to yes response and 0 to no response.

Sustainability index calculation of the hazelnut farms
The selected indicators were put through a series of procedures to calculate the composite hazelnut farms' sustainability index (CHF), including normalization to estimate the weight for each indicator, calculation of intermediate indicators within each dimension, and aggregation of these intermediate indicators using their proportion of variance. As a result, all three elements of sustainability were calculated: economic, social, and environmental. All three criteria were equally weighted and aggregated to determine the composite hazelnut farms sustainability. Giving each component equal weight was because all aspects of sustainability were similarly significant.
Step 1 When all current indicators indicate various situations and have different units, they are not interchangeable. Therefore, the indicators need to be standardized to estimate each indicator's weights. This activity results in unit fewer indicators. The min-max normalization method was used (Freudenberg 2003;Gunduz et al. 2011). The following formulas were used for normalizing an indicator whose maximum value was considered a higher contributor to sustainability. The following formula was used when the maximum value of an indicator was regarded as being more sustainable.
Similarly, the following equation was used for that indicator, whose minimum value was cataloged as being more sustainable: Here; X = Actual Value of indicator Step 2 The previous step resulted in the normalized value of each indicator between 0 and 1. As a result, the weights for each indication were estimated using factor analysis. Because there are three dimensions to sustainability, each with its own set of indicators, the factor analysis was done individually. The weight for each indicator was calculated using the loading matrix and the proportion of variation. The weight for each indicator was calculated using the equation below.
where w Lj illustrates the indicator weight L in j th component.
Step 3  Where II ijk is the intermediate indicator for each dimension of sustainability; I is the economic, social, and environmental sustainability dimension for component j and farm k; and w Lj is the weight assigned to each indicator in the preceding phase.
Step 4 The following formula was used to calculate the composite index for each component of sustainability, including economic, social, and environmental.
The intermediate composite indicators were aggregated to calculate the composite index (CI k ) of farm K. The weight j used in this equation was estimated by using the following equation.
Finally, the composite index for each dimension of sustainability was equally weighted to aggregate them for measuring the composite hazelnut farm sustainability index (CHFSI); for this, the following simple average formula was applied.

Results and discussion
In the research area, full-time and part-time farmers' ages were 54 years and 56 years; agricultural experience of full-time and part-time farmers was 30 years and 32 years, respectively. While schooling of part-time farmers was higher than that of full-time farmers (p < 0.10), full-time farmers participated in more agricultural training than parttime farmers (p < 0.01). Part-time farmers worked much more off-farm work (p < 0.10) and earned much more nonagricultural income (p < 0.05) than full-time farmers. Fulltime farmers' net income was mainly based on agriculture (p < 0.10). Farmland size of part-time farmers were larger than that of full-time farmers (p < 0.05). Regarding hazelnut production, full-time farmers produced more hazelnuts per hectare compared to part-time farmers (p < 0.10).
Overall, hazelnut sustainability scores of farms varied from 0.28 to 0.59, and the average score was 0.44 at sampled farms. Based on the overall hazelnut sustainability index, it was clear that the unsatisfactory level existed. The values are similar for both types of farms. There was no statistically significant difference between the two groups (p > 0.05). Then, economic and social sustainability index value of farms was equal, and they were higher than the environmental index value (p < 0.05). The values were 0.50 and 0.30, respectively. At the same time, the economic sustainability index score of full-time farms was higher than that of part-time farms (p < 0.05); part-time farms had higher environmental sustainability index scores than that of fulltime farms (p < 0.05) ( Table 7). In previous studies, various evaluations have been made on economic, social, and environmental aspects of full-time and part-time farming, which are the three main indicators of sustainability, but these researchers reached conclusions without creating an index (Coutu 1957;Fuller 1990;Lien et al. 2010;Loyns and Kraut 1992;Paudel and Wang 2002). In these studies, using a limited number of indicators, a positive or negative perspective was given economically, socially, or environmentally (Barbier 2000;Bollman 1982;Galiev and Ahrens 2018;Gasson 1988;Latruffe and Mann 2015;Canan and Ceyhan 2021) (Table 8).
Economic sustainability score was higher in full-time farming than part-time farming (p < 0.05) ( Table 7). In literature, many studies reported that part-time farming is Land management practices 6.0 ± 0.0 6.0 ± 0.2 6.0 ± 0.0 Hazelnut sustainability index 0.45 ± 0.01 0.42 ± 0.01 0.44 ± 0.01 Table 8 Farms' group distribution of sustainability index by sustainability level * p < 0.10, **p < 0.05, and ***p < 0.01 reflects that the difference between full-time and part-time farms is statistically significant  (Barlett 1986;Haiguang et al. 2013;Jokisch 2002;McCarthy et al. 2009;Zhang et al. 2008), while there are also studies that considered positively to part-time farming as economic aspect (Alwang and Siegel 1999;Bishop 1955;Cavazzani 1977;Galiev and Ahrens 2018;Massey et al. 1993). The common point of these studies was to make a sweeping statement by considering a few economic indicators. While researcher's reasons for the positive approach to part-time farming were to diversify income and to raise the standard of living (Cavazzani 1977;Massey et al. 1993;Upton et al. 1982); to support the labor market (Alwang and Siegel 1999;Bishop 1955;Bollman 1982;Galiev and Ahrens 2018), and to increase production by making more investment in agriculture (Li and Tonts 2014), the negative approach's reasons were lower productivity and higher production cost (Haiguang et al. 2013;Jokisch 2002;McCarthy et al. 2009) and inappropriate use of an agricultural resource (Barlett 1986;Beyene 2008;Brosig et al. 2009;Rudel 2006;Zhang et al. 2008).
The average gross margin of farms was €1300 per hectares. Full-time farms had more gross margin than part-time farms (p < 0.01). The farms' financial return and economic rantability were calculated as 0.5 and 1.7, respectively. The ratios of full-time farms were higher than part-time ones (p < 0.10). Farm's average technical efficiency score was 0.82, and full-time farms were more efficient than part-time ones. Labor productivity for full-time and part-time farmers was 53 and 49 kg/day per person. The overall average labor productivity was 51 kg/day in the study area. There was no statistically significant difference between part-time and full-time farmers groups (p > 0.05). Land productivity at full-time farms was higher than part-time farms (p < 0.10). The average land productivity of the research area was 986 kg ha-1. Similarly, full-time farmers had higher income stability and score of farmers' characteristics than part-time farmers. The average value of income stability and farmers' characteristics score were 1.2 and 3.1, respectively (Table 7). Social sustainability scores were not different in terms of farm type contrary to the findings; in most of the previous studies, the part-time farming type was advantageous in terms of social aspects (Bollman 1982;Brosig et al. 2009;Haiguang et al. 2013;Upton et al. 1982;Xu et al. 2019bXu et al. 2019aYrjola et al. 2002). A couple of studies negatively evaluated part-time farming concerning the social aspect (Barlett 1986;Swanson and Busch 1985). As in the economic evaluation, the common points of the studies that make social evaluations were given general results by considering the few social indicators. The most important reasons for this positive approach are to increase access to education and other social requirements (Bollman 1982;Brosig et al. 2009;Fuller 1975;Yrjola et al. 2002); to decrease working time (Giourga and Loumou 2006;Haiguang et al. 2013); and to reduce adverse migration effects such as abandonment of land (Upton et al. 1982;Xu et al. 2019aXu et al. 2019b. Barlett (1986) and Swanson and Busch (1985) had a negative approach to part-time farming because of decreasing farmers' motivation and the working potential of laborers in agricultural sectors.
One part of social sustainability indicators, part-time farms generated more equity than full-time farms (p < 0.05). The education level of full-time farmers was lower than that of part-time farmers (p < 0.10). The average schooling year value was 8 years in the research area. Full-time and parttime farmers had the same value for other social indicators such as social security, social involvement, social inclusion, age, and old age index. While the average age was 55 years, the old age index was 0.3. Social security and social inclusion values were 5.2 and 5.4, respectively. The social involvement value of farmers was 26.2 (Table 7).
When environmental indicators were examined in the research area, part-time farmers produced more environmentally friendly than full-time farmers (Table 7). Most of the previous studies conducted by Barbier (2000), Caraveli (2000), Ceddia et al. (2009), Lorent et al. (2008, Swanson and Busch (1985), and Yrjola et al. (2002) reported similar results except Celio et al. (2014). He propounded that there was no difference between full-time and part-time farms in terms of environmental aspect score. These studies have suggested results considering only a few environmental indicators and other sustainability aspects. Some researchers suggested that part-time farmers preferred less chemical input for their farmland (Barlett 1986;Ellis et al. 1999;Giourga and Loumou 2006;Phimister and Roberts 2006;Swanson and Busch 1985), and land use behaviors of part-time farmers helped to protect the environment (Barbier 2000;Caraveli 2000;Ceddia et al. 2009;Gasson 1988;Kristensen 1999;Latruffe and Mann 2015;Lorent et al. 2008;Salvati and Zitti 2009;Yrjola et al. 2002).
Part-time farmers used less chemical fertilizer and were more efficient in eco-efficiency scores than full-time ones (p < 0.10). Other environmental indicators' value of farms was similar in full-time and part-time farms (Table 7).
According to the composite sustainability index results, which is the average of the 3 main sustainability indicators, it is similar for full-time and part-time farming types; there is no significant difference. While only one of the previous studies (Lien et al. 2010) reached a similar result, studies were suggesting that full-time (Coutu 1957;Loyns and Kraut 1992) and part-time farming types (Fuller 1990;Paudel and Wang 2002) are generally more advantageous. Table 8 shows farms' group distribution of sustainability index by sustainability level. When focusing on the difference between the sustainability level of farm type, it was clear that the full-time and part-time farms had similar hazelnut sustainability index scores in low-and high-sustainability farms groups (p > 0.05). But, the economic sustainability index of full-time farms and the environmental sustainability index of part-time farms are higher than that of another farm type in both the sustainability level group (p < 0.01). The social sustainability index score was also nearly equal for full-time and parttime farms in low-and high-sustainability levels.
Both farmers' characteristics (age, family size, agricultural experience, schooling year, etc.) and farms' characteristics (farmland, the slope of orchards, production type, buying and selling land) associated with full-time and part-time farming by low and high-sustainability farms are presented in Table 9.
Farmer characteristics such as age and agricultural experience of full-time and part-time farmers were similar for low-and high-sustainability farms. There was no statistically significant difference between farm-type groups in both sustainability levels in terms of age and agricultural experience (p > 0.05). The family size of full-time farmers was more extensive than that of part-time farmers in low-and high-sustainability farms (p < 0.10). Part-time farmers were more educated compared to full-time farmers in the lowsustainability (p < 0.10) and high-sustainability farms group (p < 0.05). Although the family labor of full-time farmers was more elevated than part-time farmers (p < 0.10), parttime farmers were hired more labor than full-time farmers (p < 0.01). In comparison, the off-farm work ratio of parttime farmers was higher than part-time farmers in low-sustainability farms (p < 0.05); full-time and part-time farms were nearly the same off-farm rate in high-sustainability farms (Table 9).
Regarding the farm characteristics, farms in low-and high-sustainability groups, part-time farms had higher farmland than that of part-time farms. In both groups, meters above the sea levels and slope were similar for full-time and part-time farms' orchards. Low-sustainability full-time farms sold their hazelnut at a higher price than part-time farms (p < 0.05). The selling price of hazelnut of high-sustainability full-time and part-time farms was equal. In both sustainability groups, 48% of low-sustainability of part-time farms produced only crop, and 29% of the high-sustainability full-time farms had both crop and animal. Buying a new farmland ratio of low-sustainability full-time and part-time farms were 21% and 16%, respectively. There was no statistically significant difference between part-time and full-time low-sustainability farms (p > 0.05). In high-sustainability farms, full-time farms bought more farmland than part-time ones (p < 0.05) (Table 9). When examining some management practices of farms, full-time farms produced more hazelnut per hectares than part-time farms in both groups (p < 0.01). The effect of production efficiency of farms is identified in the study on sustainability of Ul Haq and Boz (2020) Similarly, full-time farmers worked more time than parttime farmers (p < 0.05). While full-time and part-time lowsustainability farms used the same amount of chemicals, in high-sustainability farms, part-time farms used more chemicals than full-time farms (p < 0.05). High-sustainability fulltime farms used more barnyard manure than part-time farms (p < 0.05). Full-time high-sustainability farms tested more leaves of their hazelnut tree than that of part-time farms (p < 0.05). Using organic manure, testing soil and stable terracing rate of farms were equal in all groups (p > 0.05) ( Table 9). Using organic manure, testing soil, and stable terracing rate of farms with high sustainability levels is similar to many studies (Kleinhanß et al. 2007; Barnes and Thomson 2014). On the other hand, some studies show differences (Buckley et al. 2015;Westbury et al. 2011).

Conclusion and recommendation
Hazelnut is of great importance in the food industry. Hazelnut farming is critical economically for Turkey, the largest hazelnut producer. The study examined the sustainability index of hazelnut farms with a set of indicators that were developed specifically relevant to the assessment of hazelnut farming in the study area and to explore the effects of part-time and full-time farming type on sustainability index hazelnut production in the Giresun and Ordu Province of Turkey.
Sustainable agriculture is defined as social equality, work, land use, protection of the environment, and biodiversity. The agricultural system is environmentally friendly, profitable, and productive and maintains the social networks of the rural population. Sustainable agriculture combines economic, social, and environmental components.
Based on the evidence from the research results, it was clear that the composite sustainability index value was an unsatisfactory level. A similarly low-level index value existed in the classification by farming type. Although there was a difference between farming types regarding environmental and economic sustainability index values, the composite sustainability index is similar. To increase the overall index score and the general policies, specific policies are needed according to the farming type.
Economic sustainability is of lower value, especially in part-time farms that devote less time to their farm. The economic sustainability index value of part-times, which have a higher gross margin due to variable costs, and less profit due to their low-cost ratio and economic profitability, is lower than full-time ones. This profitability continues to decrease every year. Since they spend less time in their gardens, it is understood from the low technical efficiency score that they are technically inadequate. Particularly, part-time farmers' better monitoring of the market to reduce their average variable costs and their participation in training to increase their technical competencies would improve their economic sustainability. Full-time farmers with higher farming characteristics, which are essential indicators of adopting farming as a profession and connecting their income to farming, are advantageous in economic sustainability. In particular, the traditional structure in the region prevents complete separation from farming, negatively affects total production, and threatens the economic sustainability of part-time farms. It will be an important factor for economic sustainability for policy-makers to employ certified workers with high technical capacity in the region, with wage support, if necessary, without disturbing the traditional structure.
Farming groups have similar values in terms of social sustainability. Foreign labor employment status and education level are better among the social sustainability indicators in part-time farms. It is crucial to solving the aging problem in agriculture to increase social sustainability in hazelnut farmers in the region. In particular, both social and economic measures to be taken towards directing young people to agriculture will prevent aging in agriculture and increase education.
The value of the environmental sustainability index was the lowest among the sustainability index, exceptionally very low for full-time farms. These are the critical reasons for the heavy use of chemical fertilizers and the low eco-efficiency score. Raising awareness of full-time farmers on the use of fertilizers would positively affect environmental sustainability. Increasing the farmers' rate for farming types, soil testing, and leaf analysis and strengthening farmers' motivation to use organic fertilizer could positively affect environmental sustainability.
The study suggested that policy-makers differentiate between part-time and full-time hazelnut farming when designing and regulating support policies. Training and extension programs must be planned to increase the level of knowledge of every willing farmer. In addition, training and certification programs must be implemented to enhance the quality of the foreign labor force.
In future work, improved practices should be developed to help farmers find win-win solutions to reduce sustainability opposition. In such cases, agricultural policy tools can help to overcome trade-offs and appropriate incentives must be identified that allow for simultaneous improvement of welfare and sustainability.