Research on Influence Factors of Pig Farmers’ Ecological Farming Behavior：Based on the TPB and SEM

, , ZG ,

1. College of Economics & Management, Huazhong Agricultural University, Wuhan 430070, China; 2. College of Life Sciences & Technology, Huazhong Agricultural University,Whuhan 430070, China

ResearchonInfluenceFactorsofPigFarmers’EcologicalFarmingBehavior：BasedontheTPBandSEM

ChengyingHAN1,2,ZhenhongQI1,DongminZHANG1,XinruiLI1

1. College of Economics & Management, Huazhong Agricultural University, Wuhan 430070, China; 2. College of Life Sciences & Technology, Huazhong Agricultural University,Whuhan 430070, China

China’s extensive development model of pig industry gives rise to serious resource and environment bottleneck restriction, it is urgent to transform the existing development model into ecological farming model. In this process, farmers’ behavior will play a key role. On the basis of summarizing development connotation and characteristics of ecological agriculture, using survey data of 323 pig farmers in 6 provinces, based on the Theory of Planned Behavior (TPB) it introduced individual characteristics and external environment constraint variables, and made an empirical analysis on factors influencing intention and behavior of pig farmers’ ecological farming using the Structural Equation Model (SEM). Results indicate: (i) ecological farming intention of pig farmers has the largest direct effect on their ecological farming behavior; (ii) ecological farming cognition of pig farmers has significant positive effect on their ecological farming behavior; (iii) external environment has the largest indirect effect on their ecological farming behavior, followed by the overall effect, and external environment has significant positive effect on ecological farming attitude and cognition of pig farmers; (iv) individual characteristics (educational level and income level) of pig farmers influence ecological farming behavior of pig farmers through their ecological farming attitude and intention, but the degree of influence is limited.

Ecological farming behavior, Influence factors, Pig farmers, Theory of Planned Behavior(TPB) ,Structural Equation Model (SEM).

1 Introduction and background analysis

Ecological agriculture is a modern agricultural development model stressing agriculture, environmental protection and resource recycling. It is characterized by low or zero emission, advocates harmonious development of environment and agricultural economy, follows the "reduce, reuse, and recycling" (3R) principle, and is a "resource - product - regenerated resources" cyclic process. In sum, ecological agriculture carries on essence of traditional agriculture, draws on the experience of modern agriculture, and takes sustainable development as basic guiding theory, to realize simultaneous optimization of agricultural economic system, rural social system, and natural ecological system, and promote sustainable use of ecological protection and agricultural resources[1]. The purpose of ecological agriculture is to integrate agricultural production, rural economic development and ecological protection and control, and resource cultivation and effective use, realize high yield, high quality, and high efficiency and sustainable development, and reach the integration of economic, ecological and social benefits[2].

Ecological farming is an essential part of ecological agriculture, and it is favorable for strengthening construction of agriculture, forestry and animal husbandry compound ecological engineering and agricultural waste recycling engineering, and favorable for increasing agricultural production efficiency, increasing farmers’ income, optimizing resource use, and improving ecological environment. Ecological farming of pig industry refers to establishing agriculture and animal husbandry integrated food chain in accordance with principle of circular economy, and forming "producer - consumer - decomposer" interdependent, closely connected and cooperative organization network system through waste exchange, resource recycling, element coupling, and industrial connection. The essence is fundamental change in production mode of pig industry, ecological development of pig industry and relevant organizational innovation with ecological transformation of structure as the core[3]. China is a large country with pig industry and pork consumption, while traditional extensive pig farming model has brought about serious resource and environment problem[4]. On the one hand, manure generated from pig farming not only brings serious public nuisance, but also wastes huge manure resources. On the other hand, traditional pig farming model is separated from planting model and seriously restricts circulation of agricultural wastes in pig farming and crop cultivation and sustainable development of agriculture. Especially in current situation of small scale and weak environmental protection awareness of pig farmers, it is extremely necessary to establish ecological pig farming model[5].

2 Related theories and research hypotheses

2.1RelatedtheoriesandresearchesTheodore Schultz (1977) believed that farmers’ economic behavior is rational, but in actual life, farmers’ rational behavior will be restricted by their subjective cognition ability, costs for acquiring information and technologies, and external policies and environment conditions[6]. According to this, farmers’ ecological farming behavior is generally the result of mutual interaction of individual cognition and external environmental conditions. Wu Lanyaetal(2014) studied factors influencing farmers participating in ecological farming from the perspective of farmers and found that cognition of poultry and agricultural information dissemination channel greatly influence farmers’ selection of ecological farming model[7]. Studies of Zhang Yu and Qi Zhenhong (2015) also indicate that behavior of external environment entities, such as government guidance, government subsidy, and technological support of scientific research institutions, has significant influence on ecological farming cognition and diffusion[8].

Ajzen (1987) put forward the Theory of Planned Behavior (TPB) through evaluating potential decisive factors of behavior[9]. This theory is helpful for understanding how farmers change their behavior decisions. Liu Xuefenetal(2013) made an empirical analysis on factors influencing ecological behavior decision of pig farmers using TPB from the perspective of ecological cognition of pig farmers, and found that five factors have significant influence on ecological behavior decision of pig farmers, namely, annual net income of pig farmer family, farming scale, participation of farming training, product quality and safety test, and ecological cognition of pig farmers[10]. Kliebenstein and Lawrence (1995) found that reducing risks, obtaining funds, and increasing income are major reasons for pig farms adopting contractual behavior[11]. Boger (2001) studied current situation of pig farming in Poland and found that vertical coordination has significant influence on quality and safety behavior of pig farmers[12].

2.2ResearchhypothesesAccording to TPB, the more positive of attitude toward behavior, the higher restraint of subjective regulation, the stronger feeling of behavior control ability, the stronger intention of implementing certain behavior, and the higher probable to finally implement certain behavior[13]. Therefore, we put forward following hypothesis:

H1: Ecological farming intention of pig farmers has positive effect on ecological farming behavior.

Ecological farming cognition of pig farmers refers to understanding, knowledge, attitude, feeling, and intention of pig farmers about ecological farming and environmental protection. Ecological farming cognition of pig farmers directly influences their attitude towards ecological farming, and consequently influences their ecological farming intention and behavior. Ecological farming cognition is mainly manifested in basic knowledge and current situations of pig farmers about ecological farming policies. On the basis of this, we put forward following hypotheses:

H2: Ecological farming cognition of pig farmers has positive effect on their ecological farming intention.

H3: Ecological farming cognition of pig farmers has positive effect on their ecological farming behavior.

Ecological farming attitude of pig farmers includes concern for quality and safety issues, intention to cooperate with surrounding farmers, and intention to participate in farming of ecological industry chain. Ecological farming attitude is an influence factor of ecological farming behavior, and there is positive correlation between attitude and behavior, the behavior intention acts as intermediary. Ecological farming attitude is a factor influencing ecological farming behavior, and it interacts with external conditions and generates ecological farming behavior. When external conditions fail to work, attitude exerts the highest influence on behavior; when external conditions exert significant positive or negative influence, it will greatly promote or impair occurrence of certain behavior, and the influence power of attitude on behavior will become weak[14]. In combination of this study, we put forward following hypotheses:

H4: Ecological farming attitude of pig farmers has positive effect on their ecological farming intention.

H5: Ecological farming attitude of pig farmers has positive effect on their ecological farming behavior.

Individual characteristics directly influence behavior attitude, subjective regulation and perceived behavioral control through influencing behavior belief, and finally influence behavior intention and behavior. In this study, individual characteristics of pig farmers include educational level and income level, and these factors may influence ecological farming intention and behavior of pig farmers. Extensive researches indicate that there is positive correlation between educational level and new technological acceptance intention of farmers (Ervin and Ervin, 1982)[15], while some researches indicate that the influence of educational level of farmers on acceptance of new technologies is not clear (Federetal, 1985)[16]. In fact, farmers with higher educational level have higher knowledge of circular farming technologies, therefore their understanding of ecological farming is deeper, so they have higher abilities of learning and using ecological farming technologies. Income level is an essential factor influencing farmers accepting new technologies (Sinden and King, 1988)[17]. Farmers with powerful economic strength will have higher ability of accepting circular farming technologies because certain funds are needed in the process of accepting circular farming technologies. According to related research findings and survey results, we put forward following hypotheses:

H6: Individual characteristics of pig farmers exert influence on their ecological farming attitude.

H7: Individual characteristics of pig farmers exert influence on their ecological farming cognition.

External environment variable include whether accepting training, whether policies and regulations of animal husbandry environment are perfect, and satisfaction with subsidies for livestock and poultry breeding. Policy support of government plays the role of guidance and encouragement in farmers’ accepting circular technologies and ecological farming model, favorable for ecological farming behavior of farmers. However, farmers lack necessary professional and technological knowledge, thus government should propagate in appropriate approaches and organize related training, which will be favorable for active acceptance of new technologies[18]. On this basis, we put forward following hypotheses:

H8: External environment exerts influence on path of farmers’ attitude towards ecological farming.

H9: External environment exerts influence on path of farmers’ ecological farming cognition.

On the basis of the above research hypotheses, we put forward the relational model for factors influencing ecological farming behavior of pig farmers, as shown in Fig. 1.

Fig.1 Relational model for factors influencing ecological farming behavior of pig farmers

3 Data source and variable description

3.1DatasourceandvariabledefinitionData in this study came from field survey of the team "Ecological Industry Chain Symbiosis Model for Pig Industry and its Performance Based on Circular Economy" in 350 pig farmers of 14 counties (cities) of 6 provinces during July and August of 2012. The field survey was carried out by random sampling method in the form of distributing questionnaire and in-depth interviews. The survey regions include Zaoyang City and Dawu County of Hubei Province, Guangzhou City and Foshan City of Guangdong Province, Yichun City, Gao’an City, Fengcheng City, and Nanchang City of Jiangxi Province, Lipu City, Nanning City, and Leye City of Guangxi Zhuang Autonomous Region, Zhumadian and Zhoukou City of Henan Province, Ji County of Tianjin City. We surveyed 30 administrative villages, distributed 350 copies of questionnaire, and collected 323 valid copies (with effective response rate up to 92.29%), and sample distribution is listed in Table 1.

Table1Distributionofsurveyedregionsandstatisticsofvalidquestionnaire

ProvincesurveyedValidcopiesProportiontototalsamples∥%DistributionofregionsofpigfarmersHubei9529.4ZaoyangCityandDawuCountyGuangdong4313.3GuangzhouCityandFoshanCityJiangxi5015.5YichunCity,GaoanCity,FengchengCity,andNanchangCityGuangxi4814.9LipuCity,NanningCity,andLeyeCityHenan4814.9ZhumadianCityandZhoukouCityTianjin3912.1JiCountyTotal323100.0--

From Table 1, we can see that samples are mainly distributed in major pig farming areas, so they are relatively representative. The questionnaire was designed mainly on the basis of literature researches and various hypotheses and specific problems. At the same time, we made small scale preliminary survey in Hubei and made correction of questionnaire in accordance with actual operation. Formal questionnaire mainly includes basic characteristics of pig farmers, operating condition of pig farms, cooperation of pig industry chain, state policies and subsidies, and performance of pig industry chain, near 60 questions in total. This is mainly to know and find out individual characteristics and social environment factors possibly influencing ecological farming behavior of pig farmers, actual recognition degree of farmers about ecological farming, ecological farming knowledge of farmers, and acquisition approach and use of farming technologies. Definition and assignment of variables in questionnaire and model are listed in Table 2.

3.2DescriptivestatisticalanalysisAmong valid samples, male samples account for 81.1%, female samples account for 18.9%, which is consistent with the fact that pig farms are mainly operated by men. In the samples, pig farmers younger than 40 years old account for 26.6%, pig farmers in 40-50 years old account for 54.8%, pig farmers older than 50 years old account for 18.6%; pig farmers with educational level of junior middle school and below account for 59.1%; pig farmers with senior middle school and college level account for 32.2%; pig farmers with university and above account for 8.7%, reflecting most pig farmers are middle aged and their educational level is low. On the whole, most farmers are interested in our survey, and their answers can assure real and reliable survey results.

Table2Definitionofvariablesinhypothesismodels

LatentvariableQuestion(code)AssignmentofvariableMeanvalueMedianStandarddeviationEcologicalfarm-ingattitudeHowmuchdoyouconcernaboutfarmingquali-tyandsafety?(Atti1)Highlynotconcerned=1;notconcerned=2;generallyconcerned=3;relativelycon-cerned=4;highlyconcerned=54.034.000.903Whetheryouarewillingtocooperatewithsur-roundingfarmers?(Atti2)Highlyunwilling=1;relativelyunwilling=2;general=3;relativelywilling=4;highlywilling=53.403.000.881Participatinginecologicalfarmingindustrychainisfavorablefordevelopmentofpigindus-try(Atti3)Highlydisagree=1;disagree=2;general=3;relativelyagree=4;highlyagree=54.224.000.714Ecologicalfarm-ingcognitionWhatdoyouthinkofinfluenceoffarmingpol-lutiononenvironment?(Cog1)Verylittle=1;little=2;general=3;high=4;veryhigh=52.282.000.897Whatdoyouthinkofinfluenceoffarmingpol-lutiononhealthoflivestockandpoultry?(Cog2)No=1;little=2;general=3;high=4;veryhigh=52.652.001.235Whataboutyourknowledgeofpoliciesofani-malhusbandryenvironmentmanagement?(Cog3)Verylittle=1;little=2;general=3;much=4;verymuch=52.853.000.950Ecologicalfarm-ingintentionWhatabouttakingecologicalfarmingbehaviorinyourpigfarm?(Int1)Highlydisagree=1;disagree=2;general=3;relativelyagree=4;highlyagree=53.074.000.999Doyouagreetomakeenvironmentimpacteval-uationbeforebuildingpigfarm?(Int2)Highlydisagree=1;disagree=2;general=3;relativelyagree=4;highlyagree=52.612.000.924Areyouwillingtoovercomedifficultiestocon-ductecologicalfarming?(Int3)Highlydisagree=1;disagree=2;general=3;relativelyagree=4;highlyagree=52.643.000.499Ecologicalfarm-ingbehaviorWhetheryouhavebuiltmethanepoolinyourpigfarm?(Beh1)No=1;Yes=21.411.000.493Thenumberoftypesofecologicalfarmingmodelyoutakeinyourpigfarm?(Beh2)No=1;1type=2;2types=3;3types=4;4types=52.241.001.041Whetheryouhavebuiltpigmanuretreatmentfacilitiesinyourpigfarm?(Beh3)No=1;Yes=21.391.000.488Thenumberoftypesofmeasuresforcontrol-lingfarmingpollutioninyourpigfarm?(Beh4)1type=1;2types=2;3types=3;4types=4;5types=52.232.000.646Individualcharac-teristicsEducationallevel(Indi1)Primaryschoolandbelow=1;juniormiddleschool=2;seniormiddleschool(secondaryschool)=3;college=4;universityanda-bove=52.512.001.159Monthlyincome(Indi2)1000-1999=1;2000-2999=2;3000-3999=3;4000-4999=4;5000andabove=53.414.001.361Externalenviron-mentWhetheranimalhusbandryenvironmentman-agementpoliciesandregulationsarewellestab-lished?(Sit1)Highlynotestablished=1;relativelynotes-tablished=2;general=3;relativelyestab-lished;highlyestablished=52.913.000.831Areyousatisfiedwithfarmingsubsidies?(Sit2)Highlyunsatisfied=1;relativelyunsatisfied=2;general=3;relativelysatisfied=4;highlysatisfied=52.803.000.886Whetherhaveyoureceivedtrainingofecologi-calfarming?(Sit3)No=1;Yes=21.281.000.452

Table3Differenceinecologicalfarmingbehaviorbetweendifferentsizesofpigfarms

SizeofpigfarmEcologicalfarmingbehaviorNo∥%Yes∥%Proportiontototalsamples∥%Lessthan100pigs12(41.5%)31(58.5%)53(16.4%)100-499pigs38(25.3%)112(74.7%)150(46.4%)500-999pigs21(34.4%)40(65.6%)61(18.9%)1000-1499pigs4(14.8%)23(85.2%)27(8.4%)1500pigsandmore6(18.8%)26(81.3%)32(9.9%)Total85(26.3%)238(73.7%)323(100%)

As to overall cognition of pig farmers about ecological farming, according to our survey, only 53.3% farmers know ecological farming, 46.7% farmers do not know or are unclear of ecological farming; 59.1% farmers know circular agriculture and 40.9% farmers do not know circular agriculture; 88.2% farmers know green foods, and 11.8% farmers do not know green foods. Therefore, most pig farmers have little knowledge of ecological farming and circular agriculture, while they have high recognition of green foods. Different sizes of pig farms have different ecological farming behavior. Through statistics of questionnaire, pig farms with 100-499 pigs account for the largest proportion (46.4%); 85.2% pig farms with 1000-1499 pigs have taken the ecological farming model.

4 Model selection and reliability and validity test

4.1SelectionofmodelsEcological farming cognition, intention and behavior of pig farmers have basic characteristics of difficult measurement and difficulty in avoiding subjective measurement error. Structural Equation Modeling (SEM) is a series of statistical methods that allow complex relationships between one or more independent variables and one or more dependent variables. In this study, we selected SEM to analyze factors influencing ecological farming intention and behavior of pig farmers.

SEM includes[22](i) measurement model, reflecting the relation between latent variables and measurable variable; (ii) structural model, reflecting structural relation between latent variables. SEM generally consists of three matrix equations:

η=Bη+Γξ+ζ

(1)

Y=Λyη+ε

(2)

X=Λxξ+δ

(3)

where equation (1) is the structural model which specifies the causal relation between hypothetical hidden exogenous variables and hidden endogenous variables;Bsignifies the effect coefficient matrix of hidden endogenous variables to hidden endogenous variables;Γdenotes the effect coefficient matrix of hidden exogenous variables to hidden endogenous variables; ζ refers to vector consisted of residual error; equation (2) and equation (3) are measurement model which specifies the relation between hidden endogenous variable η and latent endogenous variableYand the relation between hidden exogenous variable ξ and latent exogenous variableX;Λyand denote regression coefficient or factor load matrix of hidden variable η and ζ; ε and δdenote measurement error of latent variableYandX.

4.2ReliabilityandvaliditytestTo test science of the scale, we made reliability and validity test. The reliability is the degree of consistency and reliability of measurement results (data). In this study, we used Cronbach’sαas measurement indicator. Whenαis greater than or equal to 0.7, and whenαis in the range of 0.35-0.7, and below 0.35, the corresponding reliability is high, general, and low respectively. Inputting the data of this study to SPSS17.0 software, results indicate that all other four items ofαcoefficient are higher than 0.7 except ecological farming attitude and individual attitude (0.685 and 0.673 separately), reflecting high reliability of sample data.

The validity is the degree of measurement tools and means accurately measuring the objects, and it generally consist of content validity and construct validity. Content validity is the estimate of how much a measure represents every single element of a construct. Our questionnaire was designed with reference to research findings, related theoretical analysis, expert review and correction, and on the basis of preliminary survey conditions. Theoretical references mainly include TPB model, farmers’ behavior intention model, and farmers’ ecological behavior model. Thus, these basically ensure dimension of questionnaire can contain major factors influencing ecological farming behavior of pig farmers, and are representative and have excellent content validity. The construct validity is the degree to which a test measures what it claims, or purports, to be measuring. Generally, if factor analysis can be carried out on sample data, it can be deemed that sample data have excellent construct validity. KMO value of latent variables in this study is close to or greater than 0.7, Bartlett test value reaches significance effect, indicating that sample data can conduct factor analysis. The specific analysis results of sample reliability and validity of samples, as shown in Table 4.

Table4Analysisresultsofreliabilityandvalidityofsamples

LatentvariableItemsofmeasurementCronbachsαcoefficientKMOvalueBartletttest(significance)Ecologicalfarmingbehavior4items0.7040.793125.626(0.000)Ecologicalfarmingintention3items0.8120.755173.677(0.000)Ecologicalfarmingattitude3items0.6850.699103.786(0.000)Ecologicalfarmingcognition3items0.8010.798136.006(0.000)Individualcharacteristics2items0.6730.700118.106(0.000)Externalenvironment3items0.7450.764123.907(0.000)

5 Result estimation and path analysis

5.1ModeladaptationWe use Amos20.0 software as analysis tool. In the analysis of models of this study, we found that the goodness of fit is relatively low and standardized path coefficient of model estimation value is greater than 1. In other words, there is offending estimate, and statistical output estimation value of model test exceeds the acceptable range. According to application procedure of structural model, before making overall adaptation estimation for models, it is required to firstly test whether models have offending estimate (Rong Taisheng, 2010). Items of test of offending estimate include existence of negative error variance and standardization coefficient exceeds or is close to 1 (usually taking 0.95 as threshold)[19]. On this basis, we made several times of correction for models, gradually determined measurement variables leading to offending estimates, namely, farming years and nature of pig farms. Deleting these two factors, we finally obtained models with good of fit as shown in Fig. 2. The corrected structural equation model indicates that standardization coefficient of every path does not exceed or get close to 1, there is no negative error variance, and correlation coefficient of standardization estimation value of covariance does not exceed 1, and there is no extremely large or small standard error. The test results of overall model adaptation are listed in Table 5. From Table 5, we can see that all indicator results are excellent or remain in acceptable range. In other words, causal relation model in this study is consistent with actual survey data.

5.2EstimationofmodelparametersAccording to theoretical research model, with the aid of AMOS20.0, we obtained path coefficient calculation results of all structural model variables, as listed in Table 6 and shown in Fig. 2. Table 6 indicates non-standardized path coefficient (RW), standardized path coefficient (SRW), standard error (SE), critical ratio (CR), and significance.

Table5Analysisresultsofgoodnessoffitforstructuralequationmodel

GoodnessoffitIndicatorresultGeneralrequirementRemarksx2/df1.313<2IdealRMR0.046<0.05IdealGFI0.803>0.8IdealAGFI0.823>0.8IdealTLI0.852>0.9AcceptableNFI0.704>0.8AcceptableCFI0.941>0.9IdealRMSEA0.106<0.1Acceptable

Table6Estimationcoefficientofmodel

PathofinfluenceNon-standardizedpathcoefficient(RW)Standardizedpathcoefficient(SRW)Standarderror(SE)Criticalratio(CR)SignificanceEcologicalfarmingattitude<-Individualcharacteristics0.2260.3190.1231.833NotsignificantEcologicalfarmingcognition<-Individualcharacteristics-0.190-0.0970.350-0.542NotsignificantEcologicalfarmingattitude<-Externalenvironment1.4860.779***0.4293.464SignificantEcologicalfarmingcognition<-Externalenvironment4.8020.912***1.1564.154SignificantEcologicalfarmingintention<-Ecologicalfarmingcognition0.2570.602**0.0872.958SignificantEcologicalfarmingintention<-Ecologicalfarmingattitude0.3440.2920.2071.661NotsignificantEcologicalfarmingbehavior<-Ecologicalfarmingintention0.4900.934***0.1164.242SignificantInt1<-Ecologicalfarmingintention1.0000.387------Int2<-Ecologicalfarmingintention0.0690.0330.1290.531NotsignificantInt3<-Ecologicalfarmingintention1.0650.432***0.2105.074SignificantAtti1<-Ecologicalfarmingattitude1.0000.320------Atti2<-Ecologicalfarmingattitude2.0680.629***0.4125.016SignificantAtti3<-Ecologicalfarmingattitude2.4680.713***0.4805.146SignificantIndi1<-Individualcharacteristics1.0000.491------Indi2<-Individualcharacteristics1.1670.475***0.2644.419SignificantSit3<-Externalenvironment1.0000.335------Sit2<-Externalenvironment0.3780.0420.5580.678NotsignificantSit1<-Externalenvironment5.3360.697***1.0015.333SignificantCog3<-Ecologicalfarmingcognition1.0000.863------Cog2<-Ecologicalfarmingcognition0.6840.441***0.0977.027SignificantCog1<-Ecologicalfarmingcognition0.1980.175**0.0692.861SignificantBeh4<-Ecologicalfarmingbehavior1.0000.367------Beh3<-Ecologicalfarmingbehavior2.4730.897***0.4405.625SignificantBeh2<-Ecologicalfarmingbehavior0.3340.131*0.1612.080SignificantBeh1<-Ecologicalfarmingbehavior0.5990.243***0.1713.493Significant

Note:CRis the ratio of parameter estimation value to standard error of estimation value, equivalent totvalue orzvalue, ifCRis greater than 1.96, the estimation value is significant at the level below 0.05; if this value is greater than 2.58, the estimation value of parameter is significant at the level below 0.01. * denotesp<0.05, ** denotesp<0.01, and *** denotesp<0.001. The "--" denotes basis (unit 1) of SEM parameter estimation, taking this as significant path to estimate the significance of other paths.

Direction of path coefficient (positive and negative) and significance are two bases for judging whether hypotheses hold true. In this study, the model path and its estimation parameter results indicate that, all other coefficients pass thettest (t>1.96) except the path coefficient between ecological farming attitude and cognition of pig farmers and between ecological farming attitude and intention. This is shown in Fig. 2. The path coefficient t value of individual characteristic to ecological farming attitude and the ecological farming attitude to ecological farming intention is close to 1.96, and p value is 0.067 and 0.097 respectively, showing significant at the level below 10%. Also, result estimation indicates that all variables exceptInt2 andSit2 pass the significance test.

Fig.2 Corrected model path and parameter estimation results

In order to visually understand the direct, indirect and overall effect of factors influencing ecological farming behavior of pig farmers, we made an analysis by weight method, as shown in Table 7.

Table7Effectoffactorsinfluencingecologicalfarmingbehaviorofpigfarmers

PathDirecteffectIndirecteffectOveralleffectEcologicalfarmingbehavior<-ecologicalfarmingintention0.934--0.934Ecologicalfarmingbehavior<-ecologicalfarmingattitude0.4120.0510.463Ecologicalfarmingbehavior<-ecologicalfarmingcognition0.5280.0800.608Ecologicalfarmingbehavior<-individualcharacteristics--0.0860.086Ecologicalfarmingbehavior<-externalenvironment--0.7530.753

As to the direct effect, the ecological farming intention of pig farmers has the largest direct effect on their ecological farming behavior, and the next is ecological farming cognition; as to the indirect effect, external environment has the largest indirect effect on ecological farming behavior, and these bring into play through ecological farming cognition and ecological farming intention of pig farmers and accordingly influence their decisions. As to the overall effect, the ecological farming intention of pig farmers has the largest overall effect on their decision, followed by external environment, and finally is the ecological farming cognition. Generally, ecological farming behavior of pig farmers are not only influenced by their psychological factors including cognition, attitude and intention, but also influenced by individual characteristics and external environment. Especially, external environment has obvious regulation function and its action path is also proved.

5.3Analysisofmodelresults(i) Ecological farming attitude, ecological farming cognition, and ecological farming intention of pig farmers have significant positive influence on their ecological farming behavior, thus our hypotheses H1, H2, and H4 pass the test. Ecological farming intention of pig farmers has the largest direct effect on their ecological farming behavior (standardization path coefficient is 0.93), followed by the ecological farming cognition, which is consistent with individual behavior is determined by individual intention according to the TPB. Ecological farming intention of pig farmers directly influences their decision behavior of ecological farming, while ecological farming intention is also influenced by their ecological farming attitude and cognition, and the direction of influence is positive. Through survey and sample data analysis, we found that whether pig farmers concern about environmental protection of pig farms and whether there is cooperative relation between pig farms and surrounding pig farmers have significant influence on their ecological farming intention.

(ii) The influence of individual characteristics (educational level and income level) of pig farmers is not significant on ecological farming attitude and cognition of pig farmers, which is not completely consistent with hypotheses H6 and H7. Through sample data analysis, we found that the path coefficient t value of individual characteristic to ecological farming attitude and the ecological farming attitude to ecological farming intention is close to 1.96, andpvalue is 0.067 and 0.097 respectively, showing significant at the level below 10%. Possible reasons are most research objects specialized pig farmers. Although their educational level is different, due to long time of pig farm operation, their knowledge of ecological farming and environmental protection is basically consistent, showing homogeneous feature. Besides, although implementation of ecological farming behavior and acceptance of ecological farming technologies need fund input and support, pig farms provide the funds, rather than farmers. Therefore, the influence of individual income on ecological farming attitude and cognition is limited.

(iii) Ecological farming attitude and cognition of pig farmers influence ecological farming intention, then act on ecological farming behavior (the standardization path coefficient is 0.412 and 0.528 respectively), which is consistent with hypotheses H3 and H5, and the external environment has significant regulation function to ecological farming attitude and ecological farming cognition of pig farmers. Research indicates that all measurement questions of ecological farming attitude and ecological farming cognition of pig farmers pass the significance test. Specifically, pig farmers’ concern about farming quality and safety, intention to cooperate with surrounding farmers, acceptance of ecological industry chain cooperation helping the ecological farming, and influence degree of farming pollution to environment and livestock and poultry health, as well as knowledge of animal husbandry environment management policies directly influence ecological farming attitude and cognition of pig farmers, and accordingly influence their ecological farming intention.

(iv) External environment has significant positive influence on ecological farming attitude and cognition of pig farmers (standardization path coefficient is 0.91 and 0.78 respectively), which is consistent with hypotheses H8 and H9. External environment has the largest indirect effect on ecological farming behavior of pig farmers, thus ecological farming of pig industry needs government support and regulation. Model test and path analysis indicate that whether pig farmers accept ecological farming technical training influences their farming behavior decision through their ecological farming attitude and cognition; whether animal husbandry environmental management policies are well established exerts path influence on ecological farming attitude and cognition of pig farmers. However, pig farmers’ satisfaction with farming subsidies fails to pass the significance test. The reason is possibly that subsidy projects are mainly provided for breeding pigs, vaccines and hog insurance, but no subsidies for ecological farming technical training and little subsidies for construction of large-scale farms.

6 Conclusions and recommendations

6.1ConclusionsBased on survey data of 323 pig farmers in 6 provinces of China, with the aid of structural equation model, we arrived at following conclusions: (i) ecological farming intention of pig farmers has the largest direct effect on their ecological farming behavior (standardization path coefficient is 0.93); (ii) ecological farming cognition of pig farmers has significant positive effect on their ecological farming behavior (standardization path coefficient is 0.60); (iii) external environment has the largest indirect effect on their ecological farming behavior, followed by the overall effect, and external environment has significant positive effect on ecological farming attitude and cognition of pig farmers (standardization path coefficient is 0.91 and 0.78 respectively); (iv) individual characteristics (educational level and income level) of pig farmers influence ecological farming behavior of pig farmers through their ecological farming attitude and intention, but the degree of influence is limited.

6.2RecommendationsFirstly, it is recommended to strengthen propaganda and training of ecological farming model and technologies, and improve ecological farming cognition, attitude and behavior ability of pig farmers. Pig farming areas should energetically implement training of new farmers, especially new agricultural operating entities, to make more farmers learn and know new knowledge, technologies and skills of ecological farming, transform ecological farming ideas, attitude and intention, improve basic knowledge and technical level of ecological farming, and improve scientific pig farming quality. At the same time, pig farmers should positively take ecological farming decision after grasping certain farming technologies, knowledge and information. Secondly, it is recommended to improve government policy incentive and provide excellent regulation of external environment for development of pig industry. Government should improve animal husbandry environment management policies and regulations, reasonably optimize and increase subsidies for ecological farming behavior, formulate incentive policies, and use external regulations to encourage and guide pig farmers to take part in ecological farming. Improvement of policies and regulations can reinforce administration effort of government in ecological disruption, while ecological subsidies can encourage farmers to take part in ecological farming.

Finally, it is recommended to encourage pig farmers to develop large-scale farming. Developing large-scale farming is not only favorable for raising enthusiasm and increasing income of farmers, but also favorable for forming agriculture and animal husbandry combined ecological pig farming model, and favorable for reducing supervision costs of government, establishing ecological farming industry chain and realizing sustainable development of pig industry.

[1] SONG JP. The obstacles and countermeasure of sustainable development of agriculture and rural area in China[J]. China Soft Science, 2001(2):18-22.(in Chinese).

[2] CAO JJ. Analysis on the features and efficiency of several modern eco-agriculture models in Shandong Province[J]. China Soft Science, 2010,12:107-114.(in Chinese).

[3] QI ZH, WANG PC. Coupling mechanism of low-carbon agricultural ecological industry under game and interaction theory[J]. Forum on Science and Technology in China, 2010,11:136-141.(in Chinese).

[4]WU LY, QI ZH,etal. Countermeasures for internalization of environmental externalities in pig-raising industry——A case of floating dead pigs[J]. Research of Agricultural Modernization, 2013,06:694-697.(in Chinese).

[5] CHEN ZY, SHI DW. Development of the application of resource utilization technology in agricultural waste[J]. China Population,Resources and Environment, 2010,12:112-116. (in Chinese).

[6] LIN YF. Small farmers and economic rationality[J]. The Rural Economy and Society. 1988,03:31-33. (in Chinese).

[7] WU LY, QI ZH,etal. An empirical study on the influential of pig farmers to establish an eco-industrial chain[J]. Journal of China Agricultural University,2014,03:236-242. (in Chinese).

[8] ZHANG Y, QI ZH,etal. Study on the influence of family endowments on the environmental behavior of massive pig farmers under the situation of ecological compensation policy：Based on the survey of 248 massive pig farmers in Hubei Province[J]. Issues in Agricultural Economy, 2015(6):82-91, 112. (in Chinese).

[9] Ajzen I. Attitudes, traits and actions: Dispositional prediction of behavior in personality and social psychology[J]. Advances in Experimental Social Psychology, 1987, 20(1): 1-63.

[10] LIU XF, YANG ZH, WANG YP. Ecological cognition and behavior decision of poultry farmers: Based on field research of farmers in six provinces such as Shandong and Anhui[J]. China Population, Resources and Environment, 2013(10):169-176. (in Chinese).

[11] Kliebenstein, James B, Lawrence,etal. Contracting and vertical coordination in the United States Pork Industry[J]. American Journal of Agricultural Economics, 1995, 77(5): 1213-1218.

[12] Boger S. Quality and contractual choice: A transaction cost approach to the polish hog market[J]. European Review of Agricultural Economics, 2001, 28(3): 241-261.

[13] WANG HT,WANG K. Analysis on influence factor of pig farmers production safety decision-making behavior[J]. Chinese Rural Economy, 2012(11):21-30, 43.(in Chinese).

[14] Guagnano GA, Stern PC, Dietz T. Influences on attitude-behavior relationships: A natural experiment with curbside recycling[J]. Environment and Behavior, 1995, 27:699-718.

[15] Christine A, Ervin David E. Ervin. Factors affecting the use of soil conservation practices: Hypotheses, evidence, and policy implications[J]. Land Economics,1982,58(3):277-292.

[16] Feder G, Just RE, Zilberman D. Adoption of agricultural innovations in developing countries: A survey[J]. Economic Development Cultural Change,1985, 33(2): 255-298.

[17] Sinden JA, King DA. Land condition, crop productivity, and the adoption of soil conservation measures[C]. Working Paper Presented at the Australian Agricultural Economics Society Conference, Melbourne, Australia, 1988.

[18] TAO QS, HU H, WANG QJ. Analysis on the influence factors of farmers in agricultural new technology adopted will with environmental constraints[J]. Statistics and Decision, 2013(1):106-110.(in Chinese).

[19] RONG TS. AMOS and research methods[M]. Chongqing: Chongqing University Press, 2010(2): 118. (in Chinese).

December 16, 2015 Accepted: January 4, 2016

Supported by Project of National Natural Science Foundation (41171436); Key Project of National Social Science Foundation (14AZD002).

*Corresponding author. E-mail: qizhh@mail.hzau.edu.cn