.~.'.~£u- ? ‘7’:- ; . i1. , t. a .vlf.&huvptt’)a.io.u£niflpoz a . .. 1 fl .7 v.11. .35).. iv) “lit! 3‘ ’Q-h/n.'fl.»’€lrd ‘ M.» ~s.‘.xv. )3; no...) .3)! . i I... l:\}.fi 11 f... . 'IIBI 4455.... . .. i... .I. r .. 0. v... ‘ ire. [lappil - s7. rituf‘...¢. - azitin“ .3913}. 0931. . folrwi 3...... v 7v... Ivig .3273 ’t . 1. t9?T:I! IV? i=5. .5": Erin!!! 1.5 ,3!!! vyx I. .t‘ ‘ ~ . . 32.5.5 .3; . .».c.~t ... ~ It»)..5tb .....,l..) B it! 3.3 : .zu): {fir-ha!!! tilt-l... re‘4.)v VI...‘ .5 )1‘5...r!ir.{2 610779.23, tw.‘ 1.7.17: . . 19 (‘52 ‘35.? .155!" I.-. t: 0.18 331.2. (9090. I. 4|. ll?!.1!p§k3v. re. . ‘rifi .3?-.:$..51317W.iqf _ 31.19. ’1. Evil. '4 Coi. :0 ‘5‘} z): .7 ~43}... . . .t~.?\» - ,wz‘wfip; Elli—Malt? unfit-eke? .Th 1.3.5... $41.01;}... )1 3 £75.... 1 .. £33. . .9..Kpu-: ‘22:; .. . .2 u; .33.! £1 .1!» 10.7 5.3. .1. x v . 3: .135. 1 1.05:7»...7 ‘. a. x. thett :aezyh.’v.1 I!?..rrl:1ai¢. .1: r :1. .2): :2... $6 .9: ’ .x (n.1,: 5. v. . . 1:31. . 11 .1 I. . r. #5:?- u 1...; THhD-D umvsnsrvu. I IIIIIIIIIIIIIIII 300892 7380 IIIIIIIII: This is to certify that the thesis entitled THE RELATIONSHIP BETWEEN PLAN SIERRA OUTREACH ACTIVITIES AND THE ADOPTION AND CONTINUED USE OF SOIL AND WATER CONSERVATION TECHNOLOGIES BY UPLAND FARMERS presented by Michael Robotham has been accepted towards fulfillment of the requirements for MS . Resource Development degree in W Major professor Date July 27, 1993 0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY 1 Michigan State University PLACE lN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE {W00 5£§ [£13 3992 MSU is An Affirmative Action/Equal Opportunity institution cideWmS-o} THE RELATIONSHIP BETWEEN PLAN SIERRA OUTREACH ACTIVITIES AND THE ADOPTION AND CONTINUED USE OF SOIL AND WATER CONSERVATION TECHNOLOGIES B! UPLAND FARMERS BY Michael Robotham A THESIS Submitted to Michigan State University in partial fulfullment of the requirements for the degree of MASTER OF SCIENCE Department of Resource Development 1993 ABSTRACT THE RELATIONSHIP BETWEEN PLAN SIERRA OUTREACH ACTIVITIES AND THE ADOPTION AND CONTINUED USE OF SOIL AND WATER CONSERVATION,TECHNOLOGIES B! UPLAND FARMERS BY Michael Robotham Increasing land use intensity in tropical and subtropical upland areas and its far reaching consequences highlight the need for research into current upland development activities. This study focuses on one project, Plan Sierra in the Dominican Republic, which has promoted the use of soil and water conservation technologies by upland farmers since 1979. This research analyzed data collected from interviews with 82 farmers to determine the relationship between project outreach activities and the adoption and continued use of conservation practices. The information was also used to analyze the utility of grouping farmers by slope zone. The analysis identified a significant association between farmer interaction with Plan Sierra and the adoption of soil and water conservation farming practices. It also showed extremely low discontinuance rates for these practices. Analysis of the data also established that slope zone is not an appropriate criteria for farmer grouping. ACKNOWLEDGMENTS I would like to acknowledge the assistance and support I received from a number of people during my studies at Michigan State. First and foremost, Dr. Scott Witter, my major professor, provided invaluable assistance throughout my studies. This research could not have taken.place without his help and advice. The other members of my graduate committee, Dr. Michael Gold and Dr. George Axinn also provided valuable input as did a number of my fellow students. Special thanks go to Brent Simpson for his insight and suggestions and to Fidel Santos for his help with the necessary translations between English and Spanish. I owe an additional debt of gratitude to my colleagues in the Dominican Republic led by Dr. Domingo Carrasco at ISA and Licda. Inmaculada Adames and Alfredo Jimenez at Plan Sierra. I could.not have even attempted.this study without their help. Many thanks also go to the interviewers, Victor de la Paz, Junior Rincén, and Nicholas Villanueva, and to Augustin Rodriguez, our driver. Last and most important, I want to thank my wife, Betchie, for all her help and support throughout my studies and my infant son, Daniel, for adding joy and wonder to our lives. iii CHAPTER 1 1.1 CHAPTER 2 2.1 TABLE OF CONTENTS CONTEXT AND OBJECTIVES Background information . . . . . . . . . . . . 1 1.1.1 Worlwide situation in upland areas . 1 1.1.2 Situation in the Dominican Republic . 1 1.1.3 Description of the Plan Sierra project 3 Statement of the problem . . . . . . . . . . . 9 Study objectives . . . . . . . . . . . . . . . 10 Study hypotheses . . . . . . . . . . . . . . . 10 SELECTIVE REVIEW OF RELATED LITERATURE Introduction, diffusion, and adoption of innovations . . . . . . . . . . . . . . . . 14 2.1.1 Definition of terms . . . . . . . . . 15 2.1.2 Existing models . . . . . . . . . . . 16 2.1.3 Diffusion/adoption studies . . . . . 26 Integrated rural development and farming systems . . . . . . . . . . . . . . . . 32 2.2.1 Integrated rural development . . . . 33 2.2.2 Farming systems approach to development . . . . . . . . . ._. 36 Research related to recommendation domains . . 38 2.3.1 Systems of land classification based on slope . . . . . . . . . . . . . . 39 iv CHAPTER 3 STUDY METHODOLOGY 3.1 Advantages and disadvantages of methodologies . 3.7 3.1.1 3.1.2 3.1.3 Specific procedures used in this investigation 3.2.1 3.2.2 Characteristics of the survey instrument Cluster sampling Single interview survey . Face-to-face interviews Cluster sampling methodology Selection of study participants Interviewers . . . . . Data processing . . . . Definition and operationalization of variables 3.6.1 3.6.2 3.6.3 Concepts in hypothesis 1 Concepts in hypothesis 2 Concepts in hypotheses 3 and 4 Data analysis . . . . . 3.7.1 3.7.2 3.7.3 Nominal and ordinal data Interval and ratio data . Analysis methodology for 1b, and 1c hypotheses la, Analysis methodology for hypotheses 2a, 2b, and 2c Analysis methodology for hypothesis 3 Analysis methodology for hypothesis 4 41 41 41 42 43 46 47 48 49 50 50 53 54 59 60 61 62 62 63 65 CHAPTER 4 4.1 RESULTS AND DISCUSSION General description of farming system . . 4.1.1 . Physical factors . 4.1.2 Farm enterprises 4.1.3 Farm inputs and outputs 4.1.4 Farm family attributes Testing of hypothesis 1 4.2.1 Hypothesis 1a 4.2.2 Hypothesis 1b . 4.2.3 Hypothesis 1c . . . . . . . . . . 4.2.4 Summary of results for hypothesis Testing of hypothesis 2 . . . . . . . . . . 4.3.1 Hypotheses 2a, 2b, and 2c . 4.3.2 Summary of results for hypothesis Testing of hypothesis 3 . . . 4.4.1 Information from Plan Sierra . 4.4.2 Summary of results for hypothesis Testing of hypothesis 4 . . . . . . . . 4.5.1 Information from Plan Sierra . 4.5.2 Summary of results for hypothesis Possible problems with the data . 4.6.1 Testing for bias . . . . . . . . vi 66 66 67 68 69 70 70 71 72 74 76 76 76 79 79 79 81 81 81 84 84 CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS 5.1 Conclusions from this study . . . . . 5.1.1 5.1.3 5.1.4 5.1.5 5.1.6 .1. Assessment of Plan Sierra outreach activities . . . . . . Use of slope zones to delineate recommendation domains . . . Tests for bias . . . . . . General observations concerning Plan Sierra O . . O . O O . . 0 Summary of study conclusions . Generalizability of the results 5.2 Recommendations for future research . . 5.2.1 5.2.2 5.2.3 BIBLIOGRAPHY APPENDIX A APPENDIX B Research within Plan Sierra . Research outside Plan Sierra Final Recommendations . . . . . Detailed output from the various statistical analyses conducted for this study . . . . . . . . . . Copy of the survey instrument used in this study . . . . . . . . . vii 94 95 99 101 102 104 105 106 106 108 108 109 116 123 Table Table Table Table Table Table Table Table Table Table Table 10 11 LIST OF TABLES Variables used to test hypotheses 1 and 2 52 Variables used to test hypotheses 3 and 4 56-57 FAO slope classes (used to define the variable slclass) . . . . . . . . . . . . . . . 58 Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the receipt of information from Plan Sierra (pshear) . . . . . . . . . . . . 71 Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the interaction with Plan Sierra extension personnel (psextend) . . . . . . . 72 Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the participation in off-farm training activities conducted by Plan Sierra (patrain) . O O . O O O O O . O O O O O 7 3 Summary of the association between the continued use of soil and water conservation technologies (stilluse) and the various levels of interaction with Plan Sierra (pshoar, psextond, pstrain) . .p. . . . . . . . 78 Summary of hypothesis 3 test results . . . . 80 Summary of the results used to test hypothesis 4 . . . . . . . . . . . 83 Summary of the results from tests for Plan Sierra list bias . . . . . . . . . . . 86 Summary of the results from tests for Plan Sierra site bias . . . . . . . . . . . . . 91 viii Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 swcuse by pshaar for all farmers swcusa by pshear for farmers not on lists provided by Plan Sierra . swcuse by pshear for farmers in Las Piedras only . . . . . . . . . . swcuse by psaxtend for all farmers swcuse by psextend for farmers not on lists provided by Plan Sierra . . . . . . swcuse by psextend for farmers in Las Piedras only . . . . . . . . swcuse by pstrain for all farmers swcuse by pstrain for farmers not on lists provided by Plan Sierra . . swcuse by pstrain for farmers in Las Piedras only . . . . . . stilluse by pshear for all farmers . . stilluse by pshear for farmers not on lists provided by Plan Sierra . . . . . . . . stilluse by pshear for farmers in Las Piedras only . . . . . . . . . stilluse by psextend for all farmers . . . stilluse by psextend for farmers not on lists provided by Plan Sierra . . . . . stilluse by psextend for farmers in Las Piedras only . . . . . . . . . . . . stilluse by pstrain for all farmers . stilluse by pstrain for farmers not on lists provided by Plan Sierra . . . stilluse by pstrain for farmers in Las Piedras only . . . . . . . . . . . . . Summary of hypothesis 3 test results for individual variables . . . . . . . . ix 116 116 116 117 117 117 118 118 118 119 119 119 120 120 120 121 121 121 122 Figure 1 Figure 2 Figure 3 LIST OF FIGURES Location of the Plan Sierra project area . . . 4 Schematic description of the three general models of the innovation process . . . . . . . 18 Approximate boundaries of the three climate zones in Plan Sierra . . . . . . . . . 45 CHAPTER 1 CONTEXT AND OBJECTIVES 1.1 Background Information 1.1.1 werldwide situation in upland areas Throughout the countries of humid and subhumid tropics, concerns are being raised about past efforts in the field of agricultural development. Although the successes of the "Green Revolution" are many, they have only significantly affected onesmall segment of the diverse agroecosystems found in these countries, that of irrigable lowlands. Upland‘, rainfed areas have largely been neglected by both rural development agencies and agricultural planners (World Bank, 1992). Upland areas make up a considerable portion of the total land.area and support a significant and.increasing fraction of the population in many countries of the humid and subhumid tropics particularly in Asia, Central America, and the Caribbean Islands. 1.1.2 Situation in the Dominican Republic The poverty and poor living conditions experienced by many small farmers in upland areas of the developing world.and specifically in the Dominican Republic are readily apparent and well documented (de Janvry and Hecht, 1984; Hartshorn et. al., 1981; Retana, 1982; Santos, 1980). These same farmers . 1For the purposes of this study, the upland ecosystem is dEfined to include the undulating and steep lands which generally rAnge from sea level to approximately 1000m (from Garrity, 1991). 1 2 receive much of the blame for deforestation and accelerated soil erosion which exacerbates downstream problems such as flooding; siltation.of dams, and reduction.of dry season water flow (Hartshorn et. al., 1981; Plan Sierra, 1982; Retana, 1982; Santos, 1980). These same areas also provide necessary commodities for the population living in lowland areas of the country. This includes the provision of water to operate the hydroelectric power and irrigation facilities and of other inputs necessary for daily life (principally fuel in the form of wood and charcoal). Furthermore, approximately 80% of the food consumed in the Dominican Republic is grown by upland farmers (Carrasco, 1991). Poor management of upland watersheds, including, but not limited to, deforestation, improper road construction, and poor farming techniques, has resulted in extensive erosion, severe downstream :looding, siltation behind dams and in .coastal areas, and a variety of other problems (Tropical Research and Development Inc., 1992). Since downstream interests are generally more powerful than upland dwellers, problems with flooding in lowland areas and siltation behind dams have provided much of the impetus behind the design and implementation of a number of development programs in the Dominican Republic. Although they take a variety of approaches, the vast majority of these programs seek to halt or at least reduce ongoing deforestation and soil erosion. They also usually strive to increase farmer income and improve farmer livelihoods largely through preservation of soil fertility. A number of upland development activities have taken place including the Natural Resources Management Project (NARMA) financed by USAID in the 0coa Watershed (Carrasco, 1991, Kemph and Hernandez, 1987, Robotham, 1991), the.Adaptive Agricultural Research Project financed by the Dutch ministry for development cooperation (Mollema, 1988), and the Enda- Caribe Project (Witter, personal communication). This thesis Will focus on one specific ongoing upland development project directed by a non-governmental organization (NGO) known as Plan Sierra. 1.1.3 Description of the Plan Sierra project 1.1.3.1 Location The Plan Sierra project area is situated on the north slope of the Cordillera Central mountain range above the head of the Cibao valley (Figure 1). This valley is one of the main agricultural regions of the countryu The jproject originally encompassed an area of approximately 2,300 kmz, which is about five percent of the total area of the country (Santos and Quezada, c1979). Early in the project, the area of Bermudez National Park was removed from its jurisdiction resulting in a reduction in area to approximately 1,700 km? (Plan Sierra, 1982).' This area contains a range of climatic zones within upland and mountainous areas and intermontane valleys. AHm Q mama .ccmuom Ecumv mono uomnoum cuneam scam on» no coaucooq H musaflm i; ouocaEH on cmwwmz “mm“ any nuuucnnnu iiHiHi . 3.3545».— Ecumaau @ s accomumz 1.535 ® 09 <=zm>m4 3:50 «<2 r .v cucweoa AD . / Cup—mm hi. ‘5 . .. P'O W ’ Kl!) I It 's fl h iii;mv I . N . A x. U only! a. I I N . .. / 6‘. c 1.0 /I v. .0: J. k. w -/(...:. H25. . . H N 0 .~ ma uuoo cat‘s): ..... to... . o ..wa ca . ... cm: on .6 3.. . ~ . b .mw./LU:MLr\tmehMWWI, . if on: p a. I .00"! a. ounazcm Bonny mmoooua soeuc>occe one no maooos Hosanna cones ecu mo coaumwuomoc oeuoEooom m enamem a s o. 833.3 .333. neueflte a.:nco..e.e¢ ease: . co_.e~__eco_u:._uuc._. r» szms_ — co_ue:_mwm co_ue._eunc.p\ co..eu:ese_as. to» ace—a _ es sense as see e _ “was. . .u _< ea _ m ses.om .396 23:33 34.88 82:26 a going 1 iiiii iJ .839... ee.a.ao so.) ne_u.co_ca vce ._eou cn._aeunu _ . messna< . a.snco.ue.ex . guesses: . 5.333 32.32.. 3. fitoafifl o...- " u.aocme.o _ . . f iiiiii L _ nnmw- ” . co_.e~._eco_u:u.unc_ . 3:83 co_ae_.n.uc_ say? an to c a_e: nave-05%! C—Osh .806 5"“.0 Anae¢v car—m2; Oumuumcs Guns—Q>u a umOb "NO—85¢ 00.0: 8~Sb‘-O Au stone. sumac-m a usec.ocm ,ieesc. n>u_:cc. ac u_esem ou_>tem season; c .oeo a geese. ...uce_em u_nem ones .3. 3t 3 pzuxaodgua 5533.. 6.333 a _ 2:: i— co_ue:_e>u _I. :ess_..-4 w ........... m n ..... .J Emiiii uneceues< “ L oessaa< r .u eesana< . co.Mhuwwuc. ii iiiiiiii -iiiui . a eunflmnumwm.nhwm< . " “cacao.e>ea . . gaseous: . _e_uom r iiiiiiiiiiiiiiii L P iiiiiiiiiiii L p uuuuu L co..aoe< can co_m:u~_o «cacao.e>eo sot-ens: guesses: :— nee-um c. seesaw c. some.” do .oogum 19 to the diffusion process discussed in the context of the social interaction.model, this model includes the processes of both research and development of an innovation. This model has been widely used to describe and analyze the process of development and diffusion of innovations in a variety of fields. Examples exist in anthropology, general sociology, rural sociology, education, public health, communication, marketing, geography, and a variety of other fields (Rogers, 1983). In all cases, this model relies on the premise that an innovation, which has been developed and tested 'through research, will be Ibeneficial. Therefore efforts, usually extension, should be undertaken to speed the diffusion of the innovation to the appropriate group (e.g., farmers, educators, public health workers) (Mac Donald, 1976). It is important to differentiate between the research, development and diffusion model described by various authors (Havelock 1971; Mac Donald, 1976) and research which focuses on the diffusion and adoption of innovations (Feder et. al., 1985; Napier, 1991; Rogers, 1983). The major difference between the three models presented is in the area of new technology development. Each model assumes a different methodology for the development of an innovation; however, i in all cases, information about. the innovation must diffuse through the community and potential users must make a decision to adopt or reject the innovation. As a consequence, research into the processes of diffusion and adoption is important under all three models. However, most 20 of the research in these areas has been conducted regarding innovations which have been introduced from the outside, generally with the goal of improving the system. The bulk of this research has been directed toward agricultural innovations and generally employed similar techniques (survey interviews and statistical analysis) (Rogers, 1983) to those used in this studyu As a consequence, it is important to understand the models of both the diffusion and adoption processes, their components, advantages, and shortcomings. 2.1.2.4 Model of the diffusion and adoption processes Even though they are often discussed separately, the diffusion and adoption processes are interrelated. Within the many research traditions which have investigated diffusion and adoption behaviour, a number of common characteristics have been identified. Rogers (1983) provides the most widely respected overview of this subject. In his overview, Rogers identifies four main elements in the diffusion process: "(1) an innovation, (2) which is communiCated through certain channels, (3) over time, and (4) among the members of a social system” (p. 35). Using this framework, specific factors which are believed to affect the diffusion/adoption process have been determined. These include: the intrinsic and extrinsic characteristics of the innovation, the availability and use of channels of communication, characteristics of individual potential adopters, and characteristics of the existing social system. 21 This investigation focuses on the role of extension activities in the adoption process. As a consequence, channels of communication will be discussed here. Information on other characteristics which are believed to affect the diffusion/adoption process can be found in Rogers (1983). 2.1.2.4.1 Communication channels Several studies have shown that communication channels defined as “the means by which messages get from one individual to another" (Rogers, 1983, p. 17) play an important role in the diffusion/adoption process (Feder et. al, 1985; Rogers, 1983). These channels can be divided into two general types: mass media channels and interpersonal channels. Mass media channels include all the means of transferring information from one source to a large audience. Examples include radio, television, and newspapers. Interpersonal channels involve face-to-face exchange between two or more individuals. One group which commonly makes use of communication channels are change agents. Change agents are individuals who attempt to influence clients' innovation decisions in the direction of the change proposed by the group with whom the agent is involved. They use communication channels to spread information about the characteristics and benefits of the innovation. Extension workers are a traditional example of a change agent (Rogers, 1983). This investigation will focus on the relationship between farmer contact with change agents, as exemplified by Plan 22 Sierra personnel, and farmer use of the information provided by these change agents, as exemplified.by adoption of soil and water conservation technologies. 2.1.2.5 Advantages of the diffusion perspective Rogers (1983) cites four major advantages of the diffusion perspective. First of all, it is a general concept used across a variety of social science disciplines. Each discipline has a different perspective, but nearly all aspects of the social sciences are interested in the role played by innovations. Secondly, diffusion research appears. to Ibe able to provide clear answers and solve problems. Ideally it can identify the reasons why a new idea or technology is not being accepted and provide researchers with the knowledge necessary to facilitate acceptance in the future. Thirdly, the diffusion paradigm easily allows for the repackaging of empirical data into more theoretical generalizations. The paradigm also provides the basis for the comparison and contrasting a studies conducted in different geographical areas, at different time, concerning different innovations. Lastly, the research method defined by this paradigm is clear-cut and relatively easily employed. Data addressing these types of questions is not difficult to obtain, and generally accepted analytic methods exist. 23 2.1.2.6 Problems with the diffusion perspective Although the diffusion perspective has been used to analyze the adoption of innovations in.a‘wide variety of cases and will be used in this investigation, it is important to recognize some of the potential problems with this approach. The first problem is pro-innovation bias. The diffusion paradigm generally assumes that an innovation is’a good thing and should be diffused to and adopted by all members of the social system as rapidly as possible (Feder et. al, 1985; Roling, 1982, Rogers, 1983). As a consequence, diffusion research has tended to assume that adoption is the desired consequence for all potential adopters and ‘to focus on innovations ‘with. relatively' high. adoption. rates (Rogers, 1983). The reasons for this are partially historical, innovation research initially studied extremely beneficial agricultural innovations, so the reasons for non-adoption were virtually nonexistent. Other reasons include the fact that research is often funded by change agents (extension services, projects) who believe that the innovation they are promoting is a good thing and the fact that “successful" innovations leave a rate of adoption that can be studied after the fact. Largely “unsuccessful" innovations do not (Rogers, 1983). Closely related to this pro-innovation bias is individual blame bias. This is the tendency to hold the individual responsible for their problems rather than the larger society of which they are a part. Diffusion research, because of the 24 common pro-innovation bias mentioned above and its emphasis on individual choices, must take great care to not omit key social factors impacting adoption rates (Rogers, 1983; Roling, 1982) . A third related problem is the problem of equality in the diffusion process and by extension in diffusion research. There are many examples of conditions in which innovations have increased the gap between the higher and lower status groups in a society, especially in lesser developed countries (Feder et. al., 1985; Rogers, 1983; Roling, 1982). Although the diffusion paradigm does suffer from the three biases discussed above, it is possible to minimize these biases in the course of diffusion research. Authors (Feder et. al., 1985; Rogers, 1983) suggest that more research be explicitly directed toward the processes of and reasons behind rejection and discontinuance of innovations and toward the consequences of innovations. In addition, all diffusion research should attempt to determine the social and political factors having significant effects on adoption behaviour. Roling (1982) stresses the need to investigate the impact of farmers' differential access to land, water, labor, inputs, markets, capital, and information on the adoption of innovations along with the more traditional investigations of individual farmer attitudes toward change. Two additional shortcomings of diffusion research are discussed by Feder and his colleagues (1985). These are: the tendency to view adoption in dichotomous terms (adoption vs 25 non-adoption) and the lack of investigation into cases where a number of related innovations are introduced relatively simultaneously. In the first case, they note that many innovations can easily be adopted gradually by farmers. Thus, it becomes necessary to investigate the factors which may affect the level of adoption not just the use or non-use of a particular innovation” With respect to the second area, numerous examples exist where innovations have been introduced as a “package"; high-yielding crop varieties, inorganic fertilizer, and small-scale agricultural machinery for example. Very little work has been done regarding possible complementarity or conflict between adoption of various components and regarding selective adoption of specific parts of the "package“. 2.1.2.6.1 Problems with diffusion research methodology In addition to the potential problems with the diffusion paradigm, Rogers (1983) discusses two basic problems with the common methodologies used in diffusion research. The first of these is the recall problem. Because diffusion research often asks adopters or potential adopters to remember the factors which have influenced their past decisions, it is subject to variations inherent in different individuals power of recall. If diffusion research is conducted in a one-shot survey, as is traditionally done, very little can be safely concluded about the time dimension of adoption. 26 Closely related to this is the problem of causality. Diffusion research, as currently practiced in the vast majority of cases using one-shot surveys and correlation analysis, does not allow for the determination of causality. Rogers (1983) sees both of these as shortcomings of past diffusion research.and suggests greater use of time series and repeat interview studies to better understand the time component of diffusion and adoption and of field-experiment- type methodologies to determine such cause and effect relationships. 2.1.3 Diffusion/adoption studies The previous sections have provided a general overview of the adoption process. Several general models were presented providing a strong basis for subsequent analysis. Since this thesis investigation concerns the adoption of soil and water conservation technologies, this section will provide an overview of relevant work conducted in this area. 2.1.3.1 Agricultural innovations There is a large body of research related to the adoption of agriCultural innovations. Feder and his colleagues (1985) provide an extensive review of this area. In this review, they identify six major factors affecting the adoption process. They are: farm size, risk and uncertainty, human capital, labor availability, credit, and land tenure. In this typology, information acquisition is seen as a method whereby farmers can reduce uncertainty regarding' an innovation. Similar factors have also been discussed in relation to the 27 adoption of agroforestry innovations (Raintree, 1983, Repollo and Castillo, 1989). Mercer, (1992) observes that all of these factors, with the exception of risk and uncertainty are usually strong correlates with income. 2.1.3.2 Soil and water conservation Within the larger body of work regarding agricultural innovations, there is a growing amount c research regarding the adoption of soil and water conservation technologies. Much.of this research has been conducted in the United States, but an increasing amount is available from the less developed countries. Since this investigation focuses on the provision of information through extension activities in the adoption of soil and water conservation technologies, this review will focus on studies which discuss this area. 2.1.3.2.1 Adoption in the United States A significant amount of research has been conducted on the adoption of soil conservation practices in the United States in recent years. Much of this has been an outgrowth of increased concern regarding the impact of farming activities, including erosive land management practices, on water quality. Many of the studies have looked at the role of extension activities in the adoption of soil and water conservation practices; however, conclusions about their impact are mixed. In their review article, Christensen and Norris (1983) cite studies from the Great Lakes Region, Idaho, and.Missouri which show a lack of farmer knowledge concerning present erosion problems and_ the implementation of conservation 28 technologies; However, they conclude that education programs designed merely to provide general information are likely to have only limited effects. There are two reasons for this conclusion. First of all, studies show a wide variety of farmer perceptions and attitudes must be addressed by any program. Secondly, a number of other personal, economic, and social factors appear to affect an individual farmer's decision to adopt conservation practices. Ervin and Ervin, (1982) also discuss the impact of extension.efforts on farmer adoption in a Missouri watershed. However, they found no significant relationship between participation in Soil Conservation Service (SCS) programs and adoption of soil conservation practices. They suggest that perhaps some measure of the level of recent participation might have proven more discriminating. A similar study by Hoover and Wiitala in Nebraska (cited in Ervin and Ervin, 1982) also found a non-significant relationship between participation in SCS programs and adoption of conservation practices. . 2.1.3.2.2 Adoption in the lesser developed countries An growing’ body of research exists describing and evaluating soil and. water conservation programs in the countries of the lesser developed world. This literature ranges from discussions on physical methods of erosion control, to economic considerations, to studies of adoption behaviouru Most studies consider at least two of these three 29 aspects. 'This review will focus on research addressing adeption behaviour. In the vast majority of these studies, the farmer's lack of knowledge is cited as a constraint to adoption. Farmers may lack knowledge about the causes and consequences of soil erosion, the means available to resolve erosion problems, or both. (Anderson.and.Thampapillai, 1990; Blustain, 1982; Hauck, 1985; Moldenhauer and Hudson, 1988, Napier et. al., 1991). Lack of knowledge is generally not seen as the only constraint to adoption. In most of the above studies, economic and institutional factors are believed to have stronger effects on adoption. However, only a fraction of these investigations attempt to systematically determine the relative importance of these various factors (Hansen et. al., 1987; Laquihon, 1989). Dr. Hansen's study was conducted in the Dominican Republic and will be elaborated on in the next section. Laquihon's study attempted to determine if the relationships between adoption of a particular set of soil and water conservation farming technologies were correlated to a number of the variables theoretically believed to affect adoption. He presented farmers with 20 different likely determinants of practice adoption. They were asked whether or not each of these factors had played a part in their decision. From this data, he was able to identify 7 key determinants of SALT (Sloping’Agricultural Land Technology, a specific set of soil and water conservation practices) adoption. These were: 30 training and exposure to the technology, anticipated.benefits from the technology, seed availability, sincerity and capability of the sponsoring organization, source of outside farm income, and market needs (Laquihon, 1989, p. 96). Upon classification of these factors as either socio- cultural, physical, or biotic, it became apparent that socio- cultural factors played a predominant role in practice adoption. In addition to an analysis of key adoption determinants, Laquihon also performed correlation analysis between practice adoption and a number of common farming system characteristics. He identified three as being significantly correlated. These were: educational attainment, land topography, and number of animals raised. There are some potential problems with this study which must be considered. The first of these is the pro-innovation bias common to.many diffusion studies. Dr. Laquihon.works for the organization promoting the use of SALT practices. Secondly, the study contains the possibility of pro- organization bias. .All of the farmers interviewed. had received information and training regarding the use of the practices. In addition, it is highly unlikely that they were unaware of Dr. Laquihon’s affiliation with the organization. Within this context, it is not surprising that training was seen as a very important factor in the adoption of the technology. These potential problems do not, however, invalidate the conclusions of the study. 31 2.1.3.2.3 Adoption in the Dominican Republic Very little systematic investigation into the adoption of soil and water conservation technologies in the Dominican Republic has occurred. Information is only available from one study. This study was conducted by Mark Erbaugh for his Master's degree research at the Ohio State University under the supervision of Dr. David Hansen (Hansen et. al., 1987). The research was done in 1983, in the Ocoa watershed. In the Erbaugh and.Hansen study, a formal survey'was used to collect data on a number of variables 'related to the farming system and farmer adoption of soil and water conservation practices. These variables were divided into three clusters: socioeconomic (land size, income, literacy), access (extension contact, credit), and attitudinal (propensity to adopt, orientation to change). They were then compared with the adoption of soil conservation practices. The researchers found that socioeconomic factors were poor predictors of adoption behaviour, while the access and attitudinal variables showed significant correlations with practice adoption. Multivariate analysis was also used and showed that extension contacts and use of credit explained 29.2% of the variance in adoption, while attitudinal measures exPlained 29.6%. This study concluded that the use of agricultural support services and farmer attitudes appear to be the primary factors influencing farmer adoption of soil conservation practices. 32 There appear to be few potential problems with the data from this study; however, two possible areas exist. First of all, the vast majority of the farmers interviewed for the study were members of existing farmers associations. It is possible that this group was not representative of all farmers in the area. Secondly, the authors constructed several summation indices in the course of their analysis. Unfortunately, the information available indicates neither the number of questions used to create each index or the exact wording of the questions used. As a consequence, the validity of the methodology cannot be conclusively determined. However, the potential problems are expressed here do not invalidate the conclusions of the study. 2.2 Integrated rural development and farming systems Diffusion research, which was described in detail in previous sections, is seen by some authors (G. Axinn, 1988; Chambers, 1992; Flora, 1988) to be part of the outdated paradigm of technology transfer in agriculture. This paradigm assumes that the large-scale, mono-crop, capital intensive agriculture developed in the United States can and should be introduced into the agriculture of the lesser developed countries. If this occurs and the technology is adopted, people in those countries will be better off. What this paradigm often fails to consider are any of the attributes of the existing biological, social, and political situation in the place where the technology is to be introduced. In most cases, the existing conditions in an area 33 are much different from those in the United States. Farms are usually small-scale, mixed crop and livestock systems which often have a surplus of labor but little capital (G. Axinn, 1988). This situation is directly related to the pro-innovation, individual blame and equality biases discussed concerning diffusion research. When technology is considered to be highly beneficial by those who are promoting it, lack of consideration of the other important factors involved can result in completely inaccurate conclusions. In an attempt to overcome these biases and to systematically consider the wide range of conditions which can affecting “development" efforts, two approaches have been promoted. These are the integrated rural development approach and the farming systems development approach. Since the Plan Sierra project is an integrated rural development project which is committed by charter to a farming systems approach, it is important. tor examine these two important approaches to development. 2.2.1 Integrated rural development Integrated rural development can be defined as "a multi- dimensional strategy for improving the quality of life for rural people“ (Charlton, 1984, p. 179). An extensive body of literature exists covering a variety of subjects concerning the process of integrated rural development and the evaluation of various rural development projects. A very useful summary is provided by Hondale and VanSant (1985). 34 From these two analyses, a number of characteristics of integrated rural development (IRD) are apparent. First of all, IRD explicitly considers the interrelationships between all the factors contributing to the life-situations of the project participants (Charlton, 1984). This is in direct contrast to many of the initial projects in rural areas that were oriented toward one specific goal such as irrigation. .As a consequence of this consideration, IRD efforts nearly always consist of a combination of an agricultural production component with other, service related components (Hondale and VanSant, 1985). Secondly, IRD efforts usually encourage some level of participation by the project participants. The level of this participation is directly dependent on the managers of the project. Some researchers believe that project participants lack both the means and the skill necessary to transfer their hopes and aspirations into concrete activities (Williams, cited in Charlton, 1984). Others (Axinn, 1978; Chambers, 1983) express the belief that the role of outsiders is to facilitate development by local populations, not to direct development for them. There are also a number of general characteristics related to the context in which IRD efforts are conducted. First, IRD projects are often located near international borders, often for political or security reasons. Second, IRD projects can create significant differences in market potential between project participants and non-participants. 35 This is particularly' common. in. projects addressing food production. Third, IRD projects are often situated in areas dissatisfied with the current national government. This can exacerbate conflict.between.project.implementors and.the local population. Fourth, IRD projects can significantly alter the local leadership structure by creating or imposing an alternative structure related to the project. Fifth, IRD projects are often used as a proxy for or a first step in decentralization efforts. .As a consequence, project objectives may be of secondary importance to the development of management skills in rural areas. Finally, IRD projects are usually administratively complex. This can impose heavy requirements on the project staff to integrate the efforts of the project itself and the many line agencies often involved (Hondale and VanSant, 1985). The authors note that the above characteristic are usually interrelated. At times, multiple project objectives can be mutually supportive. In many cases, however, these multiple project objectives are contradictory (Hondale and VanSant, 1985). The success record of integrated rural development projects is not particularly impressive. In general, pilot projects in small areas have been successful, but attempts to expand these projects on a provincial or regional scale have seldom succeeded. Ruttan (1984) attributes this failure to two factors. First of all, small projects often have a jproportionally much higher level of human resources available 36 for all facets of their operation. Secondly, small projects can be much more easily adapted to local environments, both physical and socials These two characteristics are often lost when projects are'expanded to a larger scale. Hondale and VanSant (1985) attribute the general lack of IRD project success to a difference between rhetoric and resources. Projects with the goals of long-term success through local participation have often been given short-term budgets and specific, outsider-determined, objectives. 2.2.2 Farming systems approach to development The farming systems approach to development has been another response to the failings of the technology transfer model of agricultural development. Before characterizing this approach, it is important to define exactly’what is meant by a farming system. For the purposes of this investigation, a farming system.is defined as "a unique and reasonably stable arrangement of farming enterprises that the farm household manages according to well-defined practices in response to the physical, biological, and socioeconomic environments and in accordance with the household's goals, preferences, and resources“ (Shaner et. al., 1982, p. 64). A number of important ideas are highlighted in this definition. First of all, the farming system is a unique and reasonable stable situation. This reasonable level of stability is a necessary prerequisite for any attempts to define the system. Secondly, the system is composed of’a set of inter-related enterprises. Crops and livestock are 37 generally seen as the most important of these (FAO, 1989; Friedrich, 1992; Shaner et. al., 1982). Thirdly, the coordinating role of the farm household is highlighted. A number of aspects of the farm household are believed to be importante These include farmer characteristics and attitudes; and household resources including land, labor, and capital (Shaner et. al., 1982). Shaner and his colleagues include off-farm activities by household members within the boundaries of the farm household sub—system. Others (FAO, 1989, Friedrich and Hall, 1992) divide household activities into on-farm and off-farm components. The definition also stresses that a farming system exists within and responds to the physical, biological, and socioeconomic environments. The importance of context is also stressed.by'a number of other authors (Brush and.Turner, 1987; FAO, 1989; Friedrich and Hall, 1992). This farming systems development approach is similar to the IRD approach in that it stresses the integrated aspects of rural life. However, as is apparent from the previous discussion, integration is stressed on different levels in each approach. IRD stresses the importance of integration at a project level, combining programs focused on agriculture with those focused on other areas such as health, education, and marketing. In contrast, farming systems development stresses integration as a fundamental property of all aspects of rural life beginning at the level of the individual farm household. 38 2.3‘ Research related to recommendation domains The second goal of this investigation is to assess the feasibility of using slope zone as a basis for the determination of recommendation domains. A number of attributes have been proposed by which farm households can be aggregated. These fall into three general groups: natural factors (climate, soil, topography), historical factors (food preferences, present technology, tenure) and institutional and economic factors (access to markets, inputs). Natural factors are a commonly used method of division, but it must be remembered that economic and institutional factors can be very important (Collinson, 1982; Gilbert et. al., 1980). Other authors also support the need for aggregation of farmers to permit meaningful analysis (FAO, 1989, Ravnborg, 1992). A. number of potential problems do exist ‘with this approach. Ravnborg (1992) presents several of these. First of all, the number of characteristics which could be used to define a recommendation domain is almost unlimited. As a consequence, descriptions can easily become excessively detailed. However, if only a small number of variables are used to delineate a group of farmers, any variation within that group, which may be considerable, is lost. Thirdly, if a limited number of specific criteria are used, their relative importance in the identification process is often not explicitly stated. 39 A related problem also comes from the use of non-site- specific criteria for group delineation. What is an defining property in one region may not play an important role in another. Lastly, she also notes that many of the traditional criteria for outlining recommendation domains including use of inputs, use of machinery, and the presence and types of livestock are often oriented toward resource-rich farmers and tend group resource-poor farmers into one homogenous unit. In spite of the above limitations, authors (Collinson, 1982; Gilbert et. al., 1980; Ravnborg, 1992) are in common agreement that some level of aggregation is always necessary for farming systems research to be meaningful and for any level of technology development and diffusion to take place. However, such aggregation must rely on criteria which are chosen.deliberatelyfl Since one of the primary focuses of Plan Sierra has been on technologies to reduce soil erosion, they have used slope as a major criteria for determining recommendation domains. 2.3.1 Systems of land classification based on slope There is a long history of classifying land according to slope and basing land use recommendations on this classification. Slope is a significant component of the USDA Land Capability Classification system in the United States (Mokma et. al., 1983). This system has been widely used and adapted worldwide and the principles behind it are sound. However, it has often been used without consideration of the differences between conditions in the country of use and.those 40 in the United States, where the system was developed (Sheng, 1989). .As a consequence, Sheng (1989) proposes a "treatment- oriented" land capability classification system for use primarily in tropical countries. Variations of this system have been used successfully in Jamaica (Sheng, 1988), El Salvador, Honduras, and Thailand and is it perceived by some researchers (Hudson, 1977 and 1983) as an alternative to the USDA classification system in these and other similar areas. Other land classification systems based largely on slope have been.developed.in.Australia.(Watkins, 1991), Taiwan (Liao et. al., 1991; Lin, 1991), North Thailand (Rimwanich, 1991), Korea (Jo, 1991), and the Dominican Republic (Veloz et. al., 1985). Since slope gradient is a strong determinant of soil erosion (Hudson, 1983; Sheng, 1989) and this study focuses on the adoption.of soil andnwater conservation.practices designed to control erosion, slope appears to be an appropriate way to aggregate farmers. If the study shows slope to be an accurate determinant of farming system, these recommendation domains have the potential for further application within and outside of the Dominican Republic. They can then be used to assist in program design and implementation and in policy formation. CHAPTER 3 STUD! METHODOLOGY In order to obtain the information necessary to test the hypotheses of this study, a modified cluster sampling method- ology was use to select a sample of farmers from the total population of farmers in the Plan Sierra area. A single, structured, face-to-face interview was conducted with each farmer participant by one of three Dominican interviewers trained by the researcher. These interviewers had no professional or casual affiliation with Plan Sierra. 3.1 Advantages and disadvantages of methodologies 3.1.1 Cluster sampling There are a number of advantages to the use of a cluster sampling methodology. They include: 1) reduced costs of data collection and 2) increased ability to obtain a complete list of the target population. However, there are also significant disadvantages, principally the loss of precision due to the combination of sampling error at both the cluster selection and sample selection stages (Singleton et. al., 1988). 3.1.2 Single interview survey Use of a single interview survey also necessitates consideration of the positive and negative aspects to such an approach" The principle advantage of all types of surveys is that they can provide detailed, systematic descriptions of populations. They can also address a wide range of research topics (Singleton et. al., 1988). Single interview surveys are also the less costly when compared with other formal 41 42 methods. Such a methodology is most appropriate for gathering data on phenomena that change slowly (e.g., farm size, family size), are one time or infrequent events (e.g., dates of planting, purchases of inputs), or deal with information about farmer knowledge, beliefs, and attitudes (Shaner et. al., 1982) Unfortunately, there are also significant limitations to the use of a single interview survey methodology. First of all, survey ‘data can generally be used only to determine association between variables not to establish direct cause- and-effect relationships. Secondly, once a study has begun, surveys are not easily changed. Thirdly, surveys deal almost exclusively with reported behaviour and do not provide the opportunity to extensively examine the context in which the behaviour occurs (Shaner et. al., 1982, Singleton et. al., 1988). In addition, a single interview is often an ineffective way to gather information on sensitive topics (e.g., income) since the interviewer has little chance to build rapport with the farmer (Shaner et. a1, 1982). 3.1.3 Face-to-face interviews There are also advantages and disadvantages associated with the use of face-to-face interviews. The principal advantages of this method include: 1) generally high response rates when compared to other survey methodologies particularly for longer surveys, 2) permitting interviewers to make unobtrusive observations that may be of interest to the researcher (Singleton _et. al., 1988), and 3) allowing the 43 collection of data from poor populations who lack the formal education necessary to complete another type of survey. The major disadvantages of face-to-face interviews revolve around issues of bias. In such a method, the interviewer may inadvertently introduce bias into the data in a number of ways including, but not limited to: failing to follow 'the interview' schedule exactly' as prescribed. and suggesting answers to the respondent. In addition, bias may be introduced as a result of the respondent's reaction to the presence of the interviewer and various personal characteris- tics of both parties. The higher costs of interviewer training and the conduct of the interviews themselves consti- tute another disadvantage of this methodology (Singleton et. al., 1988). 3.2 Specific procedures used in this investigation 3.2.1 Cluster sampling methodology The choice of a cluster sampling methodology for this study was largely due to considerations of cost, including the amount of time available, and the relative inaccessibility of the study area. In addition, due to the lack of systematic population data from the rural Dominican Republic, it would have been virtually impossible to obtain the complete list of residents necessary for the use of random sample selection. As mentioned earlier, Plan Sierra has developed a system whereby the total geographic extent of the project is divided into three broad ecological zones: the humid highlands (Zona Humeda), the transition uplands (Zona Transicidn) and the dry 44 lowlands (Zona Seca) (Altieri, 1990, Jimenez and Gutierrez, 1993). The approximated extent of these zones is shown in Figure 3. All three zones were surveyed as part of a larger, ongoing study. However, for the purposes of this investigat- ion, work was confined to the Humid zone. The originally proposed sampling methodology involved the use of a geographic information system to process landuse maps drawn from.aerial photographyu This allowed the researcher to identify and locate agricultural areas. Once these areas had been determined, a random number table was used to select two of these areas in each zone for study. The initial areas selected in the Humid Zone were the villages of Las Piedras and El Gallo. However, during discussions with Plan Sierra personnel, it became apparent that this design was unaccept- able to them. In their opinion, the large size of the project area creates a significant chance that only communities in which Plan Sierra had done little or no work would be selected. .As a consequence, Plan Sierra provided the researcher with a list of communities in the humid zone where they were working. This list, which consisted of seven communities, appears to have been incomplete. Other discussions with persons associated.with.Plan Sierra indicated that at least 50 communities were involved in Plan Sierra activities. Since the humid zone occupies approximately one-third of the project area and much of the most steeply sloping land is found in 45 33 .Q .emmH .scoHoooom Bonn ooumcoov ounoflm scam ca meson ouceflao women one we moflucocsoo ouceflxoummm 5m. 2:5 E... .3 3:82 89 8.358 28 Go. 9% 3 8.232. m as. :0 women 0-1 q 0—4 NZON >¢Q MZON onHHmz<¢H mzou ants: mooc_=>\m:38. 3.503252 .32: m 1.....3 . . - - 2:3 23. 823... .gggc 338 ad .3 m uneven 46 this zone, it seems unlikely that soil and water conservation activities would be under way in only seven communities. After consideration of the probable bias involved in using the available list, the researcher decided that useful data could still be obtained. As a consequence, a random number table was used to select two communities from the list: El Gallo and Rincdn Largo. As a control, interviews were also conducted in the Las Piedras site which had been chosen under the original study design. 3.2.2 Selection of study participants Due to the constraints inherent in research conducted in rural areas of the lesser developed countries and in this particular investigation, obtaining a complete list of residents in any of the three communities selected for this investigation was not feasible. As a result, a true random sample of residents could not be determined. However, the researcher believes that a relatively representative sample was obtained. This belief is supported by Dr. Domingo Carrasco, Assistant Rector at ISA, who was closely involved with the project. In Las Piedras, interview subjects were identified using a pseudo-random selection procedure. Each of the three interviewers left the center of the village in a different direction. They were instructed to interview the person who identified themselves as the head of the household, regardless of gender, at every farm they located.where someone was home. From discussions with the interviewers, it appears that the 47 majority of the area located.near the community of Las Piedras was covered.over the course of the interview'process. A.total of 42 persons were interviewed over the course of two days in this area. In the communities of El Gallo and Rincén Largo, Plan Sierra personnel insisted that the investigation include 10 persons selected from a list of persons who had worked with the project“ Project technicians in.both.areas provided lists of 16 participants. Ten of these were selected.using’a random number generator. The interviewers were instructed to interview one person from the list and then interview their nearest neighbor, as long as that person was not also on the list. This resulted in a total of 20 interviews in each of these areas. 3.3 Characteristics of the survey instrument The initial survey instrument was designed and written by Dr. Scott G. Witter, the researcher's major professor as part of an ongoing USAID consultancy in the Dominican Republic. This original survey design was reviewed by colleagues at Michigan State University, the Universidad Catholica Madre y .Maestra (UCMM), the Instituto Superior de Agriculture (ISA), both in Santiago and by Plan Sierra officials. This design ‘was pretested by Dr. Domingo Carrasco of ISA during April, 1992. After a review of the pretest results and recommendations, the researcher made significant additions and alterations to the initial design to accommodate the research 48 goals of this study. The initial modifications to the original design were made by Dr. Karen Potter—Witter from MSU and Dr. Efrain Laureano from USAID, Dominican Republic. Final modifications were made by Mr. Fidel Santos, a bilingual graduate student at MSU. Before the survey was administered in the field, it was edited and revised by colleagues at ISA and at Plan Sierra. It was also pretested with farmers located near the ISA campus. Minor changes were made in the wording of some questions to bring them in line with the Spanish spoken by rural farmers. This survey was the first activity in a larger, ongoing research effort in the Plan Sierra area. As a consequence, it covered a wide range of topics. The researcher trained the interviewers and directed the administration of the general survey consisting of approximately 100 questions but attempted to focus his attention on the specific questions reported in this study. .As a result, only a portion of the data collected will actually be referred to in the subsequent analysis. An English translation of the entire questionnaire can be found in Appendix B. The questions which were specifically used in this analysis are presented in boldface type. 3.4 Interviewers The researcher used three persons to conduct the interviews necessary for this investigation. All three were male students at ISA.who were in either their Junior or Senior year of study. Two were pursuing bachelors degrees in Animal Science. The third was pursuing an associates degree in 49 Agronomy. ' These men were recommended by Dr. Carrasco and other ISA staff and had previous interview experience. The researcher provided additional training regarding the specifics of- this study and this particular survey instrument. The interviewers worked separately and generally completed six or seven interviews per day. The researcher accompanied one of the interviewers during each day of the process in order to obtain a better idea of any possible problems related to the survey and to gain first hand knowledge of the farming systems in the various communities. All interviews were conducted face-to-face and were generally conducted in or near the subject's home. A small number were conducted in farmer's fields during a break in their work. Interviews were conducted.with the self-selected head of the household or, if this person was not available, they were conducted with another adult who appeared to be knowledgeable about farming activities. Farmers were given the opportunity to refuse to be interviewed. If they gave their permission to be interviewed, they were assured that they would not be identified in any way during the analysis and reporting of the results. They were also told that they were free to refuse to answer any survey questions they felt were inappropriate. 3.5 Data processing' The questionnaires were collected.by the researcher each evening upon returning to the university from the field. Preliminary checks were made to insure that the interviewers 50 had indeed conducted the work, to identify problems with specific questions or interviewers, and to answer any questions the researcher had about participant responses. Daily’discussions with.the interviewers were also conducted.by the researcher to detect any problems or difficulties with the data collection. All the interviews conducted for this study were conducted in March, 1993, during the researcher's stay in the Dominican Republic. ‘Upon his return to the United States, the survey results were coded and checked by the researcher with the assistance of Mr. Fidel Santos, a fellow graduate student. .3.6 Definition and operationalization of variables In order to test the hypotheses discussed in Chapter 1, it is necessary to define the concepts outlined in the hypotheses and to operationalize these concepts into variables for subsequent analysis. 3.6.1 Concepts in hypothesis 1 As outlined previously, hypothesis 1 is subdivided into three parts, these are restated below for the convenience of the reader: HIPOTHESIS 1: Hypothesis 1a: Ho: Farmers who have received information from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this information. H1: Farmers who have received information from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this information. 51 Hypothesis'lb: Ho: Farmers who have received extension advice from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this advice. H1: Farmers who have received extension advice from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this advice. Hypothesis 1c: Ho: Farmers who have received off-farm training from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this training. H1: Farmers who have received off-farm training from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this training. These divisions reflect the three levels of increasingly intense interaction with Plan Sierra discussed earlier. The first of these levels, receipt of information from Plan Sierra in any form regarding soil and water conservation, is operationalized as the variable pshear. This variable is determined from the responses to question 27 and question 28. The second of these levels, the receipt of extension advice from Plan Sierra extension agents regarding soil and water conservation practices, is operationalized as 'the variable psextend. This variable is determined from the responses to question 33 and questibn 46. The third level of interaction, the participation by the farmer in a 'training session concerning' soil and. water conservation practices at the Plan Sierra training center in 52 Los Montones, is operationalized as the variable pstrain. This variable is determined from response to question 33. The second concept in hypothesis 1 is adoption of soil and water conservation practices. This concept is defined as the reported use at any time in the past of any practice which the farmer defines as a soil and water conservation practice. This is operationalized as the variable swcuse which is determined from the response to question 29. Details of the definition and derivation of all of these variables are shown in Table 1. Table 1: Variables used to test hypotheses 1 and 2 Variable Explanation Derivation Type Name swcuse Farmer has ever used soil and svcuse 8 “yes” if 029 a “yes“ else Nominal water conservation practices swcuse a ”no" pshear Farmer has received information pshear I “yes“ if 027-”4' and "Plan Nominal from Plan Sierra regarding soil Sierra“, GZBI'4" and “Plan Sierra“, and water conservation practices psextend I "yes", or pstrain I "yes“; else pshear - "no" psextend Farmer has received extension psextend = “yes'I if 033="3" and Nominal advice regarding soil and water "Plan Sierra" or 046-“4” and "Plan conservation practices from Plan Sierra“ else psextend I "no” Sierra extension agents pstrain Farmer has participated in a pstrain 8 “yes" if 033="4" and Nominal training session at the Plan "Plan Sierra" or “Los Nontones" else Sierra training center in Los pstrain I ”no" Nontones stilluse Farmer currently uses soil and Answer to 039 Nominal water conservation practices -=—————= 53 3.6.2 Concepts in hypothesis 2 Hypothesis 2 isidivided.into three components in the same manner as hypothesis 1. These hypotheses are restated below: HYPOTHESIS 2: Hypothesis 2a: Ho: Farmers who have received information from Plan Sierra are no more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this information. H1: Farmers who have received information from Plan Sierra are more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this information. Hypothesis 2b: Ho: Farmers who have received extension advice from Plan Sierra are no more likely to have continued to use soil and.water conservation practices on their farms than are farmers who have not received this advice. H1: Farmers who have received extension advice from Plan Sierra are more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this advice. Hypothesis 2c: Ho: Farmers who have received off-fans training from Plan Sierra are no more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this training. H1: Farmers who have received off-farm training from Plan Sierra are more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this training. The initial concept in this hypothesis, interaction with Plan Sierra is defined and operationalized in the same manner as in hypothesis 1. The second concept is the continued use of soil and water conservation practices. This is defined as the reported use of any practice which the farmer defines as ‘1 54 soil and water conservation at the present time. This is operationalized as the variable stilluse which is determined from the response to question 39. Details of the definition and derivation of this variable are also shown in Table 1. 3.6.2 Concepts in hypotheses 3 and 4 Since the same concepts are used.in both hypothesis 3 and hypothesis 4, they will be discussed together. The exact wording if these hypotheses is restated here: HIPOTHESIS 3‘ Ho: The farming systems found within an individual slope zone are not significantly different from each other. H1: The farming systems found within an individual slope zone are significantly different from each other. HYPOTHESIS 4 Ho: The farming systems found in different slope zones are significantly different from each other. H1: The farming systems found in different slope zones are not significantly different from each other The principle concept used in the formulation and testing of hypotheses 3 and 4 is the concept of a farming system. As discussed in Chapter 2, there have been a number of attempts to adequately define what is meant by a farming system and what information is necessary to formulate an accurate description of such a system. In spite of differences in these definitions and characterizations, most authors seem to agree that farming systems can be adequately described using three general classesiof information: 1) information.about the farm household, 2) information about farm enterprises, and 3) 55 infermation about linkages between the farm and the outside social, political, and environmental systems. As a consequence, the researcher has identified 31 variables which address various components of the farming system. and allow the development of a relatively accurate characterization of the predominant system.or systems in this area. These variables, their definitions, and their derivations can be found in Table 2. As is apparent from the table, the variables are divided into several groups. The first two variables, slclass and farmsise, provide physical information on the farm, The next seven variables, crdivers through crpratio, provide information. on the crop cultivation subsystem.*while the subsequent six, andivers through anmratio, provide similar information for the animal subsystem. The next five variables, labor through credit, provide information concerning a number of common material inputs to the farming system. Following these are six variables related directly to the focus of this study, problem through pstrain. These variables attempt to characterize the perception of soil erosion as a problem, the use of soil and water conservation practices, and.the varying levels of contact with Plan Sierra. The last five variables in the table, farmeryr’through income, provide information on the farm household itself. 56 sale to animals raised for home use use and sale, responses to 018.1- 018.7; divided by sum of “1“, home use and "3", both home use and sale responses to 018.1-018.7 Table 2 Variables used to test hypotheses 3 and 4 Variable Explanation Derivation Type Name slclass FAO Slope Class Reclassified values of slope according Ordinal ___.;___ g "___“___m__m _ ptg_standard FAQ sl-o- classes,” farmsize Far-size grouped as "large" Reclassified values of 01, farmsize I Ordinal or "small" “small" if 01 < 100 tareas, farmsizeI "large“ if 01 z 100 tareas crdivers Total number of different Sum of total non-zero responses to Ratio crops cultivated 09.1-09.9 cafe Cultivation of coffee cafe-1 if one of 09.1 to 09.9 I 1 Nominal yuca Cultivation of yuca yuca=1 if one of 09.1 to 09.9 I 3 Nominal mais Cultivation of corn mais=1 if one of 09.1 to 09.9 I 2 Nominal beans Cultivation of green beans beans-1 if one of 09.1 to 09.9 I 4 or Nominal or black beans 13 banana Cultivation of bananas or banana=1 if one of 09.1 to 09.9 I 5 or Nominal plantains ' 7 I crpratio Ratio of crops grown for Sum of "1", sale, and "3", both home Ratio sale to crops grown for home use and sale, responses to 013.1- use 013.9; divided by sum of "2', home use, and “3", both home use and sale responses to 013.1-013.9 andivers Total nuaber of different Total number of "1“, yes, responses to Ratio l types of animals raised 017.1-017.7 __gallinas Raising of chickens gallinasI1 if 017.1 I 1 Nominal cerdos Raising of pigs cerdosI1 if 017.2 I 1 Nominal vacas Raising of cattle vacasI1 if 017.4 I 1 Nominal mulo Raising of mules, donkeys,‘ muloI1 if 017.5 I 1 Nominal or horses anmratio Ratio of animals raised for Sum of ”2', sale and '3", both home Ratio Table 2 (con't) 1 Variable Explanation "TP_ l .___l 57 Derivation labor Use of outside labor Response to 053 Nominal ‘ pesticid Use of pesticides Response to 057 Nominal | herbicid Use of herbicides Response to 061 Nominal fertiliz Use of fertilizer Response to 065 Nominal H iUse of credit problem Perception of soil erosion Response to 014.5 Ordinal ' as a problem swcuse Use of soil and water Response to 029 Nominal conservation practices at pslist Farmer was selected from Nominal lists provided by Plan , Sierra pshear Farmer has received pshear I “yes“ if 027-“4' and “Plan Nominal information from Plan Sierra Sierra", 028I"4“ and I'Plan Sierra", regarding soil and water psextend I “yes", or pstrain I "yes"; conservation practices else pshear I "no“ psextend Farmer has received. psextend I ”yes" if 033I”3” and "Plan Nominal extension advice regarding Sierra" or 046I'4” and ”Plan Sierra”; soil and water conservation else psaxtend I “no“ practices from. Plan Sierra extension agents pstrain Farmer has participated in a pstrain I ”yes“ if 033I"4' and “Plan Nominal training session at Plan Sierra training center in Sierra” or "Los Nontenes'; else pstrain I "no“ ,"i!fl§fl” . farmeryr Age of farmer (head of Response to 085.1 Ratio household) yrsfarm Years farming on major Response to 05.1 Ratio parcel famsize Total number of family Sum of the responses to 084.1 and Ratio members 084.2 emigrate Total number of family Sum of the responses to 088.1 and Ratio members who have left the q88.2 area income Income level of farmer Reclassified values of 094, income I (high/low) "low“ if 094 S RDSSOOO/year, income I "high" if 094 > RDSSOOO/year Ordinal 58 The other major concept utilized.in hypotheses 3 and 4 is the concept of slope zone. As discussed in previous chapters, slope plays an important role in the erosion process and consequently is an important, and possibly overriding characteristic of any farming system. However, analysis of farming system similarities and differences related to slope is very difficult unless some systematic classification of slope is used. To this end, the researcher adopted the slope classification system used by the Food and Agriculture Organization of the United Nation (FAO) which is detailed in Table 3. Table 3 FAO Slope Classes (used to define the variable slclass) (From Sheng, 1989) - l- l - -_----ll -- _ , _ 7 Soil Depth Suggested Land Use ; i ! 0-12% < 15 Pasture i 7-15° > 30 Any Crop 2 12-27% < 30 Pasture 15-20° > 45 Any Crop 3 27-35% < 45 Pasture 20-25° > 55 Annual and 4 Perennial Crops 36-47% < 55 Pasture 25-30° > 60 Tree Crops 5 47—53% < 60 Forest or Tree Crops 6 > 30° All Forest Only > 58% Depths _. 59 The slope value used in this analysis is the slope of the farm as measured. by the interviewer at the farm site. Interviewers took several slope measurements at each farm. When the surveys were coded by the researcher, these slope measurements were averaged to determine an overall farm slope. These overall values were then reclassified into the appropriate FAO slope zones. 3.7 Data analysis After coding, the survey data was entered into the SPSS/PC+ statistical package for data analysis. This statistical package is widely used in social science research and allows the researcher to perform a wide range of statistical operations to test possible relationships between the collected data. I In order to conduct tests for relationships evident in the data, it is important to determine the level of measurement of various information. Blalock (1972) stresses that "The use of a particular mathematical model (or statistical procedure) presupposes that a certain level of measurement has been attained" (p. 21) . As a consequence, the type of data provided by the field.research.is the fundamental determinant of the type of statistics which are appropriate for use in testing the research hypotheses. Most of the variables used in this study provide nominal or ordinal level data. These include all of the variables used to test hypotheses 1 and 2 along with many of the variables used to test hypotheses 3 and 4. A few of the 60 variables used to test these hypotheses provide interval or ratio level data. 3.7.1 Nominal and ordinal data In the cases where the data available is at the nominal level, contingency tables will be used to compare the data and test hypothesis 1, 2, and 4. The phi statistic (oz = XZ/N) will then be used to determine whether a significant relationship between the variables exists. This statistic has a value of ‘ 0 when there is no relationship between the variables. Since there is a direct relationship between «92 and x2, the distribution of this statistic is known and significance levels are readily available (Blalock, 1972). There are problems with the accuracy of x3 and related statistics when the expected value of a significant percentage of the cells in a contingency table is less than five (Blalock, 1972). Since this is the case in some‘ of the contingency tables used in this analysis, another statistical test will be employed. This is the Fisher's Exact Test. This test, based on the hypergeometric distribution, allows for a determination of the exact probability of a given set of conditions in a 2 x 2 contingency table (Blalock, 1972). In the case of hypothesis 3, the researcher is attempting to determined the homogeneity of the farming system characteristics within specific groups defined by slope zone. As a consequence, the a statistic is inappropriate. The author will calculate frequency distributions of the appropriate variables. All of the values of the nominal and 61 ordinal variables in this study can be logically grouped to create binomial variables. The 2 statistic; z = ml-no/oz, where a, = \lrr(1-rro)/n, n1 is the measured percentage of variable, n5 is the estimated probability of one value of the variable (estimated to be 0.5), and n is the total sample size; can then be used to determine the significance of the binomial parameter, n (Ott, 1988). 3.7.2 Interval and ratio data Although the majority of the data collected in this study is at the nominal or ordinal level, a small amount of data is at the interval or ratio level. This includes the variables years farming (yrsfarm), age of head of household (farmeryr), family size (famsise), and number of family members who have emigrated from the area (emigrate). This also includes a number of indices measuring aggregate properties of the crop and animal components of the farming system (crdivers, crpratio, andivers, anmratio). When comparing these variables across groups, analysis of variance will be employed. The F statistic will then be used to test for significant differences between population means. As in the cases of nominal and ordinal level data discussed above, a different approach is necessary for hypothesis 3. The mean (u) and standard.deviation (a) will be determined for interval and ratio level variables. These allow the calculation of the coefficient of variation, V = o/p. If the variables are assumed to be normally distributed, 62 this statistic can be used to determine if the variable shows significant variation within a slope zone. 3.7.3 Analysis methodology for hypotheses 1a, 1b, and 1c Since the data available to test all three parts of hypothesis 1 is at the nominal level, contingency tables will be constructed between the use of soil and water conservation (swcuse) and the various levels of interaction with Plan Sierra (pshear, psextend, and pstrain). The a statistic will then be employed to test these contingency tables for significant differences. A significant difference is defined to exist for each part of the hypothesis when the value of 4 statistic has a probability of less than 5%. 3.7.3.1 Criteria for rejection of the null hypothesis Hypotheses 1a, 1b, and 1c will be tested individually, the null hypothesis for each hypothesis will be rejected if a significant difference is found between groups. The null hypothesis will be accepted if no significant difference is found between groups. 5 3.7.4 Analysis methodology for hypotheses 2a, 2b, and 2c The data available to test hypotheses 2a, 2b, and 2c is also at the nominal level. As a consequence, the analysis :methodology is very similar to that use to test hypotheses 1a, 1b, and 1c. Contingency tables will be constructed between the ongoing use of soil and.water conservation (stilluse) and the various levels of interaction with Plan Sierra (pshear, ,psextend, and pstrain). The a statistic will then.be employed to test these contingency tables for significant differences. 63 The same criteria for significance (pm) < .05) will be employed. 3.7.4.1 Criteria for rejection of the null hypothesis Hypotheses 2a, 2b, and 2c will be tested individually, the null hypothesis for each hypothesis will be rejected if a significant difference is found between groups. The null hypothesis will be accepted if no significant difference is found between groups. 3.7.5 Analysis methodology for hypothesis 3 Hypothesis 3 requires the use of a different type of analysis methodology to that used above. It is much more difficult to test for significant differences within the farming system in an individual slope zone. In order to test the hypothesis, subsets of the data will be created by slope zone. Frequency tables will be computed for each of the ‘ farming system variables discussed earlier in this chapter for each subset of the data (slope zone). Since all the nominal and ordinal variables can be logically reclassified as binomial, a significant difference in a farming system characteristic is defined to exist when the percentage frequency of the most common value of the variable (:11) is less than the value of 111 determined by solving the equation 2 = (n1-no)/o for H1 under the following conditions: the expected probability, "0 is assumed to equal .5 for all values of each variable; the standard deviation, 0, is equal to mosnmesspshmmssriables, n, is equal 64 to the number of farms in the slope zone; and the value of z is defined to be 1.96 (2“ for p < .05). In the case of interval and ratio level variables, a significant difference is defined to exist when the coefficient of variation, V, is greater than 0.255. This value‘was determined'using'therapproximation.that the interval p i 1.960 contains the true mean 95% of the time (Ott, 1988). From.this, it is possible to determine that V > 1/2(1.96) will indicate a significant (p > .05) level of variability. 3.7.5.1 Criteria for acceptance of the null hypothesis The binomial distribution will also be used to define the criteria for rejection of the null hypothesis. The null hypothesis will be rejected for a specific slope zone if the percentage of variables showing no significant variation (n1) is less than 69.7%. This percentage has a.probability of less than 5%. The null hypothesis will be accepted for the slope zone if the percentage of variables showing no significant variation is 69.7% or more. For the overall testing of the hypothesis, the null hypothesis will be rejected for the study area if the null hypothesis is rejected for 3 or more of the 5 slope zones. The null hypothesis will be partially rejected if the null hypothesis is rejected for 1 or 2 of the 5 slope zones. The null hypothesis will be accepted for the entire study area only if the null hypothesis is accepted for all slope zones. 65 3.7.6 Analysis methodology for hypothesis 4 The methodology used to test hypothesis 4 is very similar to the methodology employed to test hypotheses 1 and 2. In this case, the set of farming system variables will be compared across slope classes. In the case of the nominal and ordinal level variables, contingency tables will be used to compare across the various slope class groups. The a statistic will be used to test for significant differences. In the case of the interval and ratio level data, analysis of variance will be used to compare across the various slope classes. The F statistic will then be used to test for significant differences. 3.7.6.1 Criteria for rejection of the null hypothesis The criteria used to test the null hypothesis for hypothesis 4 are similar to those used to accept or reject the null hypothesis for hypothesis 3. Inn this case, the null hypothesis will be rejected if the percentage of variables within that zone showing significant variation is less than 69.7%. This percentage has a 5% probability of occurrence. The null hypothesis will be accepted if the percentage of variables showing significant variation is greater than or equal to 69.7%. CHAPTER 4 RESULTS AND DISCUSSION The methodology outlined in the previous section provides the framework for the subsequent description and analysis of the data collected during this investigation. This chapter begins with a general description of the farming'syStem. 'This is followed.by descriptions of the results from.the testing of the four hypotheses of the study. The chapter concludes with a discussion of possible problems with the data and how these problems may have affected the results. 4.1 General description of farming system The data collected and the researcher's personal observations provided sufficient information to describe the farming system in the three study areas in very general terms. 4.1.1 Physical factors The physical environment in all three study areas was very similar. As mentioned in the introduction, the study areas were located at moderate altitude (600mr800m) and in areas of reasonably consistent rainfall” The average slope of all farms in the study was 23.8° and varied only slightly across all three sites. Soils in all the areas were loam-clay loam in texture, with significant organic :matter content, and. neutral pH values. The soils in Las Piedras appeared fertile; however, they were very shallow and stony. Similar conditions were also found in Rincon Largo. El Gallo has a wetter climate. 66 67 As a consequence of the increased precipitation and resulting weathering, soils were deeper and moderately acidic. Two groups of farms could be readily identified in the study areas by size. Of the 82 farmers surveyed, 66 lived on small farms (mean size 31.4 tareas3)*while 16 lived on large farms (mean size 267.5 tareas). Land holdings were generally smaller in Rincon Largo, but were still strongly bimodally distributed. Both large and small farmers generally had between 1 and 3 parcels of land. One of these contains the family house and related structures. Tenure in these areas was very secure. Nearly all (95.1%) of farmers reported that they owned at least one parcel of their land. The remainder were all cultivating land which belonged to family members. 4.1.2 Farm enterprises All of the farmers surveyed were actively participating in agricultural activities. The principle crops grown were coffee (90.2%), cassava (69.5%), corn (59.8%), banana or plantain (53.7%), and beans (50%). In general, farmers will cultivate coffee, often intercropped with bananas, on the parcel adjacent to their house. They will then cultivate a ”conuco“, a relatively small .field of annual crops. Traditionally, the term conuco implies swidden cultivation. However, since land is increasingly scarce in these areas, the swidden/fallow system of cultivation did not appear to be 3Land in the Dominican Republic is measured in tareas. Traditionally, 1 tarea is the amount of land 1 man can cultivate in 1 day. It is approximately equal to 1/16 of a hectare (Murphy and de Castro, 1990). 68 widely used. The only place where it was evident was in El G.11.. All the farmers who reported growing coffee grew it as a cash crop; however, most farmers reported that they retain part of the harvest for home consumption. The remainder of the crops were generally grown for home consumption, either for family members or livestock. Small quantities of cassava and occasionally corn may be sold for income generation. Livestock; were also an important component of this farming system. Nearly all farmers (81 of 82) raised some type of animals on their farm. The most common animals were chickens (98.8%), mules, donkeys or horses (74.1%), pigs (65.4%) and cows (44.4%). In all cases, chickens were raised for home consumption. Mules, donkeys, and horses were raised for use as cargo animals. Pigs were raised mainly for sale, but may be used for home consumption on special occasions. Cows were raised for milk, particularly on small farms, although some are also sold. 4.1.3 Farm inputs and outputs As mentioned previously, the major source of cash income for most farmers was coffee cultivation. Additional income may come from the sale of livestock or the sale of other agricultural crops. Another important source of income in some cases were family members now living either in Dominican cities or the United States. Nearly one-half (43.9%) of the families reported at least one member living outside of the community. The most common destination (75.7%) was Santiago, 69 the nearest large city. Although people were generally reluctant to discuss the amount of money coming from the outside, it appears to provide a significant portion of some farmer's income. A number of agricultural inputs were also used in this farming system. Nearly two-thirds (62.2%) of those interviewed reported using hired labor, particularly for coffeelharvestu Pesticides (13.4%) and.herbicides (4.9%) were not widely ’used in these areas. Fertilizer use was more common (50% of farmers), but is by no means universal. .All of these inputs were mainly used on cash crops (usually coffee). Agricultural credit was also used by some farmers (30.5%). These are mainly larger farmers or those who participate in the Plan Sierra credit program. 4.1.4 Farm family attributes The data also provided information. on. a number of characteristics of the farm family. The mean age of the head of household.was just under 48 years. This person, who in all cases given.was a.man, had.been farming on his current land an average of just over 22 years. The average family size in this area was approximately 6 persons, generally 2 adults and 4 children. The median reported family income was approximately RD$5000 (approximately US$400) per year. However, the Dominican interviewers felt that a number of farmers had underreported this amount, especially those who had family members living elsewhere. 70 4.2' Testing of hypothesis 1 The first hypotheses to be tested in this analysis are the three parts of hypothesis 1 which are defined on page 11. In order to test this hypothesis, contingency tables were constructed using the variables swcuse, pshear, psextend, and pstrain which were defined in the previous chapter. These definitions are summarized in Table 1 (p. 55). Since the methodology used in this investigation contains the possibility of bias, three sub-sets of the collected data were used to test this hypothesis. 4.2.1 Hypothesis 1a The first contingency table (Table 12, Appendix A) shows the use of soil and water conservation (swcuse) crosstabulated by whether or not a farmer has received information on soil and water conservation practices from Plan Sierra (pshear) based on the entire study sample (n=82). Nineteen of .the subjects interviewed for this investigation were selected from a list of cooperators provided by Plan Sierra. There is the distinct possibility that they are not representative of the study population as a whole. As a consequence, an identical contingency table (Table 13, Appendix A) was constructed for the remainder of the study sample when these subjects are omitted (n=63). Although the removal of the 19 listed farmers is likely to have improved the representativeness of the sample, it is also possible that the two sites selected from the list provided by Plan Sierra are not truly representative of the 71 entire study area. Because of this situation, a third contingency table (Table 14, Appendix A) was created which provides identical information to that in Tables 12 and 13, but used the data collected in the Las Piedras area (n=42). This was the only area selected completely at random. In spite of the differences in effective sample size in the three different contingency tables presented, the a values in all three cases, .802, .746, and .718 were significant at the g>i< .001 level. This indicates a highly significant positive association between the two variables. These results are summarized in Table 4. Table 4: Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the receipt of information from Plan Sierra (pshear) Contingency Minimum Phi Value Approximate Table Expected Significance Frequency Table 12 14.9 .802 < .001 (n=82) 3' Table 13 13.3 .746 < .001 (n=63) Table 14 9.5 .718 < .001 (n=42) 4.2.2 ' Hypothesis 1b Similar contingency tables were constructed to test the possible association between the use of soil and water conservation and more conservative measures of contact with Plan Sierra. The first set of these tables (Tables 15, 16, 72 and 17; Appendix A) illustrate the distribution of soil and water conservation use as related to direct contact and discussion with Plan Sierra extension staff regarding the adoption and use of new technologies. The values of the a statistic for the three subsets of the data, .745, .663, and .603 respectively, are all significant at the p < .001 level. These results demonstrate a highly positive correlation between adoption of soil and water conservation and.participation in Plan Sierra extension programs. A summary of the results is provided in Table 5. Table 5: Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the interaction with Plan Sierra extension personnel (psextend) Contingency Minimum Phi Value Approximate Table Expected Significance Frequency Table 15 14.9 .745 < .001 (n=82) Table 16 8.6 .663 < .001 (n=63) Table 17 5.7 .603 < .001 (n=42) 4.2.3 Hypothesis 1c Contingency tables were also constructed to test the possible association between the use of soil and water conservation practices and farmer participation in off-site training activities conducted by Plan Sierra at their training 73 center in Los Montones. These results are illustrated in Tables 18, 19, and 20 (Appendix A). The results of statistical tests for association between the variables are summarized in Table 6. Table 6: Summary of the association between the adoption of soil and water conservation technologies (swcuse) and the participation in off-farm training activities conducted by Plan Sierra (pstrain) Contingency Minimum Phi Value Approximate Table Expected Significance Frequency 1 Table 18 7.0 .430 < .001 i(n=82) 3 Table 19 3.3‘ .370 < .005 : (n=63) Q Table 20 2.4. .350 < .005 a: Since the Minimum Expected Frequency is < 5, the Fisher's Exact Test Values were computed for both of these cases, {the values are .004 and .03 respectively. The a values for the entire sample population is .430. This value is significant at the p < .001 level. The a values for the two smaller subsets are .370 and .350 respectively. These values are both significant at the p < .005 level. lHowever, as discussed in Chapter 3, when the minimum expected frequency of a contingency table cell is less than 5, there - can be problems with the a statistic. As a consequence, the Fisher's Exact Test was used to compute the exact probability of the distributions shown in Tables 19 and 20. These values are p I .004 and p = .03 respectively. Even though these 74 values are higher than the probability of the corresponding Q values, they are both significant at the p < .05 level. 4.2.4 Summary of results for hypothesis 1 The results reported in Tables 4, 5, and 6 provide ample evidence to test all three parts of hypothesis 1. 4.2.4.1 Hypothesis 1a The results reported in Table 4 showed a very significant association between the adoption of soil and water conservation.technologies and the receipt of information from Plan Sierra. The criteria outlined in Chapter 3 allowed the researcher to reject the null hypothesis, Ho: Farmers who have received information from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this information. Furthermore, the distribution of the data as shown in the relevant contingency tables (Tables 12, 13, and 14; Appendix A) indicated that the association was a positive one. Therefore the researcher accepted the alternative hypothesis, H1: Farmers who have received information from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this information. 4.2.4.2 Hypothesis 1h The results reported in Table 5 showed a very significant association between the adoption of soil and water conservation technologies and on-farm contact with Plan Sierra extension agents regarding these technologies. The criteria outlined in Chapter 3 allowed the researcher to reject the null hypothesis, 75 Ho: ' Farmers .who have received extension advice from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this advice. Furthermore, the distribution of the data as shown in the relevant contingency tables (Tables 15, 16, and 17; Appendix A) indicated that the association was a positive one. Therefore the researcher accepted the alternative hypothesis, H1: Farmers who have received extension advice from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this advice. 4.2.4.3 Hypothesis 1c The results reported in Table 6 showed a significant association between the adoption of soil and water conservation technologies and the participation in an off-farm training at the Plan Sierra training center. The criteria _out1ined in Chapter 3 allowed the researcher to reject the null hypothesis, Ho: Farmers who have received off-famm training from Plan Sierra are no more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this training. Furthermore, the distribution of the data as shown in the relevant contingency tables (Tables 18, 19, and 20; Appendix A) indicated that the association was a positive one. Therefore the researcher accepted.the alternative hypothesis, H1: Farmers who have received off-farm training from Plan Sierra are more likely to adopt soil and water conservation practices on their farms than are farmers who have not received this training. 76 4.3 Testing of hypothesis 2 The second set of hypotheses to be tested in this analysis are the three subdivisions of hypothesis 2, which are defined on page 12. As discussed in the methods chapter, the process used to test these hypotheses is straight-forward and very similar to that used to test the three components of hypothesis 1. 4.3.1 Hypotheses 2a, 2b, and 2c Contingency tables identical to those discussed in the testing of the previous hypothesis were constructed between the variable stilluse as in Chapter 3 (summarized in Table 1) and the various levels of interaction. with Plan Sierra (pshear, psextend, pstrain). Analysis was conducted for the same three subsets of the data used to test the previous hypothesis. These tables (21-29) are included in Appendix A. The results are summarized in Table 7. 4.3.2 Summary of Results for hypothesis 2 The results presented in Table 7 provided ample evidence to test hypotheses 2a,. 2b, and 2c. Because the minimum expected cell frequency for the appropriate contingency tables was less than 5 in all cases, significance was determined using Fisher's Exact Test values. 4.3.2.1 Hypothesis 2a The initial three entries in the table were used to test hypothesis 2a. The Fisher's Exact Test values for the three' appropriate subsets of the study sample were .895, .833, and .818 respectively. Since these values were all greater than 77 .05, none of them were significant according to the criteria put forward in Chapter 3. Consequently, the researcher accepted the null hypothesis, Ho: Farmers who have received information from Plan Sierra are no more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this information. for hypothesis 2a. 4.3.2.2 Hypothesis 2h The next set of three entries in Table 7 were used to test hypothesis 2b. The Fisher's Exact Test values for the three sample subsets were .750, .600, and .545 respectively. Although these values indicated a slightly stronger relationship than that shown for the previous hypothesis, all of them were still much greater than .05. Therefore, they did not indicate a significant relationship between the variables and the researcher accepted the null hypothesis, H0: Farmers who have received extension advice from Plan Sierra are no more likely to have continued to use soil and.water conservation practices on their farms than are farmers who have not received this advice. 4.3.2.3 Hypothesis 2c The final three entries in Table 7 were used to test hypothesis 2c. The Fisher's Exact Test values for the three sample subsets were .646, .767, and .773 respectively. .As in the previous two hypothesis, these values were much greater than .05. When the criteria outlined in the methods chapter 78 Table 7 Summary of the association between the continued use of soil and water conservation technologies (stilluse) and the various levels of interaction with Plan Sierra (pshear, psextend, pstrain). [Contingency Minimum Phi Approximate Fisher's 1 Table Expected Value Significance Exact Test '1 f Freu-__nc _ of Phi _ Value {PSHEAR : Table 21 .104 .730 .895 ! (n=48) .049 Table 22 .167 .649 .833 (n=30) .083 Table 23 .182 .629 .818 (n=22) .103 J fl 4 PSEXTEND . Table 24 .250 .560 .750 (n=48) .034 Table 25 .400 .406 .600 (n=30) .152 Table 26 .455 .545 (n=22) PSTRAIN a: Table 27 .354 .454 .646 (n=48) .108 Table 28 .233 .574 .767 (n=30) .102 Table 29 .227 .579 .773 (n=22) Since the Minimum Expected Frequency is < 5, the Fisher's Exact Test Values were computed for all of these cases. 79 are applied, these results are determined to be not significant. Therefore, the researcher accepted the null hypothesis, Ho: Farmers who have received off-farm training from Plan Sierra are no more likely to have continued to use soil and water conservation practices on their farms than are farmers who have not received this training. 4.4 Testing of hypothesis 3 The third hypothesis to be tested in this analysis is hypothesis 3 defined on page 13. As discussed in the methods section, a number of variables were used to characterize the farming system. These variables were defined in Chapter 3 (Table 2). 4.4.1 Information from Plan Sierra As discussed. in the :methods section, the data. was separated into five subsets by slope zone. Because of the small number of farms in slope zone 1 (2) and slope zone 2 (5), these were combined in the analysis. Frequency distributions were calculated for all of the nominal and ordinal level variables in each slope class, while the coefficient of variation, V, values were calculated for all the ordinal and ratio level variables. The significance of the variation in each variable was determined according to the previously defined criteria. 4.4.2 Summary of results for hypothesis 3 The overall results from the testing of this hypothesis are summarized in Table 8. Only four variables (13.3%) did not show significant variation within the slope zone 1 & 2. 80 Eleven variables (36.7) did not show significant variation within slope zone 3; ten (33.3%) within zone 4; eleven (36.7%) within zone 5; and 9 (30.0%) within zone 6. Complete information on the variation in each variable is shown in Table 30 (Appendix A). These results demonstrated a considerable variation in a number of farming system description variables within each slope zone. When the testing criteria outlined in the methods section was applied, the null hypothesis, Ho: The farming systems found within an individual slope zone are not significantly different from each other. was rejected for each of the individual slope zones. Since it was rejected for all the slope zones, it was rejected for the entire study. Therefore, the researcher accepted the alternative hypothesis, H1: The farming systems found within an individual slope zone are significantly different from each other. Table 8: Summary of hypothesis 3 test results . Slope Zone Number of variables Percentage of ' showing no - all variables significant variation (n=30) 4 13.3 :(7 farmers) 1 11 36.7 ' (19 farmers) . 4 10 33.3 f (25 farmers) 5 36.7 (17 farmers) ‘ 3 (12 farmers) 81 4.5 Testing of hypothesis 4 The fourth hypothesis to be tested in this analysis is hypothesis 4, defined on page 13. The same variables used to test hypothesis 3 were used to test this hypothesis. 4.5.1 Information from Plan Sierra As in the testing of the previous hypotheses 1 and 2, contingency tables and analysis of variance were conducted for the 30 variables used to characterize the farming system by slope zone. values were computed for the appropriate test statistics (o for nominal and ordinal level variables, F for interval and ratio level variables), and the distributions of these statistics were used to determine the significance of these values. 4.5.2 Summary of results for hypothesis 4 The final results of the analyses are presented in Table 9. From this table, it is apparent that the test statistic values for all of the 30 farming system definition variables are not significant at the p < .05 level. Three variables are significant at the p < .10 level. These are andivers, a measure of the number of different animals raised; vacas, the presence or absence of cattle in the farming system; and problem, the perception of soil erosion as a problem. All three of these variables generally increase with increasing slope. The trend for the variable problem is not unexpected. Farmers working steeper land, with a higher erosion potential, are more likely to notice a erosion as problem. The trend 82 shown in the variables andivers and vacas is much less easily explained. It may reflect a greater emphasis on livestock by farmers on more 'steeply sloping land. However, this suppositionis not supported by other related variables including cerdos, the possession of pigs which has a much higher approximate significance (.302). Since none of the farming systems definition variables showed significant variation across slope zones as defined in the methods' chapter, the researcher rejected the null hypothesis, Ho: The farming systems found in different slope zones are significantly different from each other. The researcher accepted the alternative hypothesis, H1: The farming systems found in different slope zones are not significantly different from each other. 83 Table 9 Summary of results use to test hypothesis 4 Approxjests Variablea Test autistic Approxiuate _ __ __ __ V'L, ,. ___ W’L‘i__ =-==_= "L“, __,, _ _ 5‘9"”‘°'"“ M n/a his labor .142 .806 [- far-min .201. .502 pesticid .211 .471 crdivers .183 .947 herbicid .159 .732 cafe .269 .216 ferti Hz .222 .422 ca .243 .319 credit .179 .632 j III-1's -136 -830 jrwlu .434 .094. j beans .240 .330 swcuse .215 .447 I banan- -164 -708 pslist .069 .984 f crpratio 1.473 .220 pshesr .124 .875 andivers 2.088 .091 psextend .119 .888 Illinas .201 .525 pstrain .171 .674 90'9“ - 2‘3 - 302 far-erg 1 . 192 . 321 vacas . 326 .079 yrsfere 1 .419 .236 eulo .277 .195 feesize 1.303 .277 snerstio 1.263 .301 esigrsts .371 .828 incoee .203 .564 a: The descriptions and derivation criteria for each of these variables may be found in Table 2. b: For those variables which produce nominal or ordinal data, the value of the 6 statistic is given. For those variableswhich produce interval or ratio data, the value of the F statistic is given. 84 4.6 Possible problems with the data As discussed previously, the data collected in this study have proven to be sufficient for testing all four of the hypotheses. ' It is possible that the data collection procedure used may have introduced bias in favor of Plan Sierra; however the strength of this bias, if any, was not readily apparent. If it exists, bias is likely to exist on two levels. First, the nineteen farmers interviewed from the lists of farmers provided by Plan Sierra may not be a representative sample of area farmers. Second, the two sites chosen from the list of communities provided by Plan Sierra (El Gallo and Rincon Largo) may not accurately represent the area as a whole but may represent areas in which Plan Sierra has been particularly successful. In addition to the above biases, the possibility of gender bias must be addressed. The researcher, his principle Dominican colleague, and all three Dominican interviewers are men. In addition, nearly all the interviews in this study were conducted with men. As a consequence, women's knowledge regarding the survey information is lacking. 4.6.1 Testing for bias To attempt to determine the existence and extent of the first two biases, analysis of variance was conducted and contingency tables constructed identical to those used in the previous section to test hypotheses 4. This procedure allowed the comparison of various farming system attributes between farmers chosen from the lists provided by Plan Sierra and 85 farmers chosen at random. They also permitted the comparison of farming system attributes between the farmers selected at random in the sites from the list provided by Plan Sierra (El Gallo and Rincdn Largo) and those in the site chosen at random (Las Piedras). 4.6.1.1 Plan Sierra list bias The results of the comparisons between farmers interviewed who were selected from lists provided by Plan Sierra and those selected at random are summarized in Table ' 10. Seven variables in the analysis were found to be significantly different (p < .05) between the two groups. These were the variables measuring crop diversity, crdivers (p = .001); the variable indicating the presence or absence of pesticide use pesticid (p = .001); the perception of soil erosion as a problem, problem (p = .047); the use of soil and water conservation practices, swcuse (p < .001); and the receipt of information, extension advice, and training from Plan Sierra, pshear, psextend, and pstrain respectively (p < .001 for all three variables). Farmers who were selected from Plan Sierra lists grew a higher number of different crOps and were more likely to use pesticides than did farmers not on these lists. They were also more likely to view erosion as a problem. They were much more likely to have interacted with Plan Sierra at all three levels (information, extension, and training), and they were much more likely to use soil anduwater conservation practices. 86 Table 10 Summary of results from tests for Plan Sierra list bias I Test statistic Approximate Variable. Test gtatistic Approxieate I ! Value significance Value significance - _ .069 .984 labor .141 .200 I farmsize .094d .172d 993ti°idc -373 -001 crdiversc 10.932 .001 herbicid .010 .929 cafe .181 .102 fertiliz .197 .076 4yuca .175 .112 credit .139 .209 nis .156 .158 Emma“ .237 .047 I beans .144 .191 swcusec .404 <.000 I M 990909 -047 -972 pslist n/a n/a . crpratio 1.323 .131 pshearc .427 <.001 andivers .539 .465 psextendc .562 (.000 s-llinas .062 .578 pstrainc .432 <.001 cerdos ‘.026 -312 fareeryr .610 .437 . vacas .085 .446 yrsfars .139 .711 I auto .061 .580 faesize .658 .420 I anmratio,. _» emigrate .979 .326 a: The descriptions and derivation criteria for each of these variables may be found in Table 2. b: For those variables which produce nominal or ordinal data, the value of the a statistic is given. For those variables which produce interval or'ratioldata, the value of the F statistic is given. ’ c: The difference in this variable is significant at the 5% level d: Because of problems with the SPSS-x PC+ software, these values were calculated manually 87 These results reflected very well upon the operations of Plan Sierra. Farmers who were designated by the organization as cooperators had successfully incorporated the project messages: lower risk through diversity, recognition of soil erosion as a problem, and the need to use soil and water conservation practices in their farming system. There are several possible reasons for increased use of pesticides including: higher awareness of the potential benefits of pesticide application; more knowledge regarding the efficient and safe use of pesticides, access to the credit necessary to buy farm inputs when they are needed; and a stronger market orientation. However, variables which attempted. to measure use of credit,_ credit, and.:market orientation, crpratio, did not show significant differences. In addition, the lack of significance in several other variables reflected well on Plan Sierra operations. Farmers on the list appeared to reflect a typical cross-section of area farmers. Significant differences did not exist between list members and non-list members for such important variables as slope class, slclass (p = .984); farm size, farmsise (p = .172); age of farmer, farmeryr (p = .437); and income level, income (p = .287). These are typical sources of bias in many development projects (Chambers, 1983). 4.6.1.1.1 Possible effects of bias The existing differences may have affected test results for’ all four of the hypotheses. The most significant differences between list and non-list farmers were found in 88 the variables related to the adoption of soil and water conservation, swcuse and interaction with Plan Sierra, pshear, psextend, and pstrain. Since the testing of hypotheses 1 and 2 used.these variables, if bias was present it was most likely to occur in this context. Since significant differences were found in only 23% (7 of 30) of the farming system definition Variables and were not found in a number of the key variables mentioned in the previous section, the researcher concluded that this bias was not likely to affect the results determined from testing of hypotheses 3 and 4. 4.6.1.1.2 Attempts to minimize bias The potential for Plan Sierra list bias in the data was addressed by the use of both the complete data set and a subset of the data consisting of those farmers not on the Plan Sierra lists, for the testing of hypotheses 1 and 2. For hypothesis 1, the a value decreased for the smaller subset in comparisons between soil and water conservation practice use and all three levels of interaction with the project. This indicates that the inclusion of Plan Sierra list. members did strengthen the association. between the variables in all three cases. However, the significance of the a values in all cases was still much less than the .05 criteria used by this study. Therefore, the researcher concludes that, although some bias was present, it did not significantly affect the testing of this hypothesis. 89 For hypothesis 2, the results are less clear—cut. The removal of the Plan Sierra list group from the analysis does change the 6 values in the comparisons between continued use of soil and water conservation and various levels of interaction with Plan Sierra. The variables pshear, the receipt of information from the project, and psextend, the receipt of extension advice, both show a small decrease in 9 value with the removal of the listed farmers. This is unexpected because all of these farmers continue to use the practices. However, the results from the variable pstrain, the participation in‘a Plan Sierra training, show an increase in the 6 value similar to that shown for hypothesis 1. A possible reason for these inconsistent changes may be that only one farmer interviewed had discontinued use of the practices. The 9 values may have been more affected by the decreasing sample size, which. would have increased the importance of an individual farmer, than by any characteristic of farmers on the Plan Sierra lists. As is the case with hypothesis 1, the changes in Q values are not large enough to greatly affect the level of significance of the relationship tested in this hypothesis. Some bias may be present, but it has not affected this study's results. 4.6.1.2 Plan Sierra site bias The short list of communities provided by Plan Sierra is another potential source of bias. As a consequence, comparisons were made between farmers who reside in the 90 communities selected from the list provided by Plan Sierra (Rincén Largo and El Gallo) and those who reside in the community selected at random within the Plan Sierra project area (Las Piedras). A summary is shown in Table 11. Only a small number of variables were found to be significantly different (p < .05). These variables are the cultivation of cassava, yuca, possession of a mule, donkey, or horse, mulo, use of pesticides, pesticid, and farm income ranking, income. In all of these cases, residents of the Plan Sierra list sites show higher levels of the variable. There are several possible explanations for this. In the case of cultivation of cassava, yuca, and the possession of a mule, donkey, or horse, mulo, site factors appear to play a role. Both of the Plan Sierra sites were more inaccessible than the site chosen at random. In this situation, farmers might be more likely to cultivate subsistence crops like yuca and may be more dependent on cargo animals since access to public transportation necessitates a considerable walk. The difference in pesticide use, pesticid, has already been discussed in the context of Plan Sierra list bias and its significance here may be directly related to that bias since all list members reside in the selected communities. The difference in income level, income, more difficult to account for; however, the difference in farm size, farmsise, was significant at the p < .10 level. This indicates one potential factor and may partially explain the income differences across the sites. 91 Table 11 Summary of results from tests for Plan Sierra site bias Variable' Test tatistic Approxisate Variablea Test tatistic Approxieate Value Significance Value ==Si_gnificance . slclass labor .027 .830 far-size .207d .075“ pesticid° -368 .003 crdivers 1.649 .204 herbicid .163d .222d cafe .169d .146d fertiliz .035 .731 yuca‘ .306 .015 credit .025 .341 ”is -090 1'73 problee .251 .171 beans .113 .371 _ ,_ 7 . . 7 .134 .285 banana -000 1-000 pslist n/a n/a crpratio .280 .599 pshear .090 .473 andi vers 2 . 092 .153 psextend .000 . 000 dallinas 092 .471 pstrain .032 .777 99'6“ 136 - 235 fareeryr 1 .031 . 314 vacas . 080 . 527 yrsfare 1 . 707 .196 mo“ 363 .004 13.312. .033 .357 . anmratio 041 .841 eeigrate .036 .851 imnnc .3“ .mm i _ a: The descriptions and derivation criteria for each of these variables may be found in Table 2. b: For those variables which produce nominal or ordinal data, the value of the a statistic is given. For those variables which produce interval or ratio. data, the value of the F statistic is given. c: The difference in this variable is significant at the 5% level d: Because of problems with the SPSS-PC+ Software, these statistics were computed manually 92 4.6.1.2.1 Possible effects of bias If it is present, Plan Sierra site bias is likely to have a much smaller effect on the testing of the four study hypotheses. No significant differences were found between groups for the variables used in testing hypotheses 1 and 2 (swcuse, stilluse, pshear, psextend, and pstrain). As a consequence, bias was unlikely to affect these results. Significant differences were found in only 13% (4 of 30) of the farming system definition variables. However, a Significant difference was found for the variable measuring income level, income (6 = .264, p = .048). This variable has the potential to have significant impacts but, in this case, the difference was not supported by significant differences in any other variables“ .As a consequence, the researcher believes that this significant difference did not represent a significant variation in the farming system between the communities and did not significantly affect the test results for hypotheses 3 and 4. 4.6.1.2.2 Attempts to minimize hias Even though the potential for Plan Sierra site bias was much lower than that for list bias, the researcher decided to use another subset of the data to test hypotheses 1 and 2. This subset consisted of farmers from the Las Piedras area, the site chosen at random. The use of this subset had a similar impact on the results to that discussed previously in the context of list bias. However, the magnitude of the change in a values is less in all cases (hypotheses 1a, 1b, 93 1c, 2a, 2b, and 2c). This indicates that any existing site bias had a much smaller effect. ’4.6.1.3 Gender bias The potential for gender bias also exists in this study. The researcher and his Dominican colleagues are all men and the vast majority of the information was collected from male farmers. The research had assumed that information would be obtained from. both. male and female heads of household. However, in virtually all cases, the person interviewed for the investigation was a man. The reasons for this appeared to be primarily cultural. First of all, the interviewers held the opinion, which is widespread in the Dominican Republic, that farming is the exclusive province of men. As a consequence, they believed that only the male head-of-household or an adult son could provide the necessary farming system information for the survey. Secondly, since the interviewers and the researcher were male, it was socially uncomfortable for both the interviewers and for the women involved to interview a women who was home alone or with small children. 4.6.1.3.1 Possible effects of bias The data available to the researcher do not permit the determination of the presence and strength.of possible gender bias in this study. Since crop farming is considered to be primarily a male occupation in the Dominican Republic, the researcher believes that the lack of information from women has not had a significant affect on this analysis. CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS The results of this study provide a clear basis for drawing conclusions regarding the effectiveness of Plan Sierra activities to promote the use soil and water conservation technologies. Conclusions can also be drawn concerning the feasibility of using slope zone as a division for extension and policy planning. The researcher hopes that these can provide some useful recommendations for the administration of Plan Sierra in particular and for other upland development efforts in the Dominican Republic and elsewhere in the sub- humid tropics. In addition to the conclusions and relatively specific .recommendations made, a number of areas are also identified which would greatly benefit from further study which have been developed from them, there are a number of areas which would greatly benefit from.further study» The final section.of this chapter will focus on these areas. 5.1 Conclusions from this study In spite of some biases present in the data, the judicious use of subsets of the entire data set permitted reliable testing of all four of the original hypotheses. Hypotheses 1 and 2 both are focused toward the study goal of determining the impact of Plan Sierra activities on the adoption and continued use of soil and water conservation farming practices. They are presented jointly in section 5.1.1. Hypotheses 3 and 4 focused on determination of the 94 95 feasibility of using slope zone as a defining characteristic for farming system recommendation domains. They are discussed in section 5.1.2. 5.1.1 Assessment of Plan Sierra outreach activities The results from testing hypothesis 1 and 2 and of the tests for bias allow a number of conclusions to be made concerning Plan Sierra's outreach programs, particularly in the area of the promotion of soil and water conservation farming practices. 5.1.1.1 Hypothesis 1 Testing of hypothesis 1 resulted in the rejection of the null hypothesis, H0! and the acceptance of the alternative hypothesis, H1, for all three subdivisions of hypothesis 1, hypotheses 1a, 1b, and 1c. These are defined on page 11. The results showed a significant relationship between all three levels of contact with Plan Sierra: receipt of information, receipt of on-farm extension advice, and participation in off-farm training; and the adoption of soil and water conservation practices. . These results reflected very well upon Plan Sierra's outreach activities, specifically the promotion of soil and water conservation practices. Few systematic differences were apparent in the farming systems, with the exception of the use of soil and water conservation practices, between tfluase farmers who have interacted with Plan Sierra and those who have not. As a consequence, interaction with Plan Sierra 96 seemed the be the primary driving factor behind the adoption of soil and water conservation practices. 5.1.1.2 Hypothesis 2 1 Testing of hypothesis 2 resulted.in the acceptance of the null hypothesis, Ho for each of its components, hypotheses 2a, 2b, and 2c. These are defined on page 12. Unfortunately, the data available did.not permit.a detailed assessment of the impact of Plan Sierra extension activities on the continued use of these practices. If the sample is representative, the data collected indicates a very low rate of discontinuance by all farmers, without regard to the presence or absence of interaction with Plan Sierra. Only 1 of the 48 farmers surveyed who had ever used soil and water conservation practices had discontinued use of the practices. This farmer cited competition between the trees used in an agroforestry system and his annual crops as his reason for discontinuance. However, this reason is inconsistent with his reported use of dead barriers as his only 'soil and water conservation practice. He had started using the technology before the start of the Plan Sierra project and had stopped using it in the early years of the project. He reported that he had received both information and extension advice from Plan Sierra since that time, but had not resumed use of the practices. Because the above information comes from only one farmer and because his answers were inconsistent, no substantive conclusions can be drawn. 97 Over one-half of the farmer users of soil and water conservation technologies interviewed had been using the technology for fiVe years or less. As a consequence, definitive assessments are difficult regarding the potential long-term use of these practices. However, all of the farmers who reported using soil and water conservation practices indicated that these practices were beneficial. This is likely to favorably affect the continued use of these technologies and reflects positively on Plan Sierra outreach activities. 5.1.1.3 Researcher's observations During the interview process, the researcher was able to make several general observations regarding Plan Sierra outreach activities. First of all, Plan Sierra outreach activities generally followed kinship networks (Plan Sierra cooperators included the father, sons, nephews, etc.). This could reflect poorly on the project; however, there are two more likely reasons for this observation: (1) in small rural communities the vast majority of residents are connected through kinship ties, and (2) farmers with relatives involved in any activity are more likely to be aware of and involved in this activity themselves. Since the farmer's name was never recorded during the interview process, the data collected cannot be used to test the validity of this observation. Secondly, virtually all of the farmers who have interacted with Plan Sierra had highly favorable views of the jproject. However, the researcher and Dominican interviewers 98 also observed that many of the farmers who did not interact with the project had strongly negative views of Plan Sierra. The structure of this investigation did not allow for systematic assessment of this observation. Thirdly, Plan Sierra field personnel seemed well informed and enthusiastic about their work. They appeared to interact well with farmers and were generally perceived as a valuable source of information. Lastly, some differences were observed between the farms of individuals who had interacted with Plan Sierra and the farms of those who had not. Since this investigation took place during the dry season, it was difficult to observe the full effect of various conservation measures. In addition, the researcher was not able to visit several conucos where farmers reported the use of soil and water conservation practices because they were some distance from the farmer's residence. However, in several cases, the ongoing use of soil and water conservation was clearly visible on participant's fields and the lack of use was clearly evident on those of non-participants. It must be noted that a number of farmers who reported interaction with Plan Sierra and the use of soil and water conservation were employing the practices on a limited part of their farm. Farmers may be using the remainder of their land for other purposes, such as pasture or coffee, which are perceived not to require soil and water conservation 99 technologies, or they may still be evaluating the technologies for themselves. 5.1.1.4 Overall conclusions The presence of many farmers with negative feelings toward the project and the selective adoption and use of practices by farmers are causes for concern. However, the remainder of the researcher's observations strongly support the conclusions from hypotheses 1 and 2. Plan Sierra outreach activities in the area of soil and water conservation have been and continue to be very effective. 5.1.2 Use of slope zones to delineate recommendation domains The second major goal of this investigation was to assess the feasibility of using slope zones as a way to delineate recommendation domains for both development and planning purposes. As discussed in Chapter 2, slope zone is believed to be one of the most appropriate divisions, particularly in upland areas, because slope is one of the primary determinants of soil erosion rates. 5.1.2.1 Hypothesis 3 Testing of hypothesis 3 resulted in the rejection of the null hypothesis, Hb, and the acceptance of the alternative hypothesis, H... These hypotheses are defined on page 13. These results show that significant variation is found in a significant number of farming system variables within each of the five slope zones. Therefore, the researcher concludes that other factors must be used to successfully delineate 100 recommendation domains in this area. If only lepe zone is used, a great deal of variation within the farming systems will be omitted from consideration. 5.1.2.2 Hypothesis 4 Testing of hypothesis 4 also resulted in the rejection of the null hypothesis, Hb, defined on page 13. These results indicated that none of the farming system description variables showed significant variation across slope zones. The researcher concludes that slope zone is not a discriminating way to define recommendation domains. 5.1.2.3 Researcher’s observations During the course of the investigation, the researcher was able to observe all three of the humid zone study areas. As discussed in Chapter 4, the primary component of the farming system, coffee cultivation, is widespread throughout the area on all slopes. In addition, farmer decision making appears to be driven by a relatively constant set of needs: principally subsistence food and outside income to pay for other needs such as clothing, farm equipment, and education. These needs do not change significantly across slope zone, but may be impacted by factors such as income, land quality, proximity to transportation, level of education or any of a number of others. - 5.1.2.4 Overall conclusions .Although slope remains an important determinant of enVironmentally sustainable farming practices, other factors aPPear to be playing a much stronger role in determining the 101 structure and composition of the existing farming system in this area. These may include climate; farm characteristics such as farm size and production systems; and farm family attributes such as age of the farmer, family size, and income sources. From the survey data and other observations, the researcher concludes that slope zone, although scientifically appropriate, is not an accurate predictor of farming system attributes. This does not imply that the practice of recommending soil and water conservation technologies on the basis of farm slope is not useful. However, care must be taken to insure that the practices can.be successfully adapted to the wide range of farming systems present in each slope zone. 5.1.3 Tests for bias The tests conducted for biases in both the list of sites provided by Plan Sierra and the list of farmers within those sites also reflected very well on the operations of the project. These tests indicated that Plan Sierra has avoided many of' the common biases in large development projects including bias toward younger farmers with more land, flatter land, and more income (Chambers, 1983). The project does appear to focus in relatively accessible areas, which may be a form of the tarmac bias discussed by Chambers (1983). However, this may be an artifact of the study methodology since this investigation was limited to similar areas due to budget and time considerations. 102 There is also the possibility of gender bias in project operations since the ;project appears to ‘work, ‘virtually exclusively, with- male farmers regarding soil and water conservation activities. However, this is more likely to result from the general consensus that crop farming is a male activity in the Dominican Republic. It is important to note that although Plan Sierra outreach activities related to agriculture were the focus of this study, -the project is involved in a number of other activities targeted toward both men and women. These include health and education programs directed toward entire communities and specific programs for mothers and for women's groups (de Janvry and.Hecht, 1984). Assessment of the success of these programs is well beyond the scope of this study. 5.1.4 General observations concerning Plan Sierra During the course of this investigation, the author was able to make a number of observations concerning the Plan Sierra.administration. Project administration is likely to have current and ongoing impacts -on all project activities including the area of soil and water conservation. The first of these relates to the role of outsiders and insiders in the project. When discussing these terms, it is important to be aware of the frame of reference being used. Traditionally, only persons from a different country than the PrOject location were considered outsiders. However, there is a much.narrower, and, in many ways, more appropriate view of Who actually are the insiders and the outsiders. Insiders are 103 the. population in the area affected. by the development intervention, outsiders are everyone else. This group may include project staff, government officials, and foreign aid workers. When this second definition is employed, outsiders have played and continued to play the predominant roles in Plan Sierra administration. The founding members and upper level staff of the organization are largely exogenous to the target area, although not to the Dominican Republic. Some, but far from all, of the field level staff are members of the local communities. Although some attempts have been made to add more indigenous character to the organization, such as the creation of local development councils, power in the project still rests mainly in the hands of outsiders. The second relates to the administrative structure of the project. In spite of assertions to the contrary, the project is organized in centralized, hierarchal fashion. Planning activities and control over the implementation of these activities originate at the center and are transferred down the hierarchal structure to the field staff who extend the technology to farmers. Accountability for the implementation of activities then flows back up the hierarchal structure from the field technician to the central office. Although the development of district offices and local development councils has reduced the number of levels in the hierarchy, the general BYStem of organization remains the same. 104 As discussed in Chapter 2, these are common problems with Integrated Rural Development projects in general and are believed to have strongly negative effects on project sustainability, especially if funding for the project is terminated. Plan Sierra appears to have reliable sources of funding at least in the near future. However, the project is highly dependent on funds from outside the Dominican Republic. (46% of total income in 1991 (Plan Sierra, 1991)). In spite of its centralized, hierarchal structure, Plan Sierra seems to have been sensitive to the felt needs of the farmers in their area of operation and has attempted to respond to those needs. This reflects very well on project administrators. The administration has also been able to collect and retain a skilled and extremely committed staff. 5.1.5 Summary of study conclusions The test results from hypotheses 1 and 2, list bias, and site bias present a very favorable picture of Plan Sierra outreach activities, particularly as they relate to soil and water conservation technologies. All levels of project outreach had.a highly positive association to the adoption of these technologies. In addition, Plan Sierra works with a wide cross-section of farmers in the project area. The test results from hypotheses 3 and 4 indicate that slope is not a discriminating property of the farming system in‘this area. However, it still may be an appropriate way to deve10p soil and water conservation technologies. Care must be taken to provide options within these technologies so they 105 may be adapted to the wide range of farming systems within a slope zone. Some aspects of the structure and operation of the project may present future problems; however, Plan Sierra staff have been able make the necessary changes and to adapt the project to the changing situation over the past 13 years. It is likely that this will continue in the future. 5.1.6 Generalizability of the results Care must be taken.in the generalization.of these results to the entire project area. First of all, this study was confined to activities in the humid zone. Consequently, while the results may be extrapolated to other communities within this zone, extrapolation to other zones may result in an inappropriate characterization of the situation vand the recommendation of inappropriate practices. Second, two of the three communities used. in this investigation were chosen from a list of seven communities provided by Plan Sierra. As a consequence, results from these two areas can only be appropriately extrapolated to other communities on the list, not to all communities in the area. The results from the community chosen at random may be extrapolated to the entire humid zone. However, because of the large geographic extent of the Plan Sierra project area and the likely differences, access to transportation for example, between sites, the generalizability of data collected from a Single site is limited. 106 These limitations are particularly important for the results from hypotheses 1 and 2 where bias seemed to have the largest effect. Since very little evidence of bias was found over the vast majority of farming system definition variables, the researcher believes that the results of hypotheses 3 and 4 are representative of the humid zone. 1 5.2 Recommendations for future research During the analysis of the data collected for this investigation, several possible areas of future research have been identified both within and outside of the Plan Sierra project. 5.2.1 Research within Plan Sierra The first area of research within Plan Sierra is additional research related to the hypotheses of this study. This analysis found interaction with Plan Sierra to be a strong indicator of the adoption of soil and water conservation practices. However, this investigation did not address the numerous other factors which are believed to infLuence adoption. Systematic investigation. of these other factors such as farm size, income, farmer age, and family size could provide a much better understanding of the adoption process in the project area. Additional investigation is also needed in the Plan Sierra area regarding the determination of recommendation domains within the farming system. Since this study found that slope class is not a discriminating characteristic, research is necessary to help to identify what other 107 variables, if any, can be used to define recommendation domains for planning purposes at the project, regional, and national levels. In a related area to those already discussed, future research is necessary to determine the specific soil and water conservation practices being used in Plan Sierra and their actual benefits, particularly their effectiveness in reducing soil erosion. This study has shown that Plan Sierra is effectively promoting the use of these practices. However, the data collected for this study do not permit the researcher to determine the utility of various practices. Another potentially interesting area of research involves work in the other two climate zones found within the Plan Sierra project area. This study concentrated exclusively in the humid zone, but the project has also done considerable work in the transition and dry zones. Similar work to this study within those zones might prove to be instructive as would comparisons between the zones. Such comparisons could be used to determine if climatic characteristics might be a more appropriate way to delineate recommendation domains. A final area of research involves the investigation of the role of women within the farming operations in this region. In many parts of the world, women have been shown to control a wide variety of farm activities (N. Axinn, 1988). SYBtematic research into the gender-based division of labor Conld provide valuable information which could assist Plan Sierra in the planning and implementation of future programs. 108 5.2.2 Research outside Plan Sierra With respect to all of the areas discussed previously, investigation of other similar projects, particularly in the Dominican Republic, would also be extremely beneficial and allow for comparative analysis. There are a large number of upland development projects both in the Dominican Republic and elsewhere experiencing varying degrees of success. Research focusing on comparisons between the approaches used by different projects can help identify the most appropriate methods and approaches for future development efforts. 5.2.3 Final recommendations As is evident from the above ideas, there is a large amount of potential information that can be learned from Plan Sierra. A cooperative effort between project personnel and members of the academic community has tremendous potential to increase the understanding of this successful project and of upland development in general. Such understanding can enhance and improve ongoing project operations and can greatly assist ‘the11planners and implementors of future projects. B IBLIOGRAPHY BIBLIOGRAPHY Anderson, J. R, and.Thampapillai, J., 1990, “Soil Conservation In Developing Countries: Project and Policy Intervention," Policy and Research Series No. 8, Washington DC: The World Bank. Altieri, M. A., 1990, "Stabilizing Hillside Farming Systems in the Sierra of the Dominican Republic,” pp. 355-363 in Research Heads and Applications to Rmduce Erosion and Sedimentation in Tropical Steeplands, Proceedings of the Fiji Symposium, June, 1990, R. R. Ziemer, C. L. O’Loughlin, and'L. S. Hamilton, Editors, Oxfordshire, UK: IAHS Press. Axinn, G. H., 1988, “International Technical Interventions in Agriculture and Rural Development: Some Basic Trends, Issues, and Questions,” Agriculture and.Human'Values, V. 5, n. 1 and 2, Winter-Spring, 1988: pp. 6-15. Axinn, G. H., 1978, "Rural Renaissance,” Chapter 2, pp. 15-280 in New Strategies for Rural Development, Kathmandu, Nepal and East Lansing, MI: Rural Life Associates. Axinn, N. W., 1988, “Gender Related Issues in International Development Assistance for Agriculture and Rural Life,” Agriculture and Human Values, V. 5, n. 1 and 2, Winter- Spring, 1988: pp. 69-76. Blalock, H. M. Jr., 1972, Social Statistics, 2nd Ed., New York: McGraw-Hill Book Company. Blustain, H., 1982, ”Social Issues in Technology Choice: Soil Conservation in Jamaica," Journal of Soil and Water Conservation, November-December, 1982: pp. 323-325. Brush, S. B. and B. L. Turner Jr., 1987, "The Nature of Farming Systems and Views of Their Change," Chapter 2, pp. 11-48, in Comparative Farming Systems, B. L. Turner Jr. and S. B. Brush, Editors, New York: The Guilford Press. ‘ 109 110 Carrasco, Domingo, 1991, "The Role of Hillside Farmers in .Achieving Sustainable Watershed.Management: A Dominican Republic Example" , Unpublished PhD Dissertation, Michigan State University, East Lansing, MI. Chambers, R., 1992, "Methods for Analysis by Farmers: The Professional Challenge,“ Paper presented at the 12th Annual AFSR/E Symposium, East Lansing, MI. Chambers, R., 1983, Rural Development - Putting the Last First, London: Longman Group, Ltd. Charlton, S. E. M. , 1984, "Development Strategies,“ Chapter 8, pp. 175-197, in Women in Third Werld Development, Boulder, CO: Westview Press. Christensen, L. A. and P. E. Norris, 1983, "Soil Conservation and Water Quality Improvement: What Farmers Think," Journal of Soil and Water Conservation, January-February, 1983: pp. 15-20. Collinson, M. P., 1982, ”Farming Systems Research in Eastern Africa: The Experience of CIMMYT and Some National Agricultural Research Services, 1976-81," MSU International Development Paper No. 3, East Lansing, MI: Department of Agricultural Economics, Michigan State University. de Janvry, A. and S. Hecht, 1984, “Reporte de la Evaluacion de los Primeros Cinco Afios del Plan Sierra“, unpublished evaluation, San Jose de Las Matas, Dominican Republic. Ervin, C. A. and D. E. Ervin, 1982, “Factors Affecting the Use of Soil Conservation Practices: Hypotheses, Evidence, and Policy Implications,“ Land Economics, V. 58, n. 3: pp. 277-292. FAO, 1989, Farming Systems Development: Concepts, Methods, and Applications, Rome: FAO. Feder, G., R. E. Just, and D. Zilberman, 1985, ”Adoption of Agricultural Innovations in Developing Countries: A Survey, " Economic Development and Cultural Change, V. 33, n. 2: pp. 255-298. Flora, C. B., 1988, "Farming Systems Approaches to International Technical Cooperation in Agriculture and Rural Life, " Agriculture and Human Values, V. 5, n. 1 and 2, Winter-Spring, 1988: pp. 24-34. Friedrich, K. H., 1992, “A Farming Systems Approach to Agricultural Development,” pp. 107-114 in Readings in Farming Systems Development, Rome: FAO. lll Friedrich, K. H., and M. Hall, 1992, “Farming Systems Development: FAO's Reaction to the Challenge of Developing Sustainable Farm-Household Systems, " pp. 17-32 in Readings in Farming Systems Development, Rome: FAO. Garrity, Dennis P., 1991, “Sustainable Land Use Systems for the Sloping Uplands of Southeast Asia“, in Technologies for Sustainable Agriculture in the Tropics, Madison: American Society of Agronomy (forthcoming). Gilbert, E. H., D. W. Norman, and F. E. Winch, 1980, “Farming Systems Research: A Critical Appraisal,“ MSU Rural Development Paper No. 6, East Lansing, MI: Department of Agricultural Economics, Michigan State University. Hansen, D. O., J. M. Erbaugh, and T. L. Napier, 1987, "Factors Related.to.Adoption.of Soil Conservation.Practices in the Dominican Republic, “ Journal of Soil and Water Conservation, September-October, 1987: pp. 367-369. Hartshorn, G., G. Antonini, R. du Bois, D. Harcharik, S. Heckadon, H. Newton, C. Quezada, J. Shores, and G. Staples, 1981, The Dominican Republic: Country Environmental Profile, Report for USAID, McLean, Virginia: JRB Associates. Harwood, R. R., 1992, "The Structure of Biological Diversity at the Agricultural, Environmental, and Social Interface,“ Keynote address: Diversity in Food, Agriculture, Environment, and Health, Michigan State University, East Lansing, MI, June 4-7, 1992. Hauck, F. W., 1985, 7Soil Erosion and its Control in Developing Countries,“ Chapter 69, pp. 718-728, in Soil Erosion and Conservation, S. A. El-Swaify et. al., Editors, Ankenyy IA; Soil Conservation Society' of America. .Havelock, R. G., 1971, Planning for Innovation through Dissemination and Utilization of Knowledge, Ann Arbor, MI: Institute for Social Research, The University of Michigan. Hondale, G. and J. VanSant, 1985, Implementation for Sustainability: Lessons from. Integrated Rural Development, West Hartford, CT: Kumarian Press. Hudson, N. W., 1983, ”Soil Conservation Strategies in the Third World,“ Journal of Soil and Hater Conservation, November-December, 1983: pp. 446-450. Hudson, N. W., 1977, Soil Conservation and Management in Developing Countries, FAO Soils Bulletin 33, Rome: FAO. 112 Jimenez, A. and L. Gutierrez, 1993, Notes taken by Michael Robotham from a presentation by A. Jimenez and subsequent discussions with A. Jimenez and L. Gutierrez at Plan Sierra offices, 10 March, 1993. Jo, I. S., 1991, "Hillslope Farming in Korea”, pp. 81-90 in Development of Conservation Farming on Hillslopes, W. C. Moldenhauer et. al, Editors, Ankeny, Iowa: Soil and Water Conservation Society. Kemph, G. S., and Hernandez, A., 1987, “Evolutionary Conservation Project Planning and Implementation: NARMA in the Dominican Republic,“ Chapter 9, pp. 113-128, in Sustainable Resource Development in the Third World, D. D. Southgate and. J. F. Disinger, Editors, Boulder: Westview Press. Laquihon, W. A., 1989, “Some Key Determinants of SALT Adoption in the Philippines: ‘Viewpoints of Farmer Cooperators", pp. 79-116 in Social Forestry in Asia: Factors that Influence Program Implementation, N. T. Vergara and R. A Fernandez, Editors, Los Bafios, Laguna, Philippines: SEARCA. Liao, M. C., S. C. Hu, and H. L. Wu, 1991, "Soil Conservation Research and Development for Hillslope Farming in Taiwan " pp. 12-16 in Development of Conservation Farming on Hillslopes, W. C. Moldenhauer et. al, Editors, Ankeny, Iowa: Soil and Water Conservation Society. Lin, Y. L., 1991, “wusheh Watershed Management,“ pp. 270-276, in Development of Conservation Farming on Hillslopes, W. C. Moldenhauer et. al., Editors, Ankeny, Iowa: Soil and Water Conservation Society. Mac Donald, A. L. , 1976, Agricultural Technology in Developing Countries, Rotterdam: University Press. .Mercer, D. E., 1992, "The Economics of.Agroforestry,“ Chapter 7, pp. 111-143, in Social Science Applications in Asian Agroforestry, W. R. Burch Jr. and J. K. Parker, Editors, Petit Jean MOuntain, AK: Winrock International. Mokma, D. L., D. P. Krauss, and D. L Cremeens, 1983, “Soil Classification Manual for CSS-470", Department of Crop and Soil Sciences, Michigan State University, East Lansing, MI. ‘ IMoldenhauer, W. C. and N. W. Hudson, Editors, 1988, Conservation Farming on Steep’Lands, Ankeny, Iowa: Soil and Water Conservation Society. 113 Mollema, A., 1988, "Reporte de la Investigation Agrosociologica Adaptada," Universidad Agraria de Wageningen, Holanda, (unpublished project report). Murphy, Kevin, and Monica de Castro, 1992, "Case: Plan Sierra", Santiago: Instituto Superior de Agricultura. Napier, T. L. and S. M. Camboni, 1993, "Use of Conventional and Conservation Farming Practices among Farmers in the Scioto River Basin of Ohio," Journal of Soil and Water Conservation, V. 48, n. 3: 231-237. Napier, T. L., A. S. Napier, and M. A. Tucker, 1991, “The Social, Economic and Institutional Factors Affecting Adoption of Soil Conservation Practices: the Asian Experience," Soil and Tillage Research, V. 20, pp. 365- 382. Ott, L. , 1988, An Introduction to Statistical Methods and Data Analysis, Third Edition, Boston: PWS-Kent Publishing Company. Plan Sierra, 1982, "Proposal for the Overall Development Plan of the Sierra", mimeograph, San Jose de Las Matas, Dominican Republic. ' Plan Sierra, 1991, ”Resumen Sobre el Plan Sierra", unpublished manuscript, San Jose de Las Matas, Dominican Republic. Raintree, J. B., 1983, "Strategies for Enhancing the Adoptability of Agroforestry Innovations,“ Agroforestry Systems, V. 1: pp. 173-187. Ravnborg, H. M., 1992, "Resource Poor Farmers: finding them and diagnosing their problems and opportunities”, pp. 175-190 in Toward a New Paradigm for Farming Systems Research/Extension, Working Papers Set for the 12th Annual Farming Systems Symposium, Michigan State University, East Lansing, MI. Repollo, A. Q. Jr. and E. R. Castillo, 1989, ”Agroforestry Technology in Hilly Land Households: Factors Influencing Its Adoption,“ pp. 118-132 in Social Forestry in Asia: Factors that Influence Program Implementation, N. T. Vergara and R. A Fernandez, Editors, Los Bafios, Laguna, Philippines: SEARCA. Retana, G., 1982, ”Plan Sierra", Programa de Escritura de Casos, Centro de Administracion del Desarrollo Rural, Instituto Superior de Agricultura, Santiago, Dominican Republic. 114 Rimwanich, S., 1991, “Development of Hilly Land in the Northern Region.of Thailand“, pp. 61-67 in.bevelopment.of Conservation Farming on Hillslopes, W. C. Moldenhauer et. a1, Editors, Ankeny, Iowa: Soil and Water Conservation Society. Robotham, M., 1991, "The NARMA Project," Unpublished case study report, RD876, Department of Resource Development, Michigan State University, East Lansing, MI. Rocheleau, D., 1984, ”An Ecological Analysis of Soil and Water Conservation in Hillslope Farming Systems", PhD Dissertation, University of Florida, Gainesville, FL. Rogers, E., 1983, Diffusion of Innovations, Third.Edition, New York: The Free Press. Ruttan, V: W., 1984, ”Integrated Rural Development.Programmes: A Historical Perspective," Werld Development, V. 12, n. 4: pp. 393-401. Santos, B., 1980, “El Plan Sierra: una experiencia de desarrollo rural en las montafias de la Repfiblica Dominicana,” pp. 285-294, in Agricultura de Ladera an America Tropical, A. R. Novoa and J. L. Posner, Eds., Turrialba, Costa Rica: CATIE. Santos, B. and N. Quezada, c1979, "Report on Plan Sierra, Dominican Republic”, Santiago: Instituto Superior de Agricultura. Shaner, W. W., P. F. Philipp, and.W. R. Schmehl, 1982, Farming Systems Research and Development: Guidelines for Developing Countries, Boulder, CO: Westview Press. Shaxson, T. F., 1988, “Conserving Soil by Stealth,“ Chapter 2, pp. 9-17, in Conservation Farming on Steep Lands, W. C. Moldenhauer and N. W. Hudson, Editors, Ankeny, Iowa: Soil and Water Conservation Society. Sheng, T. C., 1989, Soil Conservation for*Small Farmers in the Humid Tropics, FAO Soils Bulletin No. 60, Rome: FAO. Sheng, T. C., 1988, “Demonstrating Conservation Practices on Steep Lands in Jamaica,“ Chapter 22, pp. 207-214, in Conservation Farming on Steep Lands, W. C. Mbldenhauer and N. W. Hudson, Editors, Ankeny, Iowa: Soil and Water Conservation Society. Singleton, R. Jr., B. C. Straits, M. M. Straits, and R. J. McAllister, 1988 , Approaches to Social Research, New York: Oxford University Press. 115 Tropical Research and Development Inc., 1992, “Intensive Survey of Rural and Urban Activities Impacting Water and Coastal Resources", Document PDC-5517-I-00-0105-00, USAID, Dominican Republic. Veloz, A., D. Southgate, F. Hitzhusen, and.R. Macgregor, 1985, "The Economics of Erosion Control in a Subtropical Watershed: A Dominican Case," Land Economics, V. 61, n. 2: pp. 145-155. Watkins, W. A., 1991, “Conservation Hillslope Farming in Australia", pp. 68-80 in Development of Conservation Farming on Hillslopes, W. C. Moldenhauer et. al, Editors, Ankeny, Iowa: Soil and Water Conservation Society. World Bank, 1992, “Strategies and Technologies for Asian Watersheds“, Paper prepared by the Land Resources Unit, Agricultural Division, Asia Technical Department, Washington, DC: World Bank Publications. APPENDICES Appendix Av Table 12 Table 13 Table 14 116 Detailed output from the various statistical analyses conducted for this study swcuse by pshear for all farmers PSHEAR Count Yes No Row 1.00 2.00 Total Yes 43 5 48 SWCUSE 1.00 58.5 No 3 2.00 46 56.1 swcuse by pshear for farmers not on lists provided by Plan Sierra I'— SWCUSE 1.00 PSHEAR Count Yes No 1.00 2.00 Yes 25 5 swcuse by pshear for farmers in Las Piedras only Count Yes 1.00 . SWCUSE NO 2.00 117 Table 15 swcuse by psextend for all farmers __= PSEXTEND Count Yes No Row 1.00 2.00 Total Yes 36 12 48 SWCUSE 1.00 58.5 No 0 34 34 2.00 41.5 36 48 82 43.9 56.1 100.0 Table 16 swcuse by psextend for farmers not on lists provided by Plan Sierra .1 swcuse by psextend for farmers in Las Piedras Only PSEXTEND PSEXTEND Count Yes No Row 1.00 2.00 Total : Yes 18 12 30 5 SWCUSE 1.00 47.6 ' No 0 33 33 2.00 52.4 45 63 71.4 100.0 2000 Count Yes No Row 1.00 2.00 Total Yes 12 10 22 ~ SWCUSE 1.00 52.4 No 0 20 20 Table 18 Table 19 Table 20 118 PSTRAIN swcuse by pstrain for all farmers 2.00 Count Yes No Row 1.00 2.00 Total Yes 17 31 48 SWCUSE 1.00 58.5 No 0 34 34 PSTRAIN swcuse by pstrain for farmers not provided by Plan Sierra on lists Count Yes 1.00 ‘ Yes 7 23 30 SWCUSE 1.00 47.6 ' No 0 33 33 2.00 52.4 swcuse by pstrain for farmers in Las Piedras only PSTRAIN Count Yes No Row 1.00 2.00 Total . Yes 5 17 22 . SWCUSE 1.00 52.4 No 0 20 20 2.00 47.6 119 Table 21 stilluse by pshear for all farmers PSHEAR -Count Yes No Row 1.00 2.00 Total Yes 42 5 47 STILLUSE 1.00 97.9 No 1 0 1 2.00 2.1 43 5 48 89.6 10.4 100.0 Table 22 stilluse by pshear for farmers not on lists provided by Plan Sierra PSHEAR Count Yes 1.00 24 Yes STILLUSE 1.00 No 2.00 Table 23 stilluse by pshear for farmers in Las Piedras only . STILLUSE 120 Table 24 stilluse by psextend for all farmers PSEXTEND Count Yes No Row 1.00 2.00 Total Yes 35 12 47 STILLUSE 1.00 97.9 No 1 0 1 2.00 2.1 36 12 48 75.0 _-25.0___ 100.0 Table 25 provided by Plan Sierra PSEXTEND stilluse by psextend for farmers not on lists Count Yes No Row 1.00 2.00 Total 1 Yes 17 12 29 ; STILLUSE 1.00 96.7 No 1 0 1 2.00 Table 26 stilluse by psextend for farmers in Las Piedras only PSEXTEND Count Yes No Row 1.00 2.00 Total 1 Yes 11 . 10 21 STILLUSE 95.5 ' 0 1 4.5 10 45.5 22 100.0 121 Table 27 stilluse by pstrain for all farmers _ ——=__ PSTRAIN Count Yes No Row 1.00 2.00 Total Yes 17 30 47 STILLUSE 1.00 97.9 No 0 1 1 2.00 2.1 17 31 48 35.4 64.6 100.0 Table 28 Count PSTRAIN Yes 1.00 No 2.00 stilluse by pstrain for farmers not on lists provided by Plan Sierra = STILLUSE Table 29 only Yes 7 PSTRAIN Yes 1.00 stilluse by pstrain for farmers in Las Piedras i STILLUSE 5 Table 30 Summary of hypothesis 3 test results for individual variables 122 Variable' Significant Variation in the Variable. Significant Variance in the Variable within the S lope C lass Variable wi thiB the S lope (Yes/No) Class (Yes/No) slclassc 182 3 4 5 6 slclassc 182 3 #4 I I fa rnsi 2e Yes Yes No Yes No pest i cid No No No No No crdi vers Yes Yes Yes Yes Yes herbi c i d No No No No No cafe No No No No No ferti l i 2 Yes Yes Yes Yes No yuca Yes No Yes No No credit Yes Yes No No Yes - a mai s Yes Yes Yes Yes Yes problem Ya No Yes No Yes beans Yes Yes Yes Yes Yes swcuse Yes Yes Yes Yes Yes i 1:: IE! banana Yes Yes Yes Yes Yes p,” st No No No No No c rprat i 0 Yes Yes Yes Yes Yes pshear Yes Yes Yes Yes Yes andivers Yes Yes Yes Yes No psextend Yes Yes Yes Yes Yes ggl lines No No No No No pst rai 11 Yes No Yes No No 00"“. Yes Y” Y.’ Yes m fat-”qr m Ye; Ye. YO. "a vacas No Yes Yes Yes Yes yrsfarm Yes Yes Yes Yes Yes mule Yes Yes Yes No No famsi ze Yes Yes Yes Yes Yes anmrat i 6 Yes Yes Yes Yes Yes emigrate Yes Yes Yes Yes Yes No y” I Yes No income Yes Yes Yes Yes Yes — — — —=—==- T a: The descriptions and derivation criteria for each of these variables may be found in Table 2. b: Significant variation was determined using the criteria outlined in section 3.8.5. o: The variable slclass indicates the FAO slope class, the class parameters can be found in Table 3. 123 Appendix B: Copy of the survey instrument used in this study‘ 2-- - -- x: ---- Title Page MSU-ISA EXPLORATOR! SURVE! PLAN SIERRA MARCH, 1993 Questionnaire Number Interview Zone Slope Climate Zone Date of the Interview Interviewer Number Page 1 Section I FARH CHARACTERISTICS Sub-section 1.1 Sketch of the Farm ‘The survey was administered in Spanish, an English translation is provided here for the convenience of the reader. A complete Spanish text is available from the author upon request. 124 Sub-section 1.2 Basic Farm Information 1. How many tareas do you cultivate? 2. How many parcels does your farm contain? 3. What is the size of each parcel? 3.1 Parcel #1 tareas 3.2 Parcel #2 tareas 3.3 Parcel #3 tareas 3.4 Parcel #4 tareas 4. Approximately how far is each parcel from you house? (specify the units used by the farmer, minutes or meters) 4.1 Parcel #1 4.2 Parcel #2 4.3 Parcel #3 4.4 Parcel #4 5. Approximately how long have you operated your farm? 5.1 Parcel #1 years 5.2 Parcel #2 years 5.3 Parcel #3 years 5.4 Parcel #4 years 125 Sub-section 1.3 Land ownership Please respond to the following QUBSIlOKS for each parcel: 6. Who owns the parcel? _ r s Number of the parcel 1 2 3 4 1) You 2) Member of your immediate family 3) Local owner 4) Owner from outside the community 5) The government 6) Other, please specify 7. In what form do you pay the rent and how much do you pay per year? Number of the parcel 1 2 3 4 1) Pay in cash 2) Pay part of the harvest 3) Other, please specify w 8. How did you acquire the rights to use the land? Number of the parcel ' 1 2 3 4 1) Purchase 2) Inheritance 3) Received as a gift i 4) Other, please specify 5 0W9“ M- __J__J 126 Sub-section 1.4 Agricultural Production Please complete the table below with the farmer's responses to the following questions. 9. What crops do you plant? 10. 'When do you plant each crop? 11. When do you harvest each crop? 12. What type of agriculture do you use? eg: Monoculture, Intercropping (on the same field), Alternate cropping 13. What is the principle use of each crop? eg: sale, home use, both Type of Crop Approximate Approximate Type of Principle use Planting Date Harvest Date cultivation 9.1 10.1 11.1 12.1 13.1 9.2 10.2 11.2 12.2 13.2 9.3 10.3 11.3 12.3 13.3 9.4 10.4 11.4 12.4 13.4 9.5 10.5 11.5 12.5 13.5 9.6 10.6 11.6 12.6 13.6 9.7 10.7 11.7 127 13.7 9.8 10.8 11.8 12.8 13.8 9.9 10.9 , 11.9 12.9 13.9 i, 127 14. Classify the level of importance of the different problems with your crops in the following table: No. Description of Problee Very Inportant Not Important Important 14.1 Difficult to obtain seeds 1 2 3 14.2 Difficult to obtain inputs (fertilizer, 1 2 3 herbicides, pesticides) 14.3 Difficult to obtain credit 1 2 3 14.4 Difficult to sell the harvest 1 2 3 14.5 Soil Erosion 1 2 3 14.6 Flooding 1 2 3 14.7 Drought 1 2 3 14.8 Pests 1 2 3 14.9 weeds 1 2 3 14.10 Other, please specify 1 2 3 —— L j—L 14.11 What is the most important problem with your crops? (number) 14.12 Why? 128 15.1 Does your land have different types of soil? 1. Yes 2. No, If "no“ go to question 16.1 15.2 Do you plant your crops in relation to the type of soil? 1. Yes i 2. No 15.3 Why or why not? 16.1 Does your land have different levels of slope? 1. Yes 2. No, if "no“ go to question 17 15.2 Do you plant your crops in relation to the level of slope? 1. Yes 2. No 15.3 Why or why not? If the farmer uses the system of coffee cultivation under the shade of guama (lag; ygng), respond to questions 97-101. Sub- 129 Sub-section 1.5 Animals 17. Do you have animals? 1. Yes 2. No, if "no" go to question 20 Respond yes or no for each animal and indicate the use or uses in the table below TypeofAnimal Yes No UseorUsesofthcAnimal 17.1 Chickens 18.1 17.2 Pigs . 18.2 17.3 Goats 18.3 17.4 Cows 18.4 17.5 Horse, donkey or mule 18.5 17.6 Other 18.6 130 Suhésection 1.6 Questions relating to trees 20. Do you have trees on your farm? 1. Yes 2. No, if ”no" go to question 26. 21. Classify the level of importance of the different uses of trees in the following table. Fuclwood or charcoal 2 21.2 Construction materials 1 2 3 21.3 Sell to make money 1 2 3 2L4 Ham. 1 2 3 21.5 Forage 1 2 3 21.6 To improve the harvest 1 2 3 21.7 Shade 1 2 3 21.8 $011 and water conservation 1 2 3 2L9 resume 1 2 3 21.10 Cord or Rope 1 2 3 21.1 1 Agricultural implements l 2 3 21.12 Kitchen utensils - 1 2 3 21.13 Fencing 1 2 3 21.14 Other, please specify 1 2 3 ——---:==--——=w 21.15 What is the most important use of your trees? (number) 21.15 Why? 22. 23. 24. 25. Have 131 you planted any trees in the past five years for any of the following uses? 22.1 22.2 22.3 22.4 22.5 22.6 Have past 23.1 23.2 23.3 23.4 23.5 23.6 23.7 What 24.1 24.2 24.3 24.4 24.5 N (D 01 No Fuelwood or charcoal Shade Fruit Fencing Lumber Other HHD—‘i—‘I—‘l—‘l NNNNNN you sold any of the following tree products in the year? X§§ £2 Fuelwood or Charcoal 1 2 Lumber 1 2 Fruit 1 2 Fibers 1 2 Wood products (eg furniture) 1 2 Medicine 1 2 Other 1 2 do you think about the following statements? Xfifi £2 Trees improve your harvest. 1 2 Trees improve your livestock. 1 2 Trees reduce the space available for your crops. 1 2 Trees reduce soil loss. 1 2 Trees reduce the water available for your crops. 1 2 Do you have sufficient trees on your farm? 1. 2. Yes NO 132 Sub-section 1.7 Soil and Water Conservation and 26. 27. 28. 29. Management Practices. Have you heard about soil and water conservation practices? 1. Yes 2. No if "no" go to question 53 Where did you hear about these practices for the first time? (Circle only one response) Family Friend Government NCO, please specify University or institute Newspaper/radio/television Other, please specify \lOtU'thNl-J Where else have you heard about these practices? (Circle all responses which apply) 1. Family 2. Friend 3. Government 4. WOO, please specify 5. University or institute 6. Newspaper/radio/television 7. Other, please specify Do you use any soil and water conservation practices on your farm? 1. Yes 2. No, 29.1 Why not? Go to question 53 133 IF THE FARMER HAS EVER USED SOIL AND WATER CONSERVATION PRACTICES ON THEIR FARM 30. Which soil and water conservation practices have you used? (Circle all the responses which apply) Contour cultivation 1. 2. Crop rotation 3. Vegetative barriers 4. Agroforestry practices 5. Dead barriers 6. Diversion canals 7. Terraces 8. Other, please specify‘ 30.1 Which of the practices mentioned above did you use first? 30.2 Why? 31. In what type of parcels do you use soil and water conservation practices? In all parcels In parcels with steep slopes In parcels with moderate slopes In parcels with shallow slopes #NNH 32. In which type of crops do you use soil and water conservation practices? All crops Crops grown for home use Crops grown for sale Trees #WNH 33. 34. 35. 134 Where did you obtain information about how to implement these practices? (circle all the responses which apply) \JO‘U‘ D “NH 33. 33.2 Why? From my family From my friends From an extension agent who visited my farm Specify from where Attended a training Specify where and when Written materials Experimentation Other, please specify Which of the above sources provided the most useful information? (number) Approximately how long ago did you first use these practices? Years Why did you first use these practices on your own farm? (Circle all the responses which apply) 1. 2. 3. 4. 5. Incentives Recommendation from family or friends Recommendation from an agency or NGO Which agency? Because I wanted to resolve problems on my land Other, please specify 35.1 What was the most important reason? (number) 35.2 Why? 36. 37. 38. 135 ' Did the practices create benefits? 1. Yes 2. No 36.1 What benefits do the practices create? (circle all the responses which apply) Erosion control Increase soil moisture Increase soil fertility Provide forage for livestock Provide lumber for home use Other, please specify GUI-#001014 36.2 Which of these is the most important? (number) 36.3 Why? Were there problems with the practices? 1. Yes 2. No 37.1 What were the problems? 37.2 What did you do to resolve them? Did you make changes to try to improve the practices? 1. Yes 2. No 38.1 What did you do? 38.2 Did it work? 136 39. Do you use soil and water conservation practices at the present time? 1. Yes, if "yes" answer questions 44, 46, y 485 2. No, if "no" answer question 49 FOR FARMERS WHO STILL USE SOIL AND WATER CONSERVATION PRACTICES. 44. Do you use the practices every year or only in some years? 1. Use every year 2. Use some years but not every year 44.1 Why? 46. If you have questions concerning soil and water conservation practices, who do you ask? (Circle all responses which apply) 1. Family 2. Friends 3. Government agency: Which one? 4. WOO: Which one? 5. Private voluntary organization: Which one? 6. Other, please specify 46.2 Where do you obtain the most useful information? (number) 46.3 Why? 48. Do you believe that the limitation of soil erosion on your farm has resulted in: 1&8 E2 1. Increase in production 1 2 2. Decrease in production 1 2 ' 3. Increase the life of the soil 1 2 4. Increase soil moisture 1 2 5. No effect 1 2 Go to question 53 5Questions 40-43, 45, and 47 were eliminated after the first day of surveys on the advice of the interviewers. The answers provided for these questions were identical to the answers given for'earlier questions. 137 FOR FARMERS WHO HAVE USED SOIL AND WATER CONSERVATION PRACTICES IN THE PAST BUT NO LONGER USE THESE PRACTICES 49. How long did you use soil and water conservation practices? years 50. When did you stop using the practices? (year) 51. Were there problems with the practices? Yes No NH e e 51.1 If "no“ why did you stop using them? 51.2 If ”yes“ What problems did you have? (Circle all those which apply) 1. The practices reduce the space for your crops 2. The practices cost too much to maintain 3 The trees compete with your crops 4: Other, please specify 51.3 What was the biggest problem? (number) 51.4 Why? 52. Did you try to modify the practices? 1. Yes 2. No if "no“ go to question 53 52.1 What did you do? 52.2 Did it work? 1. Yes 2. No 52.3 Why or why not? 138 Sub-section 1.8 Use of Inputs Sub-section 1.8.1 Use of Labor 53. Do you need additional people to complete all the work needed on your land? 1. Yes 2. No, go to question 57 54. For which tasks do your require help? 1. Soil preparation 2. Planting 3. Weeding 4. Harvesting 5. Other, please specify 55. During which months do you require extra help? 1. January to March 2. April to June 3. July to September 4. October to December 56. Who provides the assistance? 1. Family members 2. Friends 3. Hired help 139 Sub-section 1.8.2 Use of Pesticides 51. 58. 59. What type and what amount of pesticides do you utilize in your land?. 1. I do not use pesticides (go to question 60). 2. , amount/tarea 3. , amount/tarea 4. , amount/tarea During which months of the year do you use pesticides? January to March April to June July to September October to December kNNH In what type of crops do you use pesticides? 1. All crops 2. Crops grown for food 3. Crops grown for sale 4. Trees IF THE FARMER DOES NOT UTILIZE PESTICIDES 60. If you do not use pesticides, why not? 1. They are not necessary 2. I can not find them when I need them 3. They are very expensive 140 Sub-Qsection 1.8.3 Use of Herbicides 61. 62. 63. What type and what amount of herbicides do you use on your farm? 1. I do not use herbicides. (go to question 64). 2. , amount/tarea 3. , amount/tarea '4. , amount/tarea During which months of the year do you use herbicides? . January to March April to June July to September October to December fiUNl—i On what type of crops do you utilize herbicides? 1. All crops 2. Crops grown for food 3. Crops grown for sale 4. Trees IF THE FARMER DOES NOT USE HERBICIDES 64. If you do not use herbicides, why not? 1. They are not necessary 2. ' I can not find them when I need them 3. They are very expensive 141 Sub-section 1.8.4 Use of fertilizers 65. What type and what amount of fertilizers do you use on your farm (chemical and organic)? 1. I do not use fertilizer. (go to question 67). , amount/tarea , amount/tarea , amount/tarea 66. In what type of crops do you utilize fertilizers? hWNI—J All crops Crops for consumption Crops for the market . Trees IF THE FARMER DOES NOT USE FERTILIZERS 67. If you do not use fertilizers, why not? 1. They are not necessary 2. I can not find them when I need them 3. They are very expensive Sub-section 1.8.5 Training 68. Have you received any kind of training with relation to the application of: 68.1 68.2 68.3 Pesticides Yes No From whom? Herbicides Yes No From whom? Fertilizers Yes No From whom? 142 Sub-section 1.8.4 Use of Credit 69. 70. 71. 72. 73. Do you use credit for your agricultural practices? 1. 2. Yes No, if “no" answer question 74 How much do you borrow every year? \lO’iUl-FUJNH less than $1,000 pesos $1,000-$3,000 pesos $3,001-$5,000 pesos $5,001-$7,000 pesos $7,001-S10,000 pesos $10,001-S15,000 pesos More than $15,000 pesos Where do you obtain it? bUNI—l Agricultural Bank (Dominican Government) Commercial Bank NGO, Which one? Other, specify In what form is the credit given? #NNH What (”#105214 74.1 74.2 Cash Seeds Inputs (fertilizers, herbicides, Others) Others, please specify types of crops require credit? All crops Crops grown for food Crops grown for sale Trees Animals What are the interest charges on your credit? 1. pesos for each 100 pesos 2. portion of the harvest This credit is: For how long? How many times? 75. 143 If you do not use credit, why not? It is not necessary I am afraid of debt/afraid of losing my land I cannot obtain credit Interest rates are too high Other, please specify Ul-hUND-i e e e e e 144 Section 2 Use and Availability of Water 76. 77. 78.1 79. Where do you obtain your drinking water? 1. From your own well 2. From a river or stream 3. From a lake 4. From rain water 5. From a community well If you obtain the water from a well, how deep is the 1. 3 to 5 meters 2. 6 to 10 meters 3. 11 to 15 meters 4. 16 to 20 meters 5. 21 meters or more Do you have enough water for the entire year? 1. Yes 2. No 78.2 If you do not have enough water for the entire year, which are the months when water is scarce? January to March April to June . July to September . October to December bWNH Do you use agricultural practices that you believe increase the availability of soil moisture? 1. Yes 2. No (go to question 84) 145 80.1 What praCtices do you use? (Circle all the answers that apply) 81. 82.1 83. 1. 2. 3. 4. 5. 80.2 80.3 What 1. 2. 3. 4. 5. Mulching Minimum tillage Canals and catch basins to collect rain water Irrigation Other, specify Which one of these practices do you use the most often? (number) Why? type of irrigation do you use on your farm? I do not use (go to question 83) Drip irrigation Flooding Spraying Other, please specify Do you have, either by yourself or with some neighbors, a reservoir to store rain water? 1. Yes 2. No 82.2 If you have one, How big is it? x meters or gallons Who insures that the irrigation system is working? 1. You 2. You and your neighbors 3. Other, please specify 146 Section 3 Demographic Information 84.‘ How many members are in your household? 1. adults 2. children 85. What are their ages? 1. Yourself 2. Your wife/husband 3. Your children , , , , 4 . Others, specify 86.1 Where were you born? 86.2 Where was your wife/husband born? - 87. Approximately how long have you lived in this area? years 88. How many people have left your household to go and live outside of the community? 1. adults 2. children 3. None 88.1 Where did they go? Santiago Santo Domingo Another part of the Dominican Republic please specify The United States Other, please specify U'I-F NNH 147 Section 4 Land Inheritance 89. 90. 91. 92. 93. Do you think you have enough land to support your family? 1. Yes 2. No If you responded "no“, what amount of land do you need 1. 5 - 10 tareas 2. 11 - 20 tareas 3. 21 - 40 tareas 4. More than 40 tareas Will your children continue cultivating the land after you stop doing so? 1. Yes 2. No How much longer will it be before one of your children starts cultivating the land? They are doing it now 1 - 5 years 6 - 10 years 11 - 15 years More than 16 years U'lthUNH What will happen to your land after your death? 1. It will be divided among the children in equal parts It will be divided among the children in unequal parts It will be divided as it is (in existing parcels) It will be divided in 2 parts It will be divided in 3 parts It will be divided in 4 parts I have not yet decided #0101430) N 148 Section 5 Income 94. 95. 96. Approximately how much income did you and your family receive from work on your farm during the past year? Less than $1,000 pesos $1,001-$3,000 pesos $3,001-$5,000 pesos $5,001-$10,000 pesos $10,001-$15,000 pesos $15,000-$25,000 pesos More than $25,000 pesos \IQU‘IbNNl-‘l In the past year, did you or a member of your immediate family receive payment from off-farm employment? 1. Yes 2. No If you answered “yes“, off-farm income account for approximately what part of your total income? Much less than one-half Less than one-half One-half More than one-half . Much more than one-half U'Iwal-J 149 If the Farmer Uses the System of Coffee Cultivation under the Shade of Guama (Inga verg) 97. Where did you obtain information about how to implement the system of coffee cultivation under the shade of guama trees on your farm? (Circle all responses which apply) 1. From my family 2. From my friends 3. From an extension agent who visited my farm Specify where the agent was from 4. Attended a training Where and when 5. Written materials, radio, or television 6._ Other, please specify 97. Which one of these sources provided the most useful information? (number) 97.2 Why? 98. Why did you first use the system of cultivating coffee under the shade of guama trees on your own land? (Circle all the responses which apply) UNH 5. Incentives Recommendation from family or friends Recommendation from a government agency or non-government organization Please specify which organization? Because I wanted to resolve problems on my land Other, please specify 98.1 What was the most important reason? 98.2 Why? (number) 150 99. 'Where did you obtain the coffee and guama seedlings to start your coffee cultivation system? a —— Coffee Guama Purchase From whom? Family Friends Government (eg. the Bureau of Forestry) Specify Non-government organization (eg. El Plan Sierra) Specify Other, Specify w 100. Did you need credit to start cultivating coffee? 1. Yes 2. No * . 100.1 If you responded "yes”, Where did you obtain the credit? Agricultural Bank (Dominican Government) Commercial Bank Non-government organization, please specify . Other, specify uh UNH 101. If you have questions about coffee cultivation, who do you ask? (Circle all responses which apply) 1. Family 2. Friend 3. Government agency: Please specify? 4. Non-government organization: Please specify? 5. Private volunteer organization: Please specify? 6. Other, please specify 101.2 Where did you receive the most useful information? (number) 101.3 Why? "11111111111111“