Content uploaded by Jules N Pretty
Author content
All content in this area was uploaded by Jules N Pretty on Feb 10, 2018
Content may be subject to copyright.
UNCORRECTED PROOF
2Gender and Social Capital: The Importance of Gender
Differences for the Maturity and Effectiveness of
Natural Resource Management Groups
OLAF WESTERMANN
Danish Institute for International Studies (DIIS), Copenhagen K, Denmark
JACQUELINE ASHBY
Rural Innovation Institute, Cali, Colombia
and
JULES PRETTY
*
University of Essex, Colchester, UK
Summary. —This paper seeks to contribute to an improved understanding of the gender aspects of
social capital manifested in groups for natural resource management (NRM). We investigated how
gender differentiated social groups differ in their activities and outcomes for NRM. A total of 46
men’s, mixed, and women’s groups were analyzed in 33 rural programs in 20 countries of Latin
America, Africa, and Asia. Significant gender differences were found in relation to group maturity
and NRM achievements and approaches as well as important differences in experiences of collab-
oration and capacity to manage conflict. Overall, we found that collaboration, solidarity, and con-
flict resolution all increase in groups where women are present. In addition, norms of reciprocity
are more likely to operate in women’s and mixed groups. Similarly, the capacity for self-sustaining
collective action increased with women’s presence and was significantly higher in the women’s
groups. The results demonstrate the importance of gender analysis for collective NRM and partic-
ularly the role of women for collaboration in and sustainability of NRM groups.
Ó2005 Elsevier Ltd. All rights reserved.
Key words — collective action, gender, social capital, natural resource management
27
1. INTRODUCTION
29 It is increasingly well established that social
30 capital is an important factor in building and
31 maintaining collective action (Krishna & Uph-
32 off, 1998; Pretty, 2003; Pretty & Ward, 2001;
33 Putnam, Leonardi, & Nanetti, 1993; Scoones,
34 1998; Woolcock, 1998), which is in turn funda-
35 mental to substantial and long-term changes in
36 natural resource management (NRM) (Agra-
37 wal & Gibson, 1999; Baland & Platteau, 1996;
38 Bromley, 1992; Korten, 1986; Ostrom, 1990;
39 Pretty, 2002; Pretty & Smith, 2004; Reddy,
40 2000; Steins & Edwards, 1999; Wade, 1987).
41 Analysis of causal relationships among
*We would like to thank all the people who have res-
ponded to the questionnaire survey, especially those who
have dedicated time and energy to complete the ques-
tionnaires. We would also like to thank James Silva and
James Garcia (CIAT) for their assistance with the stat-
istical analysis and Christine Kolb (CIAT) for her assi-
stance. We are also grateful to Rachel Hine (University
of Essex), Sheri Arnott (IDRC), Helen Raij (IDRC-
MINGA), Liliana Rojas (CIAT-REDECO), Peggy
McKee (PRGA), and Ana Maria Ponce (CONDE-
SAN—InfoAndina) for input regarding mailing lists and
list servers and to Anna Knox (PRGA), Christine Okali,
Gordon Prain, and Helen Hambly for their comments
on earlier versions of the paper. Final revision accepted:
April 15, 2005.
World Development Vol. xx, No. x, pp. xxx–xxx, 2005
Ó2005 Elsevier Ltd. All rights reserved
Printed in Great Britain
0305-750X/$ - see front matter
doi:10.1016/j.worlddev.2005.04.018
www.elsevier.com/locate/worlddev
1
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
42 improved resource management and collective
43 action has hitherto centered on the existence
44 or creation of appropriate institutional and
45 property arrangements (Bromley, 1992; Leach,
46 Mearns, & Scoones, 1999; Olson, 1965; Os-
47 trom, 1990), but there is an emerging recogni-
48 tion that relations of trust and common
49 values are important to collective action (Har-
50 ris & Renzo, 1997; Lyon, 2000; Pretty & Ward,
51 2001; Uphoff, 2000). Particular attention has
52 been given to the concept of social capital,
53 broadly understood as a social resource ‘‘upon
54 which people draw when pursuing different
55 livelihood strategies requiring coordination
56 and collective action’’ (Scoones, 1998, p. 8).
57 However, as Krishna (2000) concludes in an
58 analysis of the implications of differences in so-
59 cial capital, little is known about how to tailor
60 programs to building social capital based on
61 such differences. In this paper, we argue that
62 the role of gender differences may be of partic-
63 ular importance to understand and create social
64 capital in order to sustain NRM groups.
65 Although the gender dimensions of NRM have
66 been identified as key factors shaping peoples
67 access to and use of natural resources (Agra-
68 wal, 2000; Cleaver, 1998a; Poats, 2000), most
69 discussion of social capital so far appears to
70 have been almost gender blind (Molyneux,
71 2002) or even critical toward women’s role in
72 the formation and maintenance of social capital
73 (Riddell, Wilson, & Baron, 2001). Conse-
74 quently, analysis of gender biases of social cap-
75 ital, understood as collective action that
76 (re)produces gender discrimination, that is,
77 reinforces male dominated power structures
78 and excludes women from participation and
79 decision making, is also almost nonexistent.
80 Thus, the hypothesis that gender influences
81 NRM through different, gender-related stocks
82 and usages of social capital requires further
83 examination and empirical testing.
84 Classifying social capital as ‘‘institutional,’’
85 based on transactions governed by roles, rules,
86 procedures, and organizations or as ‘‘rela-
87 tional,’’ and so governed by norms, values, atti-
88 tudes, and beliefs, suggests that different
89 strategies are needed for building social capital
90 to support collective action for NRM. Krishna
91 (2000, p. 79) indicates that in situations where
92 relational social capital is strong but institu-
93 tional capital is weak, collective action inter-
94 ventions will need to introduce rules,
95 procedures, and skills to build institutional cap-
96 ital on a relational capital base. Conversely,
97 where rules, procedures, roles, and organiza-
98tions are in place to support collective action,
99but mutual trust is low and little value is placed
100on collaboration, interventions will need to
101build trust and willingness to work together,
102and create relational social capital (Krishna,
1032000, pp. 80–88). In this paper, we conclude
104that the distinction between relational and
105institutional social capital is highly pertinent
106to understanding the implications of gender dif-
107ferences. Thus, neglect of the gender dimen-
108sions of social capital might lead to
109misleading conclusions about optimal interven-
110tion strategies. We investigate the different pro-
111cesses and outcomes in 46 men’s, mixed, and
112women’s groups in 33 rural programs in 20
113countries of Latin America, Africa, and Asia.
1142. GENDER, SOCIAL CAPITAL, AND
THE ENVIRONMENT
116Discussion on the gender aspects of develop-
117ment and environment has its origins in the the-
118ories of Women, Environment, and
119Development (WED), which highlight women
120as having a special relationship with the envi-
121ronment due to their responsibilities for the
122family and concern for the well being of future
123generations (Jackson, 1993; Manion, 2002;
124Martine & Villarreal, 1997). In this approach,
125women are seen as ‘‘a transcultural and trans-
126historical category of humanity with an inher-
127ent closeness to nature’’ (Jackson, 1998, p.
128314) and thus likely to be the principal manag-
129ers of the environment at local level (Green,
130Jokes, & Leach, 1998).
131A number of alternative perspectives have
132also emerged that are less biologically deter-
133minist about women’s roles in development
134and environmental management. These include
135gender analysis (Jackson, 1993), feminist polit-
136ical ecology (Rocheleau, 1995), feminist envi-
137ronmentalist (Agrawal, 1992), and the
138micropolitical economy of gendered resource
139use (Leach, 1991). All support the argument
140that gender differences in NRM are not due
141to women’s inherent closeness to nature but
142due to ‘‘...dynamic and complex gender identi-
143ties in which men and women experience both
144shared and divided interests’’ (Jackson, 1998,
145p. 315). According to these authors, gender dif-
146ferences in environmental relations and man-
147agement should be understood as, and
148equated with, social relations.
149Gender differences in needs and endowments
150may be key determinants of ways in which men
2 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
151 and women manage natural resources. The
152 relationship between women and nature is fre-
153 quently analyzed in terms of the increasing
154 dependency on natural resources poor rural
155 women experience due to poverty. In what
156 has been termed the feminization of poverty,
157 women have been identified as often carrying
158 the main burden of poverty due to the overrep-
159 resentation of female-headed households
160 among the poor who depend more on common
161 pool resources (Jackson, 1993; Martine & Vil-
162 larreal, 1997). Simultaneously, it has been
163 claimed that the household division of labor
164 and women’s responsibility for family provision
165 of household resources such as water and fuel
166 wood makes women both more dependent on
167 common property or open access to natural re-
168 sources and at the same time more vulnerable
169 to the negative effects on rural livelihoods of re-
170 source degradation (Manion, 2002).
171 Despite the case for viewing gender differ-
172 ences and gender relations as influential in
173 NRM decisions, gender has been largely absent
174 from efforts made to define social capital
175 (Molyneux, 2002; Riddell et al., 2001). How-
176 ever, several studies have found that men and
177 women may have different kinds and qualities
178 of social capital based on differences in their so-
179 cial networks, values of collaboration, levels of
180 conflict and capacity for conflict management.
181 With respect to social networks, a number of
182 researchers have found that women often de-
183 pend more on informal relations and so form
184 stronger kinship and friendship relations than
185 men, who tend to rely more on formal relation-
186 ships (Agrawal, 2000; Molyneux, 2002; More,
187 1990; Riddell et al., 2001). However, structural
188 variables (such as number of children, marital
189 status, age, employment status, income and
190 occupation) can be more important for explain-
191 ing differences in their social networks than
192 gender (More, 1990).
193 Molinas (1998) found that successful collec-
194 tive action is dependent on the degree of wo-
195 men’s participation. This is consistent with the
196 argument that women exhibit more cooperative
197 behavior than men due to greater interdepen-
198 dency and altruism (Folbre, 1994; Sharma,
199 1980; White, 1992). However, Jackson (1993)
200 emphasizes that the assumption of women’s
201 greater altruism is evidence of a common fail-
202 ure to scrutinize the private interest of women
203 adequately. Women cannot be seen as a uni-
204 form category but a diverse group of people
205 who vary according to class and culture as well
206 as resource endowments and decision-making
207power both between and within households.
208Molyneux (2002) also criticizes the assumption
209that women are more altruistic for not ques-
210tioning the power relations that limit women’s
211participation in formal organizations and so
212cause women more to rely on informal net-
213works. Hence, the ‘‘naturalization’’ of women’s
214cooperative behavior could be abused by tar-
215geting women for voluntary ‘‘unpaid’’ work.
216Agrawal (2000, p. 292) on the other hand,
217without rejecting possible gender differences in
218informal relationships and altruism, finds that
219the key to understand such gender differenti-
220ated social capital has to be found in the depen-
221dency of social networks and value of
222collaboration as the gender division of labor of-
223ten obliges women to work in groups. She sug-
224gests that
225‘‘...women have a greater need to build up social
226capital through localized networks, since women’s
227avenues for accumulating economic resources and
228their physical mobility is typically more restricted
229than men’s. They also have a greater need to sustain
230these networks, given their fewer exit option and les-
231ser intra-household bargaining power.’’
232Gender differences in conflictiveness and
233capacity to resolve conflicts may also reflect
234power relations that make women more vulner-
235able than men to the negative effects of conflict.
236According to several authors (Agrawal, 2000;
237Cleaver, 1998a, 1998b; Moser & Mcllwaine,
2381999), women are often more affected by con-
239flict because they are more dependent on infor-
240mal networks of collaboration. But Agrawal
241(2000) suggests that such interdependence helps
242to overcome social division and to facilitate
243conflict resolution.
244In summary, gender relations have been iden-
245tified as important determinants of the capacity
246for collective action for NRM. Gender differ-
247ences in several aspects of social capital have
248also been identified or hypothesized, but these
249two strands of analysis in the literature have
250not been well integrated. Several important
251and unanswered questions have practical impli-
252cations for policy and program design. To what
253extent do women and men demonstrate differ-
254ent NRM outcomes based on collective action?
255Do women tend to build and use social capital
256more readily than men, and if so, is this associ-
257ated with gender differences in NRM? More-
258over, gender-differentiated social capital may
259not be inherently beneficial to NRM if social
260capital upholds or increases exclusion and dis-
261crimination. Thus, if gender-differentiated so-
262cial capital exists, is this due to innate gender-
THE IMPORTANCE OF GENDER DIFFERENCES 3
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
263 related attributes, the poverty of most rural wo-
264 men, or the underlying differences between men
265 and women in power, influence over decision
266 making, and control over assets? Thus, do
267 NRM interventions relying on collective action
268 for success need to include gender-differentiated
269 strategies for building and using social capital?
270 In this paper, we analyze the different and
271 complementary roles of women and men in so-
272 cial capital formation and its use, and explore
273 the potential consequences of gender differences
274 for NRM. We bring empirical evidence to bear
275 on some aspects of the questions posed above.
276 The analysis focuses on three broad proposi-
277 tions about the characteristics of gender differ-
278 ences in social capital:
279 (i) Women and men commonly depend on
280 different kinds of social relations or net-
281 works (Agrawal, 2000; More, 1990; Neuh-
282 ouser, 1995). Women are often more
283 dependent on informal networks based on
284 everyday forms of collaboration such as col-
285 lecting water, fetching fuel wood, and child
286 rearing. Such informal networks provide
287 solidarity and access to household resources
288 like water and firewood. Men are often
289 engaged in more formal networks, such as
290 project groups and community councils that
291 improve access to economic resources and
292 decision making (or power) (Agrawal, 2000).
293 (ii) Women and men may value collabora-
294 tion differently. Women often have more
295 everyday experiences of informal collabora-
296 tion based on reciprocal relationships and
297 higher dependence on social relations for
298 access to household resources (Agrawal,
299 2000; Cleaver, 1998b). At the same time, it
300 is often assumed that women reveal more
301 relational and altruistic behavior due to
302 their role and responsibility for reproduction
303 (Folbre, 1994; Sharma, 1980; White, 1992),
304 and are less motivated by selfish individual-
305 ism (Molyneux, 2002), while men are more
306 individualistic and more engaged in formal
307 collaboration, decision making and orga-
308 nized power structures.
309 (iii) Women are better able to overcome
310 social division and conflicts (Agrawal,
311 2000; Cleaver, 1998b; Moser & Mcllwaine,
312 1999), because of their greater interdepen-
313 dency and their everyday experiences of col-
314 laboration. As a consequence, women are
315 expected to perform better in groups,
316 and—also as a result of their greater depen-
317 dency on natural resources due to the house-
318hold division of labor—to achieve better
319outcomes from collective NRM (Agrawal,
3202000). 321
322
3. FRAMEWORK AND METHODOLOGY
323Examination of the complex causal relation-
324ships between gender and collective NRM
325through different gender-related stocks and
326usage of social capital requires an innovative
327three-dimensional framework that combines
328elements of gender analysis, collective NRM,
329and social capital based on previous frame-
330works developed for environmental collective
331action (Agrawal, 2000; Krishna, 2000; Pretty
332& Frank, 2000; Pretty & Ward, 2001): The
333three dimensions employed for our analysis
334comprise
335(1) The effects of gender on social capital
336based on the three propositions on gendered
337social capital described above that refer to
338social relations in networks, collaboration,
339and conflict management.
340(2) The impact of gender on the effective-
341ness of collective action measured in terms
342of the maturity of groups (Pretty & Ward,
3432001).
344(3) The effects of gender on the results of
345NRM measured in terms of a group’s learn-
346ing approach to NRM (Pretty & Frank,
3472000)348
349
Combining these three dimensions of the
350relationship between gender, collective action,
351and NRM, our analysis assesses five features
352of collective action in NRM groups: (i) collab-
353oration; (ii) social relations in networks; (iii)
354conflict management; (iv) group maturity; (v)
355impact on NRM. Each of these five variables
356is now defined in detail below.
357(i) Collaboration—defined here in terms of
358five dimensions: frequency, value, purpose,
359type, and structure of collaboration. To
360examine the frequency by which group
361members collaborate, we examined (1) how
362often the groups get together for meetings
363or specific activities and (2) how often the
364group’s members work together outside the
365group (by collaboration ‘‘among group
366members outside the group’’ we refer to sit-
367uations where two or more members of the
368group get together and collaborate on infor-
369mal basis on activities that are not necessar-
370ily related to the specific objectives of the
371group). To measure the value placed on col-
372laboration, we analyzed the dichotomy
4 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
373 between altruism and working for the com-
374 mon good vs. selfishness and participation
375 for personal benefit. This was measured
376 through respondents’ assessment of whether
377 group members participated for the purpose
378 of individual gains (such as resources and
379 higher personal status) and/or whether their
380 main motive of participation was a desire to
381 contribute to group or community benefit.
382 Because such judgment is naturally prone
383 to bias, we have sought to revise the analysis
384 through triangulation of similar issues.
385 These include level of solidarity among
386 group members in situations of emergency
387 or need, as well as group members’ specific
388 incentives or purposes of collaboration
389 including access to monetary resources or
390 credit, access to agricultural inputs or land,
391 access to decision making, access to collabo-
392 ration and mutual help and opportunities to
393 socialize (i.e., psychological benefit of
394 belonging to a group). Based on the three
395 propositions about gender differences in
396 social capital discussed above, we would
397 expect to observe higher values related to
398 collaboration, and higher frequency of col-
399 laboration in groups where women are pres-
400 ent, and the highest levels of collaboration in
401 women-only groups.
402 (ii) Social relations or networks—defined
403 here as a set of people (or organizations or
404 other social entities) connected by a set of
405 social relationships (such as kinship, friend-
406 ship, labor groups) that enable the flow of
407 resources and information through them
408 (Garton, Haythornthwwaite, & Wellman,
409 1997). We focus specifically on bonding con-
410 nections (between individuals in the group)
411 and less on bridging connections (horizontal
412 between the group and other local groups)
413 and linking connections (vertical connec-
414 tions between the group and external organi-
415 zations). To examine bonding connections,
416 we analyzed in-group relationships among
417 individuals dividing them into relational
418 (family, friends, and neighbors), functional
419 (cooperatives, community councils, and
420 external projects), symbolic (ethnic, reli-
421 gious, and political), and place-based rela-
422 tionships (historical, and cultural
423 attachment).
424 (iii) Conflict management—defined here
425 according to the frequency of differences in
426 groups that lead to conflict, and capacity
427 to resolve disagreements. If gender differ-
428 ences influence capacity to manage conflict
429we should, according to the propositions
430on gender and social capital and women’s
431capacity to overcome social division and
432conflict, expect to find fewer incidences of
433conflict and a higher capacity (from nonex-
434isting to very high) to resolve the existing
435ones among women’s groups as compared
436to men’s and mixed groups. This again
437should coincide with dependency on and
438values of collaboration previously measured
439in terms of frequency of collaboration as
440well as values of altruism and solidarity.
441(iv) Group maturity, here defined as NRM
442groups’ ‘‘potential for self-defining and
443self-sustaining activity’’ (Pretty & Ward,
4442001, p. 209), has been operationalized in
445previous research into a series of criteria
446which can be found at three levels of devel-
447opment termed reactive dependence, realiza-
448tion independence, and awareness
449interdependence. We measure these stages
450of maturity on the basis of seven criteria:
451(1) group objectives in relation to NRM
452which reflect whether the group is reactive,
453regenerative or innovative; (2) the group’s
454views on change (whether avoiding change,
455adjusting to change, or creating new oppor-
456tunities); (3) whether the group monitors
457and evaluates its own progress; (4) the
458degree of reliance on external facilitators to
459solve problems; (5) collective or individual
460planning and testing; (6) the importance of
461external aid for the formation of the group;
462and (7) resilience or likelihood of the group
463breaking up. Effectiveness or the potential
464for self-defining and self-sustaining activity
465is operationalized principally in terms of
466increasingly supportive values and attitudes
467toward self-organizing collective action. If
468there are gender differences in social capital
469that strengthens internal group relations,
470then we would expect group maturity to be
471positively related to the proportion of
472women in a group.
473(v) NRM impact is defined here in terms of
474the management and learning approach to
475NRM adopted by the groups. The three cat-
476egories used are (a) reactive (focused on eco-
477efficiency by reducing cost and environmen-
478tal harm); (b) regenerative (adoption of
479regenerative technologies and some princi-
480ples of sustainability); (c) redesign (innova-
481tion according to ecological principles, no
482longer adopting new technologies to fit the
483old system, but innovating to develop
484entirely new systems of management). The
THE IMPORTANCE OF GENDER DIFFERENCES 5
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
485 three categories of NRM are indicators of
486 the evolution of the capacity of a group to
487 engage in a progressively more sophisticated
488 learning process approach to NRM (Argyris
489 & Scho
¨n, 1978), and is evinced by a progres-
490 sion along a continuum from remedial mea-
491 sures to changing current practice and
492 ultimately to fundamental innovation
493 (Pretty & Frank, 2000). NRM innovation
494 in a collective action situation requires high
495 levels of trust and networking to promote
496 knowledge sharing and confidence in reci-
497 procal support from the group in the face
498 of risk. If there are gender differences in
499 the stock and usage of social capital, and if
500 these affect innovation in NRM, then we
501 would expect to find that groups with a
502 higher proportion of women have a higher
503 probability of being at the innovation stage
504 in the continuum of NRM.
505
506 We examined the NRM outcomes that
507 groups achieved, providing respondents with
508 12 options from which they could select freely.
509 These options were designed to show whether
510 groups had adopted a reactive or regenerative
511 learning approach to NRM. In this analysis,
512 we could not include options for redesign, as
513 these should be innovative beyond current
514 knowledge. However, respondents had the pos-
515 sibility to describe such novelty in the ‘‘others’’
516 category of the questionnaire.
517 In total, we examined 46 different randomly
518 selected groups (responses received from ques-
519 tionnaires conveyed to more than 500 NRM
520 programs all over the world) related to 33 pro-
521 grams working on NRM in 20 countries of La-
522 tin America, Africa, and Asia (see Appendix
523 A). All groups had at least three years of work-
524 ing experience, and the majority were small
525 groups with less than 50 participants. The pro-
526 grams were drawn from databases on NRM
527 from the CGIAR systemwide program on
528 PRGA, the IDRC MINGA program, the
529 World Bank, and the University of Essex (Pret-
530 ty, Morison, & Hine, 2003). The groups were
531 concerned with a variety of NRM issues,
532 including agrobiodiversity, agroforestry, coast-
533 al resources, food crop production, integrated
534 pest management, irrigation, soil management,
535 and watershed and catchment management be-
536 sides a number of programs that work on a
537 variety of multipurpose activities with the
538 objective to alleviate poverty through sustain-
539 able NRM. The groups represent all of the ma-
540 jor categories of NRM groups identified by
541 Pretty and Ward (2001) from eight countries
542in Asia, eight in South America, and four in
543Africa.
544The membership of six program’s groups was
545solely men, eight were solely women, and 32
546were mixed. We recognize the difficulties work-
547ing with a category like ‘‘mixed groups,’’ in
548which women’s and men’s respective degree of
549participation may vary considerably and in
550which the exact gender composition of the
551groups and position of the group members is
552not explicitly measured by the questionnaire.
553Thus the ‘‘mixed group’’ should be seen as a
554distinct category of group formation, creating
555different dynamics and providing unique
556opportunities for participation that are differ-
557ent from purely men’s and women’s groups.
558When compared with men-only and women-
559only groups, these mixed groups represent a
560phenomenon that allows us to investigate rela-
561tionships between the presence of women (or
562men) in groups and overall group behavior/per-
563formance.
564In all, the groups in the sample contained
565some 1,015 families, representing an average
566of 22 members per group. Despite the con-
567straints of sample size, we were able to identify
568significant differences between some categories
569(maturity of groups, NRM achievements/ap-
570proach, and homogeneity of groups) and trends
571in others (frequency of collaboration, solidar-
572ity, and capacity to manage conflicts).
573These groups were surveyed using a question-
574naire instrument containing 31 questions di-
575vided into the five themes mentioned above.
576These included (1) collaboration (characteriza-
577tion of the organization of the group, values
578of collaboration, and experience of collabora-
579tion outside the group); (2) social capital in net-
580works (local, local–local and local–external
581connections); (3) capacity to overcome social
582division and conflict; (4) indicators of the matu-
583rity of the group; and (5) NRM achievements
584and approaches. The questionnaire was pre-
585pared for response by key informants who were
586external program facilitators (program direc-
587tors or workers with e-mail access), and the an-
588swers are based on their assessment of types
589and levels of social capital in the groups and
590not on evidence of specific actions and relations
591as experienced by group members themselves.
592Reliance on a key informant assessment rather
593than a direct assessment by the people directly
594involved, that is, in our case, the members of
595the NRM groups, is a limitation of the data
596in particular when a subjective and qualitative
597assessment is involved. To moderate potential
6 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
598 biases, we have put emphasis in the analysis on
599 more tangible issues of social capital like fre-
600 quency of collaboration and number of connec-
601 tions that can easily and objectively be reported
602 by external observers. At the same time, we
603 have avoided going into some of the more
604 intangible aspects of social capital, such as trust
605 and reciprocity that would require an insider
606 perspective.
607 Another limitation of the study is the rela-
608 tively small size and varied composition of the
609 sample, which required careful selection of the
610 statistical methods applied and triangulation
611 of different measures. To analyze the data, we
612 used three different statistical methods to deter-
613 mine differences between the groups: (i) fre-
614 quency distributions of response for women’s,
615 men’s, and mixed groups; (ii) Fisher’s exact test
616 of association between responses and type of
617 groups; and (iii) least significant different test
618 (LSD test) for women’s, men’s, and mixed
619 group’s averages in cluster of responses.
620 In order to compare the frequency of re-
621 sponse for women’s, men’s, and mixed groups,
622 we have calculated and compared how often
623 women’s, men’s, or mixed groups, have chosen
624 a given option on the average. This analysis
625 was conducted for all responses. We used Fish-
626 er’s exact test to calculate whether the differ-
627 ences in frequency of response for women’s,
628 men’s, and mixed groups are significant. In
629 our case, Fisher’s exact test shows significant
630 association when the value Pr <= Pis equal
631 or less than 0.05 (95% level of confidence) or
632 when the Pr <= Pis equal or less than 0.10
633 (90% level of confidence). To determine the dif-
634 ference among averages for cluster of re-
635 sponses, we use the least significant difference
636 test, as it is useful for the comparison of five
637 groups or less (we have three). To interpret
638 the results of the LSD test, the LSD value is
639 compared with the observed average differ-
640 ences. Means with the same t grouping letter
641 are not significantly different because their dif-
642 ferences are less than the LSD value. The
643 LSD test was used to analyzes clusters related
644 to maturity of groups, local connections, types
645 of collaboration outside the group, and NRM
646 achievements. Cluster analysis was carried out
647 to analyze the relationship between group com-
648 position (women only, men only, or mixed) and
649 group maturity. First, multiple correspondence
650 analysis was conducted of the seven variables
651 used to define group maturity to generate
652 scores for each group on dimensions that repre-
653 sent a combination of the proportion of the
654shared variance, and then cluster analysis of
655these scores was conducted using Ward’s meth-
656od (SPSS, 1994).
6574. RESULTS FROM NRM GROUPS
658(a) Similarities between men’s, mixed, and
women’s groups
660Despite the clear differences in these pro-
661grams and their geographic locations, there
662were many similarities among the groups. With
663respect to motives for collaboration (altruistic
664vs. selfish), no significant differences were de-
665tected among men’s, women’s, or mixed
666groups—a key aspect of relational social capital
667and of the argument used by some early ecofe-
668minists (Folbre, 1994; Sharma, 1980; White,
6691992). We found that in around half of the
670groups, group members came together for com-
671mon good and community purpose mainly
672(50% of the women’s groups, 66.7% of the
673men’s groups, and 46.9% of the mixed groups),
674while in approximately one-third of the groups,
675the majority of group members were collabo-
676rating because of the individual benefits such
677as resources and status they could gain from
678this. Moreover, 25% of the mixed groups and
67912.5% of the women’s groups had other reasons
680for collaboration, emphasizing participation
681for both selfish and altruistic reasons. This in-
682cludes sharing of ideas and more sustainable
683management of natural resources.
684More surprisingly, we did not find any evi-
685dence that women had stronger informal rela-
686tions as indicated by kinship, friendship, and
687neighborhood relations (20% for men’s groups,
68822.5% for women’s groups, and 24.4% for
689mixed groups), despite the well-documented re-
690search on social networks that suggests the fact
691that women have more informal and kinship re-
692lated networks than men (Agrawal, 2000;
693More, 1990; Neuhouser, 1995). In general, the
694LSD test on group member’s relationships did
695not reveal any significant differences among
696women, men, and mixed groups’ local connec-
697tions (relational, functional, symbolic, and
698place-based relations) except for the functional
699category where men had a significant higher
700score than both the women’s and mixed groups.
701However, it is worth noting that the principal
702relationship for all three groups is place based,
703which is consistent with the relatively high level
704of altruism and orientation toward the commu-
705nity by the majority of the groups. At the same
THE IMPORTANCE OF GENDER DIFFERENCES 7
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
706 time, all groups reported a very high level of
707 cohesion with little likelihood of breakdown
708 even after initial objectives had been fulfilled.
709 (b) Gender differences in collaboration
and solidarity
Although we found no gender differences in
712 the value placed on altruism, the analysis of
713 collaboration identified some gender differences
714 in collaborative behavior. Comparing fre-
715 quency of collaboration, women’s groups tend
716 to meet more often than men’s and mixed
717 groups (Figure 1). Half of the women’s groups
718 meet on an average 1–2 times a week, much
719more frequently than the others: 83.3% of the
720men’s groups and 71.9% of the mixed groups
721meet at most bimonthly. Women’s groups also
722collaborate more frequently outside the group
723(Figure 2). Members of half of the women’s
724groups collaborate on an everyday basis or 1–
7252 times a week, where only 16.7% do so in
726men’s and 31.3% in mixed groups. Members
727of half of the men’s groups collaborate outside
728only 1–2 times a year.
729One explanation for the greater frequency of
730interaction by the women’s NRM groups is re-
731vealed by analysis of group members’ principal
732activities for collaboration outside the group
733(Table 1). Women’s groups collaborate more
Figure 1. Meeting rate for men’s, mixed, and women’s groups.
Figure 2. Frequency of collaboration outside the group.
8 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
734 on everyday household activities like cooking
735 and child rearing than both men and mixed
736 groups. In contrast, men’s groups work to-
737 gether on less frequent activities like community
738 infrastructure projects as well as in community
739 boards and in externally facilitated projects.
740 There is no significant difference among the
741 group members’ collaboration outside the
742 group over agricultural activities (sowing, weed-
743 ing, and harvesting) and NRM (fetching fuel
744 wood and collecting water). The main purpose
745 of the group for collaboration is to also provide
746 some insight into why women’s groups meet
747 more frequently. Access to monetary inputs or
748 credit is the main purpose for meeting for the
749 women groups and tends to be more important
750 to women’s groups than to men’s groups only or
751 mixed groups (31.3% compared with 16.7% for
752 men’s groups and 15.3% for mixed groups)
753 compared with other motivations for collabora-
754 tion. Credit tends to require a high level of mon-
755 itoring and turn-around.
756 Finally, solidarity tends to increase in groups
757 where women are present. A majority of all
758 types of groups report that they always or usu-
759 ally help fellow members in case of emergency,
760 but mixed and women’s groups report more sol-
761 idarity compared to the men’s groups: 90.7% of
762 mixed groups and 87.5% of women’s groups say
763 they always or usually help fellow group mem-
764 bers in case of emergency or need, while 66.7%
765 of men’s groups indicate that they do this.
766 In summary, frequency of collaboration in-
767 side and outside the NRM group as well as sol-
768 idarity is higher in groups including women,
769 and this is mainly associated with collaboration
770 in gender-specific tasks, responsibilities, and
771 needs. Our findings do not suggest that collab-
772 oration among women is related to the special
773 value they place on altruism.
774 (c) Gender differences and conflict
775 Fisher’s exact test of the differences among the
776 group shows a significantly higher homogeneity
777among members of women’s groups than men’s
778and mixed groups (37.5% of the women’s groups
779claim not to have significant differences among
780group members, while the figures for the men’s
781groups and mixed groups are respectively, 0%
782and 6.3%). On the other hand, our results show
783no gender differences in the incidence of conflict
784(50% of both men’s and women’s groups have no
785experience of serious conflict), but reported that
786the capacity to manage conflict tends to be high-
787er in women’s groups. Overall, 73.9% of all
788groups report having demonstrated capacity
789(medium to very high) to overcome differences
790and conflict, but where women are present,
791groups tend to be better at managing conflict
792(50% of the women’s groups have high or very
793high capacity to manage differences and conflict,
794while only 33.3% of men’s groups and 40.6% of
795mixed groups do so) (Figure 3). However, these
796results may be biased by the fact that the term
797conflict and the severity of these could have been
798interpreted in different ways in the question-
799naire. Consequently, respondents may have gi-
800ven their answers based on different
801understandings of the term ‘‘conflict.’’
802(d) Gender differences in group maturity
803Group maturity refers to the effectiveness of
804groups to sustain collective action, measured
805according to seven criteria. Table 2 presents
806the frequencies for men only, mixed, and wo-
807men only groups on each of the seven criteria.
808Overall, men’s groups are at an early, realiza-
809tion-independence stage of group maturity.
810Women’s groups, by contrast, have a strong
811capacity for sustained collective action indi-
812cated by the stage of awareness interdepen-
813dence in group maturity. It is important to
814note that maturity in this context does not refer
815to or correlate with age or duration of the
816group: our analysis also found that women’s
817groups have fewer years of experience (3.9
818years) than both the men’s (5.3 years) and the
819mixed (5.7 years) groups. Table 2 shows that
Table 1. Types of collaboration beyond specific group activities
a
Type of external collaboration Men’s (N= 6) Mixed (N= 32) Women’s (N=8)
Natural resource management (%) 8.3 9.4 0
Household (%) 0 0 31.3
Agriculture (%) 11.1 25.0 20.8
Community infrastructure (%) 33.3 18.7 8.3
Community organization (%) 9.8 7.0 3.1
a
Least significant differences test for cluster of responses.
THE IMPORTANCE OF GENDER DIFFERENCES 9
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
820 women’s groups are more forward looking in
821 terms of NRM and appear to have less fear
822 of change. Women’s groups also conduct self-
823 analysis more regularly, which is consistent
824 with their generally more frequent rate of meet-
825 ings and collaborative actions. And finally, wo-
826 men’s groups more often organize on their own
827 behalf and with less external assistance than
828 both the men and the mixed groups.
829 The conclusion that the presence of women in
830 groups is likely to increase maturity or capacity
831 for self-defined and self-sustaining collective
832 activity on a long-term basis is supported by
833 the results of the cluster analysis shown in Fig-
834 ure 4. Multiple correspondence analysis was
835 carried out with two dimensions that are com-
836 posites of group type and the criteria in Table
837 2. Dimension 1 represents ‘‘collective capacity’’
838 as it is a composite of four variables: indepen-
839 dence from external facilitators for problem
840 solving, engagement in group planning and
841 group testing, resilience, and women’s presence
842 in group composition. Collective capacity goes
843 down reading from left to right in Figure 4.
844 Dimension 2 is predominantly characterized
845 by the variable ‘‘group formation’’: Thus,
846 groups formed because an external agency
847 asked them to do so cluster at the top as shown
848 in Figure 4; groups formed without external
849 agency cluster at the bottom. The most impor-
850 tant cluster identified is Cluster G1, which is
851 associated with groups formed without external
852 agency. Groups in this cluster have women
853 members, independence from external facilita-
854 tors; group planning as well as group testing;
855and are considered unlikely to break down. This
856cluster is characterized by the presence of wo-
857men in the groups: it includes all except one of
858the women-only groups and includes only one
859of the men-only groups. Cluster G3 differs from
860G1 in that its groups have been formed by exter-
861nal agency, and includes the one remaining wo-
862men-only group. Cluster G2 and Cluster G4
863consist of most of the men-only groups and
864are located mainly on the right-hand side of Fig-
865ure 4, showing that collective capacity is lower
866in these clusters which are characterized by reli-
867ance on outsiders to solve problems as well as
868individual planning or testing.
8(e) Gender differences and NRM
achievements
871Gender differences were identified in the type
872of NRM achievements reported by the groups.
873Actual NRM achievements of the groups were
874classified in terms of their relation to different
875learning approaches—reactive and regenera-
876tive. Women’s groups report a significantly
877higher proportion of regenerative outcomes
878than men’s groups. There is no significant dif-
879ference among groups in their reactive achieve-
880ments. This difference among women’s and
881men’s groups is consistent with their responses
882to the survey question designed to elicit which
883NRM approach the group applies. These find-
884ings support the assumption of Pretty and
885Ward (2001) that NRM learning approach is
886related to group maturity. We show that the
887more mature women’s groups apply more
Figure 3. Group capacity to manage conflict.
10 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
Table 2. Proportion of men’s, mixed, and women’s groups measured according to seven criteria in each of three stages of
maturity (stage 1: reactive dependent; stage 2: realization independent; stage 3: awareness interdependent)
Seven criteria measured Group type
Men Mixed Women
%N%N%N
Group objective
Stage 1: To conserve or restore
a natural resource or resources
to a previous status (the goal
of the group is to restore
what once was)
0 0 12.5 4 12.5 1
Stage 2: To adapt to a change
in the status of a natural resource
or resources (the goal of the group
is to adjust to new realities)
66.7 4 46.9 15 12.5 1
Stage 3: To create new opportunities
in managing a natural resource
or resources (the goal of the group
is to introduce something completely new)
33.3 2 40.6 13 75.0 6
Views of change
Stage 1: The group is fearful of
change, it is defensive
0 0 3.1 1 0 0
Stage 2: The group is adjusting to
change, it is reactive
50.0 3 46.9 15 12.5 1
Stage 3: The group is creating new
opportunities, it is proactive
50.0 3 50.0 16 87.5 7
Self-analysis
Stage 1: The group has never evaluated
its progress in meeting its objectives
16.7 1 0 0 0 0
Stage 2: The group sometimes evaluates
its progress in meeting its objectives
66.7 4 59.4 19 25.0 2
Stage 3: The group regularly evaluates
its progress in meeting its objectives
16.7 1 40.6 13 75.0 6
Problem solving
Stage 1: Usually relies on help from
outsiders to solve a problem
33.3 2 21.9 7 25.0 2
Stage 2: First tries to solve a problem
itself before seeking help from outsiders
33.3 2 71.9 23 75.0 6
Stage 3: The group does not need outside
facilitators to solve its problems
33.3 2 6.3 2 0 0
Planning and testing
Stage 1: Individual planning and testing 16.7 1 12.5 4 0 0
Stage 2: Group planning and then individual testing 66.7 4 40.6 13 62.5 5
Stage 3: Group planning and group testing 16.7 1 46.9 15 37.5 3
Group formation
Stage 1: Because an external agency asked it to 50.0 3 43.8 14 12.5 1
Stage 2: Because one or more of its members took
the initiative and there was external
agency support to help it form
33.3 2 43.8 14 50.0 4
Stage 3: Because one of more of the members
took the initiative to form the group
without external support
16.7 1 12.5 4 37.5 3
(continued next page)
THE IMPORTANCE OF GENDER DIFFERENCES 11
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
888 regenerative measures of NRM compared with
889 the less mature men’s groups.
890 5. CONCLUSIONS AND POLICY
891 IMPLICATIONS
892 The analysis of different and complementary
893 roles of women and men in social capital forma-
894 tion and the potential consequences of such dif-
895 ferences for collective NRM in this study were
896 guided by the proposition that women tend to
897 build more relational social capital than men,
898that is, informal social relations and networks
899based on norms of collaboration and conflict
900management. The reason for this, it is argued,
901is that women supposedly value collaboration,
902altruism, and conflict resolution more highly.
903Gender differences in stocks and use of relational
904social capital may translate into different NRM
905outcomes because norms of reciprocity facilitate
906collective management of natural resources by
907providing trust. Trust and reciprocity among
908actors at a personal and generalized level facili-
909tate information exchange (and thus limit
910transaction costs) and so collaboration needed
Table 2—continued
Seven criteria measured Group type
Men Mixed Women
%N%N%N
Resilience
Stage 1: It is possible that group
breaks down before its goals are achieved
16.7 1 6.3 2 0 0
Stage 2: It is possible that the group
breaks down after achievements of initial goals
0 0 21.9 7 28.6 2
Stage 3: It is unlikely that the group
breaks down. The purpose of the group is
redefined when initial goals are achieved
83.3 5 71.9 23 71.4 5
Dimension 1: Collective Capacity
Dimension 2: External agency in group formation
Cluste
r
G1
G2
G3
G4
Hi
Hi
Lo
Lo
Figure 4. Cluster analysis of group maturity.
12 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
911 for collective action is enhanced. Collaboration
912 which values solidarity and generosity may al-
913 low access to resources and discourage stake-
914 holders from applying certain management
915 practices that would affect others negatively.
916 Women’s relational social capital and values
917 which support solidarity with other women
918 may enable them to organize more effective col-
919 lective action than men.
920 Our analysis reveals mixed results in relation
921 to these propositions. On the one hand, propo-
922 sitions about the tendency of women to have
923 altruistic values and informal relationships that
924 endow them with higher social capital than men
925 are not supported by this study. In groups
926 formed for collective action in NRM, we did
927 not find significant gender differences in rela-
928 tional social capital in terms of the reported va-
929 lue placed on altruism and the extent of
930 informal kin relationships. On the other hand,
931 we did find gender differences in the frequency
932 of collaboration, solidarity, and capacity to
933 manage conflict where the data reveal an effect
934 on group behavior of the presence of women in
935 groups. We found that collaboration, solidarity
936 and conflict resolution all increase with wo-
937 men’s presence in the groups, which is congru-
938 ent with Molinas’ (1998) and Odame’s (2002)
939 finding that women’s participation increased
940 cooperation. Our findings suggest that norms
941 of reciprocity are more likely to operate in
942 groups where women are present and that this
943 may be the result of women’s work responsibil-
944 ities that rely on frequent collaboration.
945 Similarly, the capacity for self-sustaining col-
946 lective action increased with women’s presence
947 and was significantly higher in the women’s
948 groups. The analysis demonstrated a significant
949 relationship between (a) maturity of groups and
950 gender and (b) NRM approaches/achievements
951 and gender characteristics of groups. This re-
952 sult supports the finding from another study
953 that maturity of groups is positively related to
954 performance and management of natural re-
955 sources (Pretty, 2003; Pretty & Ward, 2001).
956 At the same time, we would be cautious about
957 concluding that a more regenerative approach
958 to NRM found in women’s groups compared
959 with men’s more reactive approach is a direct
960 result of higher awareness of ecological princi-
961 ples. This situation more likely reflects women’s
962 potentially higher dependency on common
963 property resources and their limited access to
964 external inputs.
965 The analysis provides clear evidence of the vi-
966 tal role of gender analysis for collective NRM
967and points to the importance of diagnosing
968gender differences in social capital in a commu-
969nity or a group before intervention in order to
970match the existing level of social capital with
971the need to organize for specific collective activ-
972ities. It does not make sense to assume that wo-
973men will automatically possess a higher stock
974of social capital than men. It will be important
975to examine how different gender-related needs,
976responsibilities, and endowments, and in partic-
977ular the gender division of labor, affect commit-
978ment to norms of reciprocity and collaboration.
979However, we recommend careful analysis of
980the potentials for the spillover effect of gen-
981der-differentiated social capital in order to iden-
982tify ways for taking advantage of the existing
983levels of social capital to strengthen the organi-
984zation of collective NRM. Such an analysis
985must pay ample attention to the private interest
986of women and should consider whether and
987how collective action represents a resource
988from which different types of classes of women
989will benefit in different ways. Specifically, it will
990be important not to exploit women’s potential
991for collective action to implement NRM pro-
992jects that are not in their direct interest and to
993avoid reinforcing the ‘‘dark side’’ of gender-
994specific social capital, which may be exclusion-
995ary and discriminatory. Women may depend
996more on some forms of relational social capital
997simply because they are excluded from male-
998dominated formal networks and organized
999power structures where institutional social cap-
1000ital is built and exercised.
1001Moreover, informal networks are needed to
1002cope with multiple responsibilities for household
1003provisioning, reproduction, childcare, and risk
1004management. Women’s capacity for organizing
1005effective collective action may not be related to,
1006or depend on gender differences in the values,
1007attitudes, and informal relations that constitute
1008relational social capital but on their opportunity
1009for participation and even the sheer pressure of
1010their workload. Based on the finding that mixed
1011groups are an important type of organization
1012where women’s presence has an effect on group
1013performance, we would rather suggest that
1014attention should be exercised in forming and
1015supporting mixed groups to ensure that women
1016are given both a clear voice and decision-making
1017power. In mixed groups, women and men are
1018likely to have different needs, capabilities, and
1019preferences, and to the extent that these differ-
1020ences are respected the presence of women in
1021mixed groups is likely to raise the level of matu-
1022rity and solidarity in the groups and so improve
THE IMPORTANCE OF GENDER DIFFERENCES 13
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
1023 NRM outcomes. This would imply that an
1024 important focus of gender-sensitive capacity
1025 building and interventions to promote collective
1026 action would be to ensure that there is appropri-
1027 ate opportunity for women to participate.
1028 Consequently, we recommend that interven-
1029 tions to promote collective action for NRM di-
1030 rectly address the gender composition of group
1031 organization, and in particular the groups’ rela-
1032 tional and institutional social capital, and any
1033 norms, rules, or networks that exclude women
1034 from participation and decision making. Such
1035 a recommendation necessarily implies readiness
1036 to challenge the structural positions from which
1037 women participate. To do so, it is critical to
1038 diagnose the power relations among men and
1039 women and comprehend their patterns of inter-
1040 dependence to be able to influence and facilitate
1041 gender relations and dynamics in collective ac-
1042 tion groups. Likewise, it is essential to assess
1043 the meaning of participation to women and
1044 men and understand better the dynamics and
1045 processes of how they draw on collective action
1046 resources in gender-differentiated groups.
1047 Further research could usefully examine
1048 these issues to flesh out the dynamics underly-
1049 ing our finding that the presence of women in
1050 NRM groups tends to increase their effective-
1051 ness. Based on our results, we suggest that
1052 understanding gender relations is important
1053for the sustainability of groups and how they
1054may improve NRM. Krishna’s question about
1055how to tailor programs to build collective ac-
1056tion based on recognition of gender differences
1057in social capital remains unanswered by our
1058analysis. However, one implication of our find-
1059ings is that in cases where women are high on
1060relational social capital as our analysis re-
1061vealed, but weak on institutional social capital,
1062and where men have strong institutional social
1063capital but are short of relational social capital,
1064their capacity to organize effective group pro-
1065cesses for collective action in NRM will vary.
1066Gender differences in social capital imply that
1067some form of intervention is required to con-
1068struct institutional social capital in the form
1069of enforceable rules, procedures, and sanctions
1070that can be used by women, or alternatively
1071that relational capital is built in the form of en-
1072hanced to trust, norms of collaboration, and
1073conflict management for men. Thus, we con-
1074clude that capacity building and interventions
1075to promote collective action for NRM need to
1076be gender differentiated.
6. UNCITED REFERENCE
1078Uphoff (1998).
1079 REFERENCES
1080 Agrawal, A., & Gibson, C. C. (1999). Enchantment and
1081 disenchantment: The role of community in natural
1082 resource conservation. World Development, 27(4),
1083 629–649.
1084 Agrawal, B. (1992). The gender and environmental
1085 debate: Lessons from India. Feminist Studies, 18(1),
1086 119–158.
1087 Agrawal, B. (2000). Conceptualising environmental
1088 collective action: Why gender matters. Cambridge
1089 Journal of Economics, (24), 283–310.
1090 Argyris, C., & Scho
¨n, D. (1978). Organisational learning.
1091 Reading, Mass: Addison-Wesley.
1092 Baland, J.-M., & Platteau, J.-P. (1996). Halting degra-
1093 dation of natural resources: Is there a role for rural
1094 communities. Oxford: Clarendon Press and FAO.
1095 Bromley, D. W. (Ed.). (1992). Making the commons
1096 work: Theory, practice and policy. San Francisco: ICS
1097 Press.
1098 Cleaver, F. (1998a). Incentives and informal institutions:
1099 Gender and the management of water. Agriculture
1100 and Human Values, (15), 347–360.
1101 Cleaver, F. (1998b). Choice, complexity, and change:
1102 Gendered livelihoods and the management of water.
1103 Agriculture and Human Values, (15), 293–299.
1104Folbre, N. (1994). Who pays for the kids: Gender and the
1105structures of constraint. London: Routledge. 1106Garton, L., Haythornthwwaite, C., & Wellman, B. 1107(1997). Studying online social networks. Journal of
1108Computer-Mediated Communications, 3(1), Available 1109from www.ascusc.org/jcmc/vol3/issue1/garton.html.1110Green, C., Jokes, S., & Leach, M. (1998). Questionable 1111links: approaches to gender in environmental re- 1112search and policy. In C. Jackson, & R. Pearson 1113(Eds.), Feminist visions of development. London and 1114New York: Routledge. 1115Harris, J., & Renzo, P. (1997). Policy arena—Ômissing 1116link’ or analytically missing? The concept of social 1117capital—an introductory bibliographic essay. Journal
1118of International Development, 9(7), 919–937. 1119Jackson, C. (1993). Doing what comes naturally? 1120Women and environment in development. World
1121Development, 21(12), 1947–1963. 1122Jackson, C. (1998). Gender, irrigation and environment: 1123arguing for agency. Agriculture and Human Values,1124(15), 313–324. 1125Korten, D. C. (1986). Introduction: Community-based 1126resource management. In D. C. Korten (Ed.), Com-
1127munity management: Asian experience and perspec-
14 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
1128 tives (pp. 1–15). West Hartford, CT: Kumarian
1129 Press.
1130 Krishna, A. (2000). Creating and harnessing social
1131 capital. In P. Dasgupta, & I. Serageldin (Eds.),
1132 Social capital a multifaceted perspective. Washington,
1133 DC: The World Bank.
1134 Krishna, A., & Uphoff, N. (1998). Mapping and
1135 measuring social capital: A conceptual and empirical
1136 study of collective action for conserving and developing
1137 watersheds in Rajasthan, India. Ithaca: Cornell Uni-
1138 versity.
1139 Leach, M. (1991). Engendered environments: Under-
1140 standing natural resource management in the West
1141 African forest zone. IDS Bulletin, 22(4), 17–24.
1142 Leach, M., Mearns, R., & Scoones, I. (1999). Environ-
1143 mental entitlements: Dynamics and institutions in
1144 community-based natural resource management.
1145 World Development, 27(2), 225–247.
1146 Lyon, F. (2000). Trust, networks and norms: The
1147 creation of social capital in agricultural economics
1148 in Ghana. World Development, 28(4), 663–681.
1149 Manion, H. K. (2002). Ecofeminism within Gender and
1150 Development. Ecofem.org—The eJournal. Available
1151 from http://www.ecofem.org/journal/.
1152 Martine, G., & Villarreal, M. (1997). Gender and
1153 sustainability: Re-assessing linkages and issues, sus-
1154 tainable development department (SD)—Food and
1155 Agricultural Organization of the United Nation
1156 (FAO). Available from http://www.fao.org/sd/wpdi-
1157 rect/wpan0018.htm.
1158 Molinas, J. R. (1998). The impact of inequality, gender,
1159 external assistance and social capital on local level
1160 cooperation. World Development, 26(3), 413–431.
1161 Molyneux, M. (2002). Gender and the silence of social
1162 capital: Lessons from Latin America. Development
1163 and Change, 33(2), 167–188.
1164 More, G. (1990). Structural determinants of men’s and
1165 women’s networks. American Sociological Review,
1166 55, 726–735.
1167 Moser, C., & Mcllwaine, C. (1999). Gender and
1168 rebuilding social capital in the context of political
1169 violence: A case study of Colombia and Guatemala.
1170 Paper presented at the conference, Armed Conflict
1171 and Political Violence, June 10–11, 1999, The World
1172 Bank.
1173 Neuhouser, K. (1995). Worse than men: Gendered
1174 mobilization in an urban Brazilian Squatter Settle-
1175 ment, 1971–91. Gender and Society, 9(1), 38–59.
1176 Odame, H. H. (2002). Men in women’s groups: A gender
1177 and agency analysis of local institutions. In F.
1178 Cleaver (Ed.), Masculinity matter: Men, gender, and
1179 development. London: Zed Books.
1180 Olson, M. (1965). The logic of collective action. Cam-
1181 bridge: Harvard University Press.
1182 Ostrom, E. (1990). Governing the commons: The evolu-
1183 tion of institutions for collective action. New York:
1184 Cambridge University Press.
1185 Poats, S. (2000). Ge
´nero en el manejo de los recursos
1186 naturales con referencia al programa Minga del
1187 CIID, Informe final de Consultarı
´a, IDRC, Grupo
1188 Randi, Randi—FUNDAGRO, Quito, Ecuador,
1189 Marzo 2000.
1190Pretty, J. (2002). Agriculture: Reconnecting people, land
1191and nature. London: Earthscan. 1192Pretty, J. (2003). Social capital and the collective 1193management of resources. Science, 302, 1912–1915. 1194Pretty, J., & Frank, B. R. (2000). Participation and 1195social capital formation in natural resource manage- 1196ment: Achievements and lessons. Plenary paper for 1197international landcare 2000 conference, Melbourne, 1198Australia, March 2–5, 2002. 1199Pretty, J., Morison, J. I. L., & Hine, R. E. (2003). 1200Reducing food poverty by increasing agricultural 1201sustainability in developing countries. Agriculture,
1202Ecosystems and Environment, 95(1), 217–234. 1203Pretty, J., & Smith, D. (2004). Social capital in biodi- 1204versity conservation and management. Conservation
1205Biology, 18(3), 631–638. 1206Pretty, J., & Ward, H. (2001). Social capital and the 1207environment. World Development, 29(2), 209–227. 1208Putnam, R. D., Leonardi, R., & Nanetti, R. Y. (1993).
1209Making democracy work: Civic traditions in modern
1210Italy. Princeton, NJ: Princeton University Press. 1211Reddy, V. R. (2000). Sustainable watershed manage- 1212ment—institutional approach. Economic and Politi-
1213cal Weekly, 35, 3435–3444. 1214Riddell, S., Wilson, A., & Baron, S. (2001). Gender, 1215social capital and lifelong learning for people with 1216learning difficulties. International Studies in Sociology
1217of Education, 11(1). 1218Rocheleau, D. (1995). Gender and biodiversity: A 1219feminist political ecology perspective. IDS Bulletin,
122026(1), 9–16. 1221Scoones, I. (1998). Sustainable rural livelihoods: A
1222framework for analysis. Brighton, UK: Institute for 1223Development Studies. 1224Sharma, U. (1980). Women, work and property in North-
1225West India. London: Tavistock. 1226SPSS Categories 6.1. (1994). SPSS Inc. 1227Steins, N. A., & Edwards, V. M. (1999). Platforms for 1228collective action in multiple-use common pool 1229resources. Agriculture and Human Values, 16,1230241–255. 1231Uphoff, N. (1998). Community-based natural resource 1232management: Connecting micro- and macropro- 1233cesses and people with their environments. Paper 1234presented at the international workshop on commu- 1235nity-based natural resource management, Washing- 1236ton, DC, May 10–14. 1237Uphoff, N. (2000). Understanding social capital: Learn- 1238ing from the analysis and experiences of participa- 1239tion. In Dasgupta, Partha, Serageldin, & Ismail 1240(Eds.), Social capita: A multifaceted perspective.1241Washington, DC: The World Bank. 1242Wade, R. (1987). The management of common property 1243resources: Collective action as an alternative to 1244privatisation or state regulation. Cambridge Journal
1245of Economics, 11, 95–106. 1246White, S. (1992). Arguing with the crocodile: Class and
1247gender in rural Bangladesh. London: Zed Books. 1248Woolcock, M. (1998). Social capital and economic 1249development: Toward a theoretical synthesis and 1250policy framework. Theory and Society, 27, 151–208.
1251
1252
THE IMPORTANCE OF GENDER DIFFERENCES 15
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
1253
APPENDIX A. PROGRAM INVENTORY
Agrobiodiversity
Strengthening the Scientific Basis for In Situ Agricultural
Biodiversity Conservation on Farm
Vietnam
Community based Biodiversity Development and
Conservation of Indigenous Vegetables of Kenya through
Sustainable Use
Kenya
Incorporation of the Chain of Users of Potato
in the Participatory Improvement Program
Ecuador
Agroforestry
Conservation of Medicinal and Aromatic Plants
for Sustainable Livelihood
Nepal
Combining Ecological Knowledge and Socio-Economic
Perspectives in the Participatory Improvement of Multistrata
Agroforestry Systems at the Forest Margin
Indonesia
Coastal resources
Programa Ecoplata: Apoyo a la Gestio
´n Integrada de la
Zona Costera Uruguaya del Rı
´o de la Plata
Uruguay
la Marginalizacio
´n de las Comunidades Costeras Mexico
Food crop production
Whole Family Training in Maize, Bangladesh
Local Committee for Agricultural Research
(CIAL Spanish Acronym)
Colombia
Accelerating Adoption of Zero Tillage in Rice–Wheat
Systems in the Indo-Gangetic Plains
Nepal and Pakistan
Risk Management Project Zimbabwe
Revaluation of Native Potato Varieties with Emphasis on
Gender in High-Risk Climatic Zones in Ayllu Chullpas
Bolivia
Linking the Formal and Informal Systems: Exploring the
Potential for Crop Development and Biodiversity Enhancement
China
Improved Irrigation and Productivity for Organic Aromatic
Herbs Farmers in the Provinces of Sihuas y
Pomabamba, Department of Ancash
Peru
Integrated Pest Management
Motivating Farmers to Reduce Insecticide Use Philippines and Vietnam
Integrated Management of Potato Pests: Refining and
Implementing Local Strategies through Farmer Field
Schools. The Case of San Miguel
Peru
Desarrollo Agrı
´cola de la Poblacio
´n Indı
´gena de la Zona
de Influencia de Mitu
´—Monfort (Vaupe
´s): Control de
Pudriciones en Yuca Mediante Investigacio
´n Participativa
Colombia
Irrigation
Farmer Managed Irrigated Agriculture in Sindh Province Pakistan
Multiple purpose
Sustainable Improvement of Marginal Lands
in Lebanon: Aarsal, a Case Study
Lebanon
Mainstreaming Marginalized and Disadvantaged
Community through Gender and Developmental
Activities in Morang District
Nepal
(continued next page)
16 WORLD DEVELOPMENT
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS
UNCORRECTED PROOF
Appendix A—continued
Apoyo la Familias de Baja Renta de la Regio
´n
Semi-A
´rida del Estado de Sergipe
Brazil
Participatory Rural Development Project Nepal
Manejo de Recursos Naturales en la Sierra sur
del Peru
´—MARENASS
Peru
Produccio
´n Sostenible de Flores de Anturios Como
Alternativa de Diversificacio
´n, Conservacio
´n y Paz Para
la Mujer Rural en el Municipio de Caldono Cauca
Colombia
Disen
˜o e Implementacio
´n Participativos de un Prototipo
de Reconversio
´n de Fincas a la Produccio
´n Sostenible de
Hortalizas en el Municipio de Cota, Cundinamarca
Colombia
Developing Effective Institutions for Sustainable Natural
Resources Management in Deduru Oya River Basin
Sri Lanka
Soils management
Integrated Soil Productivity Initiative through
Research and Education
Uganda
Control de Erosio
´n de la Micro-cuenca Toralapa
Alta—Tercera Fase
Bolivia
Alternativas Para la Recuperacio
´n de Suelos Degradados en
Zonas de Ladera del Departamento del Valle del Cauca
Colombia
Watershed and Catchment Management
Mainstreaming of Gender Concerns in Village Panchayats India
Participatory Innovation Development in Chivi Southern Zimbabwe
Support Program for the Recovery of the San Roque
Lake Watershed
Argentina
Regional Development Plan for the Chicamocha
River Watershed
Colombia
THE IMPORTANCE OF GENDER DIFFERENCES 17
WD 1690 No. of Pages 17, DTD = 5.0.1
25 August 2005; Disk Used ARTICLE IN PRESS