Examining the relationship between community residents' economic status and the outcomes of community development programs.

In designing and implementing community development interventions the economic status of targeted participants is a demographic characteristic worth considering. The findings from this research indicate that even within the limited economies of rural Mexican villages there are variations in economic status that affect the ways in which the outcomes of community development programs are perceived. The poorest of the poor are likely to be less satisfied with development projects than those with average or better-off economic status. This is true whether a development project uses a bottom-up approach or a top-down approach. The more participatory approach does not attenuate the relationship between economic status and satisfaction with development programs. On the contrary, it may exacerbate it.

Key words: community development, economic status, rural Mexican villages

**********

According to the United Nations Human Development Report (1998) the world population has become increasingly stratified economically with people living in poverty experiencing social and geographic segregation based on class. Jameson and Wilber (1996) provide strong evidence that this stratification and distancing causes hardship.

In Mexico and other developing economies wealth accumulation and capital investment are usually concentrated in urban areas, leaving rural areas largely characterized by poverty and underdevelopment in relation to cities (Lerner, 1958; Todaro, 2000; Coordinating Committee of CASA, 2001). The growing divide separating the rural poor from the sophisticated urban professional or university-educated worker is obvious and its impact upon development goals is well documented. However, the scope and types of economic differences among rural community residents and how these differences may also impact the achievement of development goals is less understood (Lawson, McGregor, & Saltmarshe, 2000).

The purposes of this paper are to estimate the scope and types of differences in economic status that exist in seemingly homogenous rural communities and to examine the relationship between the community residents' economic status and the success of community development programs, with a specific interest in whether program type mediates this relationship. To accomplish these purposes community residents in rural Mexico participating in two types of development programs were surveyed. One of the programs sought community participation and input (bottom-up), while the other did not (top-down). Findings indicate that there are variations in economic status within the limited economies of rural communities in Mexico and that these variations are associated with different perceptions of community development programs regardless of program type.

Economic Status

A frequent problem with development projects has been their exacerbation of inequality, either through their concentration of power and benefit in local elites or their failure to address unjust power dynamics (Deere & Leon, 1982; Esman & Uphoff, 1984; Griffen, 1987; Ugandan Ministry of Finance, 2002). This problem of unbalanced distribution of benefits also occurs in rural communities even though nearly everyone seems to fit the standard textbook picture of poverty (Todaro, 2000). Anthropologists studying poor communities have shown that despite uniform appearances of economic status among families in poor communities, differences in status do exist (Lerner, 1958; Lewis, 1959; Scheper-Hughes, 1992). However, there is no consensus on the how these differences influence the course of development projects particularly on the community level.

It is not difficult to imagine ways in which subtle class distinctions at a local level could impact development project processes and outcomes. In any disadvantaged community among the poor and near-poor there are some who participate in the modern economy's formal sector and others who are alienated from it (Lerner, 1958). There are some people who are desperately poor in the absolute sense and others living above poverty thresholds who face relatively little hardship. Often within society the wealthier benefit disproportionately from government incentives and programs. This general trend is likely to be reflected in the limited economies of rural communities. As a result of these variations in economic status, community development projects probably can not enjoy equal participation from--or deliver equal benefits to--all strata (Norgaard, 1994).

Program Type

The economic status of participants in a development project may determine the form that a project takes, especially when the development experts organizing the project use a bottom-up or empowering approach. In these sorts of Freirian approaches to development the oppressed populations decide what they want to do and how they will do it (Nederveen Pieterse, 2001). Programs using a bottom-up approach may find that psychological and logistical issues play a role in determining who participates most, even when organizers try to elicit equal participation from all community members (Kahn, 1994; Norgaard, 1994).

Expert-driven and top-down development has often been criticized because outsiders do not understand (or ignore) local social realities. As a result, outside experts are thought likely to overlook the needs of the most disadvantaged groups (Esman & Uphoff, 1994; Griffen, 1987). However, it is conceivable that an outside expert who genuinely cares about the best interests of a community or a particularly oppressed social stratum might devise programs that succeed at achieving development goals congruent with community goals. Even if bottom-up programs are generally better at eliciting appropriate development plans from community members there remains a risk that a flawed selection processes within a community will unjustly determine who participates and who benefits.

Setting

The sample of research participants came from communities in rural Mexico with populations ranging from 89 to approximately 2,000. The local hard currency economies in these communities consisted primarily of three activities: selling small crops of potatoes or coffee, owning a small store that carried basic goods (e.g., canned beans, toilet paper, dried rice, candy, liquor, tobacco), or selling a service (e.g., telephone calls, truck rides from the village to the nearest bus station, or transportation of crops). Beyond these occupations a barter system was in place, and most families engaged in subsistence farming. As a result of land reform many of the community residents owned small plots of land. However, there was some tenant farming as well as communal sharing of land, both done in an attempt to boost salable crops. Monthly bills that required hard currency generally consisted of electricity, transportation, staple food items, and costs associated with subsistence farming (INEGI, 1998a; INEGI, 1998b, Larrison, 2002).

On the surface these communities seemed to match a textbook example of typical poverty communities (Todaro, 2000; Coordinating Committee of CASA, 2001). They did not exhibit obvious extremes of economic stratification. There were no indicators of gross inequality such as the villas within gated compounds surrounded by shacks of earth, cardboard, and sheet metal. On the contrary, most housing was of modest size, and construction materials seemed uniform. There were no signs of political or religious factionalism, as these communities were fairly uniformly Catholic and the districts were politically uncontested. Neither observation nor conversation with them revealed general trends of farmers, shopkeepers, or service providers doing economically better than other groups. In fact, it was common for households to mix their economic activities, doing both farming and some service work.

Programs

The remoteness and poverty of these communities attracted interventions from two university-based development programs. The two programs, Brigadas Universitarias en Servicio Social (Brigadas) and Universidad Veracruzana Proyecto (UNIR), sent students into the region to carry out service projects or assist local residents in development projects. Yet, the two programs embodied contrasting philosophies regarding who should control the development process. The presence of two development programs with similar workers from similar backgrounds working in the same region in comparable communities gave this study a remarkable opportunity to investigate development outcomes.

Examples of services provided by students included setting up clinics and teaching people about basic dental hygiene, nurses providing basic public health services, helping local farmers diversify crops in an attempt to build local self-sustaining economies, developing educational programs, and reviving local traditional arts and culture. Neither program had the resources to undertake major infrastructure projects such as building and paving roads. However, there were some important distinctions between the two programs' approaches to community development.

Brigadas. The Brigadas program was government-funded serving approximately 100 communities around the state of Veracruz, Mexico, and structured using a top-down development model. At the time of data collection the Brigadas had twelve employees that oversaw 156 students (Brigadas, 1998). The specific program interventions used by students were derived from service plans, which were based upon their professional course work and individually developed in consultation with the Brigadas staff and their professors. The plans were ultimately approved by the Brigadas director and implemented by the students when they reached their communities. Supervision was limited, with visits from Brigadas staff occurring once every two or three months, and interactions with professors limited to the times the students could find transportation from their communities to the university campus.

Although the students' plans did not include input from community residents, the Brigadas staff's past experiences with this particular set of communities insured that students' plans reflected the communities' needs. The experienced staff provided insight that most students were just starting to gain, while in a bottom-up project the students might have been expected to draw such information from the people they were helping. This lack of input from community residents differentiated the Brigadas program from the UNIR program.

UNIR. The UNIR program was an experimental program providing services similar to the Brigadas program to nine communities in the microregion of the Cofre del Perote, located in the mountain region of Veracruz. It received outside funding from the Kellogg Foundation, which required that a bottom-up model be used to provide direction to services. There were eight staff members overseeing 76 students. The program's interventions were based upon the belief that communities have the capacity for self-directed development (Nackerud & Brooks, 1996; Blanchard, 1988).

By using a bottom-up approach, UNIR attempted to identify and include community residents interested in improving the local quality of life. UNIR further identified common problems that affected the community's health and economic stability, and helped residents develop possible solutions to these problems. Based upon this information, which was gathered through community meetings and informal interviews with community residents, plans were developed and implemented with students providing technical support and knowledge. Monthly meetings among community residents, UNIR staff, and students maintained a form of supervision over the students' plans.

Purpose

Data were analyzed to answer the following questions: 1) What (if any) differences in economic status exist within these seemingly homogenous communities? 2) How does economic status influence community perceptions of the development process and program outcomes? 3) How do various economic status groups within communities affect community residents' perceptions of women's participation, leaders' responsiveness to community needs, and level of community involvement? 4) Do the patterns of satisfaction among the different economic status groups hold when program type is distinguished?

Methods

The project was a case study. The lead author and a hired data collector/translator collected data from community residents. Researchers directly administered surveys and field work lasted between October 1998 and March 1999. The generalizability of the findings presented is limited by the cultural context of one particular region in Mexico.

Sampling

A convenience sample of 701 individuals from 21 villages completed questionnaires about the impact of each program's interventions. The sample sizes of the community residents surveyed and the number of villages included were approximately equal for the Brigadas (residents N = 357, villages N = 12) and UNIR (residents N = 344, villages N = 9) programs. Chi-squares were conducted to compare the two program sub samples on a nmnber of demographic characteristics. The two sub samples were comparable in terms of age, level of hunger, and ability to pay monthly bills. The two sub samples were different in gender distributions with women making up 64.6 percent of the Brigadas sample and 54.7 of the UNIR sample [[chi square] (1, N = 700) = 7.208, p = .007].

The only requirements for individual residents' participation in the survey were a willingness to participate, knowledge of the development programs, age over 18 years, and residency in the community. These parameters necessitated a conscious decision to include people in the survey who may only have had second-hand knowledge about the programs' functioning. Beyond the aforementioned boundaries, all were welcomed to participate in the survey whether their perceptions of the programs were positive, negative, or neutral. The distributions of the demographic variables for the whole sample show a diversity of research participants, with a particularly good representation of various ages (26% - 18 to 25, 26% - 26 to 35, 22% - 36 to 45, 14% - 46 to 55, 7% - 56 to 65 and 5% - 66 plus), and gender (59.7% female).

Variables

The variable, economic status was finally measured using two economic indicators. These included individuals' perceptions of their ability to pay bills and affirmations of having three meals per day most of the time. The variables of program outcomes and satisfaction were measured by two paper-and-pencil instruments, a satisfaction scale and the Goals of Community Development Scale (GCDS), which was developed for this research project.

Economic Status of Community Residents. Initially, four statements were employed to quantify the variable economic status. Respondents used a four point Likert scale (strongly agree, agree, disagree, strongly disagree) to rate their level of agreement with the following statements: 1) Your family eats three meals a day most of the time. 2) The primary breadwinner of the family works in the village that you live in. 3) Your family makes enough money to pay the bills. 4) Your family makes enough money to pay the bills and save money (see table 1).

A factor score generated by principle components analysis of the four economic status items served as a scale of social and economic status. Factor analysis showed that the items concerning paying bills and eating three meals per day had the highest loadings (of .74 and .68 respectively for the unrotated loadings) on the one factor extracted from the four items. These two items related to each other significantly (p-value < .0001) with a Spearman rank correlation of .28 and a Kendall rank correlation of .25. Responses to these two items were therefore used to rank order the sample according to economic status as a complimentary measure to the factor score.

Answers to the economic questions about paying bills and eating three meals per day allowed us to divide the sample into rank-ordered groups according to economic status. At the bottom were the participants who strongly disagreed with both statements (n = 28), and just above them were those who disagreed with both statements without strongly disagreeing to both (n = 35). The top category strongly agreed to both statements (n = 98) and the category just under them agreed with one and strongly agreed with the other (n = 182). An additional group agreed with both statements without strongly agreeing to either (n = 51). A middle group mixed agreement and disagreement (n = 297). These six groups could be reduced to three by combing those who consistently disagreed into a poor group (n = 63), a better-off group that consistently agreed and gave at least one "strongly agree" response (n = 280), and a middle group (n = 348).

Community Development Outcomes. The GCDS was structured using 35 statements with a four point Likert response format. The scale's statements encompassed four general themes, economics, health, education, and social development. Of the 35 questions, 24 were used to create the GCDS, giving the instrument a range of possible scores between 24 and 96. The remaining eleven statements were to collect a variety of demographic information, such as gender, age, economic standing, and community of residence. The GCDS had a Cronbach's alpha of .87 and a Guttman's Lambda 2 of .89, showing a good level of reliability (Spector, 1972). It was also affirmed to have face validity among the staff and program directors of the UNIR and Brigadas programs confirming that the identified outcomes were indeed the ones sought by both programs.

Three specific items were chosen from the GCDS for chi-square tests with the economic status categories. These items were the statements: 1) More people are involved in making important decisions; 2) Local political leaders have been more responsive to your community's needs; and 3) Women have taken a more active role in community decision making.

Satisfaction. The satisfaction scale was based upon Attkison's (Department of Psychiatry at the University of California, San Francisco) Client Satisfaction Questionnaire--8 (CSQ-8). The CSQ-8 has been normed on a wide range of clients receiving social services of various forms in the United States. It is an eight-question, standardized paper and pencil measurement instrument that uses a four point Likert scale. The range of possible scores is between 8 and 32. A score of 24 or more indicates that the respondent has been mostly to highly satisfied with the services they received (Fischer & Corcoran, 1994).

The questions originally contained in the CSQ-8 were transformed to meet the needs of the research, which were oriented towards satisfaction with community outcomes rather than individual satisfaction. The questions were then translated into Spanish. The transformed satisfaction scale had a Cronbach's alpha of .77, and a Guttman's Lambda 2 of .79, indicating the translated scale maintained an acceptable level of reliability.

Translation of Measurement Instruments. The GCDS and the satisfaction scale were initially written in English and then translated verbatim to Spanish by a hired translator. The two program directors and the staff from the UNIR and Brigadas programs had an opportunity to review the initial translation. They were encouraged to provide feedback about the instruments' validity, cultural sensitivity, and accuracy in translation. The Brigadas' director and several UNIR and Brigadas staff offered suggestions. Based upon these suggestions changes were made to the GCDS and the satisfaction scale. The instruments were then pre-tested with ten community residents. Utilizing the findings from the pretest, the GCDS and satisfaction scale were retranslated by the research project's field assistant.

Limitations

The research design is limited in several important ways. First, it is a single case study, meaning that the generalizability of the results beyond the immediate case is limited. The findings should therefore be considered preliminary; however, the findings indicate that it is possible to build a better understanding of the relationship between community residents' economic status and the outcomes of community development programs.

Second, the study used a convenience sample, which means that generalizing to the population should be done with care. While collecting data in the field every house within communities was typically visited to ask for participation in the research. Furthermore, only five people approached to participate refused. However, there were instances when residents were not home.

Findings

The univariate results for the two economic status questions suggest that only about nine percent in this sample regularly experiences poverty hardships related to both hunger and an inability to pay bills. About 48 percent of the sample denies regularly experiencing either type of hardship, and 43 percent sometimes experiences some hardships. Thus, the sample is not uniformly poor. On the contrary, nearly half the sample avoids severe hardship while only a small but significant minority seems to experience absolute poverty. This low level of absolute poverty hardship reflects Mexico's position between low-income and middle-income societies, but does not match the desperation revealed in some national survey data on rural poverty (Coordinating Committee of CASA, 2001).

Dividing the sample into six rank-ordered economic status groups allowed the use of an ANOVA to compare average satisfaction and GCDS scores. This yielded a significant difference with both outcome variables. The pattern of lower satisfaction among persons in the lower status groups (F-value = 9.75, p < .0001) was mirrored by lower scores on the GCDS for the poorer groups (F-value = 8.53, p < .0001). Post-hoc tests revealed that the poorest two categories responded with significantly lower satisfaction than the other four categories with both outcome measures, while the top two categories also scored significantly higher than the "mixed response" categories.

Results suggested collapsing the six economic status categories into three (poor, middle, and better-off) would not oversimplify the measures of economic status, so this was done. These three groups were then compared on their responses to specific items in the GCDS. Asked if they perceived that women had become more active in community decision-making, 86 percent of the economically advantaged strongly agreed while only 63 percent of the poor did so. Asked whether local leaders had become more responsive, only 43 percent of the poor strongly agreed while 71 percent of the better off did so. In evaluating whether the development programs were helping more community members get involved in important decision-making again the poor were not as strongly positive as the better off, with 41 percent of the poor strongly agreeing while 70 percent of the better off community residents did so. In all these areas the poor were more likely to strongly disagree, disagree, and agree. The better off were more likely to strongly agree.

All these apparent differences were statistically significant. Using the three-level economic status variable, table 2 reports the ANOVA results for mean scores on the GCDS and satisfaction scale as well as the Chi-square results for responses to the three specific items from the GCDS. Only in the UNIR (bottom-up) group does any item demonstrate independence from economic status; no relationship was detected between economic level and one's perceptions that the UNIR program had increased community participation. Also in the UNIR group, economic status was only marginally significantly related to perceptions of women's involvement. However, for both the UNIR and Brigadas groups the mean scores for satisfaction are very clearly different according to economic status.

For the total sample the correlation between the factor score for economic status (using the four economic indicators, rather than just the two concerning paying bills and eating three meals per day) and satisfaction measured by the satisfaction scale was 0.18, while for GCDS this correlation was 0.19. These are nontrivial but weak relationships. The correlation for 342 cases responding to the UNIR programs was higher, at 0.25 for both outcome scales (p < .0001), while for the 348 cases responding to Brigadas the correlations were only 0.11 for the satisfaction scale (p < .05) and 0.15 for the GCDS (p < .01).

Inspecting best fit lines on a scatter chart in which scores on the satisfaction scale and economic status have been standardized (see Figure 1), it appears there may be a curvilinear relationship so that economic status increases with satisfaction and perceptions of development goals being met until one reaches a point of status slightly above the average, at which point the association is weaker (or vanishes). This is illustrated by the locally weighted scatterplot smoother (lowess) best fit lines for each development program's group (with tension at 66 percent of the data) shown in figure 2. The lines also show a steeper association for the UNIR group that for the Brigadas group. These UNIR/Brigadas patterns of steepness and leveling off of effect were observed with both the GCDS and satisfaction measures. The relationships are also seen when the independent variable of economic status is made a categorical variable. When using the GCDS or satisfaction scale as a dependent variable in ANOVAs and Chi-Square tests there is always a clear difference between those with the highest economic status scores and those with the lowest, but the highest (least poor) are rarely statistically significantly different from the medium group(s), and the middle is often statistically different from the poorest group. The middle economic group looks much like the highest group in perceptions of development while the poor group is clearly different from the better-off groups.

Finally, regression models controlling for age and gender found a consistent pattern of economic status relating significantly but modestly to the GCDS and satisfaction scales (see Table 3). An interesting finding of these regressions was the tendency of the younger respondents in the Brigadas (top-down) group to be more favorable about outcomes while in the UNIR (bottom-up) it was the older persons who were more favorable. The substantive strength of the models ranged from a paltry two percent of variance explained (predicting satisfaction in the Brigadas group) to a more impressive ten percent (for satisfaction in the UNIR group). In both development groups the models explained about six percent of the variance in the GCDS.

Discussion

The communities included in this study were rural, with small populations and limited local economies. The data collected show first that variations in economic status do exist even within these limited economies. A small but significant minority of community residents regularly faced hunger and an inability to pay bills, while a larger fraction of the population survived without such hardship. Second, the data suggest that these variations do affect the way community residents perceive community development programs. Residents with worse financial conditions were less impressed by the development outcomes. Residents with middle or better off economic status within the sample were more satisfied. This relationship holds true whether the development programs are organized with bottom-up or top-down models. Third, the relationship between economic status and satisfaction with a development project seems stronger in the group that participated in the bottom-up (UNIR) projects. However, on the specific issues of community participation and women's involvement, the class differences in perceptions were not significant for the UNIR participants, but were stronger and significant for residents participating in the Brigadas projects.

With the growing focus on the importance of demographic characteristics of program participants such as gender, age and culture, the findings reported in this research point to economic status as a demographic variable also worth considering when designing and implementing community development programs (Cook, 1990). Indeed, findings from the regression models suggest economic class is more important than age in how development outcomes are experienced. Furthermore, the findings suggest that whether or not purposely targeted by community development programs, people in the lowest economic strata seem to benefit less from program interventions then those individuals who already have resources of their own. Other researchers (Lawson, & et. al., 2000; Macdonald, 1995; Beck, 1989; Lecomte, 1986), as well as the United Nations Human Development Programme (1998), have cited similar problems with programs not assisting people from the poorest strata of society.

There are a number of plausible scenarios for how economic status may influence perceptions of community development programs. For example, better off community residents who grow surplus crops and run small stores may be more aware of how development projects can help them. As they are presumably more integrated into the modern economy and more aware of opportunities for improvement they may find it easier to appreciate small gains or improvements that bring indirect or long-term benefits. Complimenting this, persons in greater material hardship may be more acutely aware of project failings. Such impoverished persons may have immediate survival needs, and projects that do not help them directly satisfy those needs may be seen as fatally flawed.

Participation may also be a mediating factor. Those who are better off have more time to participate in development projects, and participation relates to satisfaction. Those in the greatest poverty may be too ill or too busy to get involved with local events, or they may feel alienated or intimidated, or they may be overlooked or avoided by others in their community and the development workers (Ugandan Ministry of Finance, 2002). Psychological variables may also play a role, as forms of hope and optimism are sensitive to life stresses with the poorest of the poor being more pessimistic or depressed, and this in turn influences how they evaluate project success, making them more critical (Pacini, Muir, & Epstein, 1998; Taylor & Armor, 1996).

A well-designed project utilizing sensitive and dedicated workers may be sufficient to elicit satisfaction, with community input being of only slight importance. This may be especially true when emphasizing the perceptions of the poorest groups. The Brigadas approach did not use inputs from the community, yet achieved satisfaction levels nearly equivalent to those achieved by the UNIR programs in which communities determined the development projects. If one compares the satisfaction of the better off half of the sample receiving Brigadas services to that of the poorer half receiving the UNIR services the satisfaction levels are nearly the same. This means that other variables, perhaps quality of the development work and the dedication and wisdom of the development workers probably exert stronger influences on satisfaction outcomes.

As expected, there were some instances that were contrary to the trend found in the data. One example was a single mother who had few economic resources, and was not recognized as a formal or informal community leader until after her interactions with UNIR. The result of her involvement was a rise in prestige of her family, which had some tangible outcomes such as her daughter receiving a scholarship to study at a university in Costa Rica. This example, however, appeared the exception rather than the rule (Larrison, 2002). It is probable that development workers remember other exceptional cases such as this woman's, but the data collected here show that such anecdotes could be misleading if they are used in generalizations about successes in helping the poorest of the poor.

Conclusions

The findings of this research show that even in poor communities, people experience variation in poverty and material hardship and that these variations in economic status play a role in how community residents perceive the outcomes of development programs. This means that community development programs need to be sensitive to how economic differences among community residents affect the achievement of community development goals (Lawson, et. al., 2000; United Nations Development Programme, 1998; Iatridis, 1994).

From a program model standpoint, bottom-up community development strategies that attempt to elicit participation from the targets of the project should be aware that emerging leaders may represent different status backgrounds, even in seemingly homogenous rural communities. If local participation is elicited, a careful development worker will encourage the very poorest to participate rather than simply allowing whoever comes forward to lead the process. Community organizers such as Khan (1994) and Homen (1999) have made this observation before, and this study offers new empirical evidence supporting the validity of this warning.

As well, the top-down model is not immune from neglecting the poorest strata in communities as demonstrated by the long history of development programs that have done little to help reduce poverty (Iatridis, 1994; Jameson & Wilbert, 1996). Programs using outside experts to devise community interventions must consider whether their projects will help the most desperately poor or the more secure persons within a disadvantaged community. This means top-down programs must have a high level of familiarity with local culture across a number of dimensions including economic status.

Table 1

Distribution of economic factors

                           Income
                           earners
                           worked in     Families
           Ate three       village of    without enough    No long
           meals per day    residence    money to pay      term savings
           (N = 691)        (N = 691)    bills (N = 690)   (N = 690)

Strongly    45 (6.5%)       67 (9.7%)    243 (35.2%)       517 (74.8%)
Disagree
Disagree    36 (5.2%)       22 (3.2%)     99 (14.3%)        69 (10.0%)
Agree      177 (25.6%)      61 (8.8%)    228 (33.0%)        80 (11.6%)
Strongly   433 (62.7%)     541 (78.3%)   120 (17.4%)        25 (3.6%)
Agree

Table 2

Relationships between economic status and satisfaction across
development approaches.

Comparison            Total Sample        UNIR

                                          Bottom-Up

ANOVA on              F = 9.9.            F = 9.4.
Satisfaction Scale    p < .0001           p < .0001
ANOVA on              F = 18.4.           F=18.2.
GCDS                  p <.001             p <.001
Community             Chi-Square =        Chi-Square = 4.6
being more            21.5                p = .59
involved.             p = .0015           Cramer's V = .08
                      Cramer's V = .13
Chi-Square on         Chi-Square =        Chi-Square =
Leaders being         24.9                16.4
more responsive       p = .0004           p =.012
                      Cramer's V = .13    Cramer's V = .15
Chi-Square on         Chi-Square =        Chi-Square
Women being           24.8                11.2
more active           p = .0004           p = .082
                      Cramer's V = .13    Cramer's V = .13

Comparison            Brigadas

                      Top-Down

ANOVA on              F = 4.6
Satisfaction Scale    p < .0005
ANOVA on              F = 5.3
GCDS                  p < .006
Community             Chi-Square =
being more            30.1
involved.             p < .0001
                      Cramer's V = .21
Chi-Square on         Chi-Square =
Leaders being         19.7
more responsive       p = .004
                      Cramer's V = .17
Chi-Square on         Chi-Square =
Women being           21.3
more active           p = .0016
                      Cramer's V = .18

Table 3

Regression Models Using Economic Status to Predict Outcome
Perceptions

                       Standardized coefficients

                                             Brigadas only
                    Total Sample                (n=331)

               GCDS      Satisfaction     GCDS      Satisfaction

SES            .18 *         .18 ***      .15 **        .11 ***
Age           -.04           .08 *       -.14 **       -.01
Gender         .04          0.07          .13 *         .11 *
Model         8.3 ***       9.9 ***      7.5 ***       3.3 *
F-value
Model Adj.     .03           .04          .06           .02
r-squared
                    Standardized coefficients

                  UNIR only
                   (n=340)

                GCDS      Satisfaction

SES            .25 ***        .27 ***
Age            .13 *          .22 ***
Gender         .01            .06
Model         8.0 ***       13.6 ***
F-value
Model Adj.     .06            .10
r-squared

* sig at p < .05. ** sig. at p <.01. *** sig at p < .001

References

Beck, T., Survival strategies and power amongst the poorest in a west Bengal village. In K.P. Jameson and C.K. Wilber (Eds.). The political economy of development and under development (6th ed.), (pp. 477-493). New York: McGraw-Hill, Inc. 1996.

Blanchard, D. Empirical strategies of bottom-up development. ICA International, IERD Regional Development Symposia. 1988. pp. 318-338.

Binnendijk, A. L. Donor agency experience with the monitoring and evaluation of development projects. In K. Finsterbusch, J. Ingersoll, and L. Llewellyn, (Eds.). Methods for social analysis in developing, countries, (pp. 165-184). Boulder, Colorado: Westview Press. 1990.

Brigadas. Programa general: Brigadas uiversitarias en servicio social. Veracruz, Mexico: Direccion de Vinculacion General Departamento de Vinculacion Social. 1998.

Buzzard, S. Surveys: Avoiding the common problems. In K. Finsterbusch, J. Ingersoll, and L. Llewellyn, (Eds.). Methods for social analysis in developing countries, (pp. 71-88). Boulder Colorado Westview Press. 1990.

Carmines, E. G. & Zeller, R.A. Reliability and validity assessment. London: Sage Publication. 1979.

Cochran, W. G. & Cox, G. M. Experimental designs (2nd ed.) New York: John Wiley & Sons, Inc. 1992.

Cook, C. C. Demography of the project population. In K. Finsterbusch, J. Ingersoll, L. and Llewellyn, (Eds.). Methods for social analysis in developing countries, (pp. 89-106). Boulder, Colorado: Westview Press. 1990.

Coordinating Committee of CASA Mexico. Adjustment and poverty: Twenty years of crisis in Mexico. Mexico City Structural Adjustment Participatory Review International Network. 2001. (Retrieved on June 4, 2003 from http://www .saprin.org/global_rpt.htm)

Deere, C. D., & Leon de Leal, M. Peasant production, proleterianization, and the sexual division of labor in the Andes. In L. Beneria, Ed., Women and development: The sexual division of labor in rural societies (pp. 65-93). New York: Praeger. 1982.

Esman, M. J., & Uphoff, N. T. Local organizations: Intermediaries in rural development. Ithaca, NY: Cornell University Press. 1984.

Fischer, J. & Corcoran, K. (Eds.) Measures for clinical practice: Volume 1 (2nd ed.). New York: The Free Press. 1994.

Green, Salkind, & Akey. Using SPSS for Windows: Analyzing and understanding data. New Jersey: Prentice Hall. 1997.

Gott, W. E., Kilpatrick, A. C. & Nackerud, L. Social development and social constructionism: Strangers or bedfellows? Social Development Issues 20(1). 1998. pp.

Griffin, K. World hunger and the world economy. New York: Holmes & Meier. 1987.

Homan, M. S. Promoting community change: Making it happen in the real world (2nd ed). Pacific Grove, CA: Brooks/Cole Publishing Company. 1999.

Iatridis, D. Social policy: Institutional context of social development and human services. Pacific Grove, CA: Brooks/Cole Publishing Company. 1994.

Instituto Nacional de Estadistica Geografia E. Informatica--INEGI. Anuario estadistico del estado de Veracruz, Vols. 1 & 2. Mexico City: Instituto Nacional de Estadistica, Geografia e Informatica. 1998a.

Instituto Nacional de Estadistica Geografia E. Informatica--INEGI. Veracruz tabulados basicos ejidales por municipio: Programa de certificacion de derechos ejidales y titulacion de solares urbanos, PROCEDE 1992--1997. Mexico City: Instituto Nacional de Estadistica, Geografia e Informatica. 1998b.

Jameson, K.P. & Wilber, C.K. The political economy of development and underdevelopment. New York: McGraw-Hill, Inc. 1996.

Kahn, Si. How people get power, Revised Edition. Washington, DC: National Association of Social Workers. 1994

Larrison, C.R. A comparison of top-down and bottom-up community development interventions in rural Mexico: Practical and theoretical implication for community development programs. Lewiston, New York: Edwin Mellen Press. 2002.

Lawson, C.W., McGregor, J.A. & Saltmarshe, D.K. Surviving and thriving: Differentiation in peri-urban community in Northern Albania. World Development 28(8), 2000. pp. 1499--1514.

Lecomte, B. J. Project aid: Limitations and alternatives. Paris: Development Centre of the Organisation for Economic Co-operation and Development (OECD). 1986.

Lerner, D. The passing of traditional society. New York: The Free Press. 1958.

Lewis, O. Five families: Mexican case studies in the culture of poverty. New York: Basic Books. 1959.

Macdonald, L. NGOs and the problematic discourse of participation: Cases from Costa Rica. In D.B. Moore and G.J. Schmitz's (Eds.). Debating development discourse: Institutional and popular perspectives, (pp. 201--229). New York: St. Martin's Press, Inc. 1995.

Nackerud, L. & Brooks, R. UNIR project: An external program evaluation. Unpublished manuscript submitted to the W.K. Kellogg Foundation and Universidad Veracruzana. 1996.

Norgaard, R. B. Development betrayed: The end of progress and a coevolutionary revisioning of the future. New York: Routledge. 1994.

Nederveen Pieterse, J. Development theory: Deconstructions/reconstructions. Thousand Oaks, CA: SAGE Publications. 2001.

Ott, L. An Introduction to statistical methods and data analysis (4th ed.). Duxbury Press. 1993.

Pacini, R., Muir, E & Epstein, S. Depressive realism from the perspective of cognitive-experiential self-theory. Journal of Personality and Social Psychology, 74, 1998. pp. 1056-1068.

Rubin, A. & Babbie, E. Research methods for social work (2nd ed.). Pacific Grove, CA: Brooks/Cole Publishing Company. 1993.

Scheper-Hughes, N. Death without weeping: The violence of everyday life in Brazil. Berkeley, CA: University of California Press. 1992.

Spector, P. E. Summated rating scale construction: An introduction. London: Sage Publishing. 1992.

Taylor, S. E., & Armor, D. A. Positive illusions and coping with adversity. Journal of Personality, 64 (4): 1996. pp. 873-898.

Todaro, M. P. Economic development (7th ed.). Reading, MA: Addison Wesley. 2000.

Ugandan Ministry of Finance Uganda Participatory Poverty Assessment Process, Second participatory poverty assessment report: Deepening the understanding of poverty. Kampala, Uganda: Ministry of Finance, Planning and Economic Development 2002. (retrieved at http://www.uppap.or.ug/nationalPPA2 .htm on June 4, 2003)

United Nations Development Programme Human development report. New York: Oxford University Press. 1998.

CHRISTOPHER R. LARRISON

ERIC HADLEY-IVES

University of Illinois at Urbana-Champaign

School of Social Work