In a stratified sampling design, the sample size is determined separately for each stratum, and the total sample size is the sum of the sample sizes across all strata. The calculation of the sample size for each stratum was performed using the formula proposed by [6] :

N = [Nz²p (1 − p)]/[d² (N − 1) + z²p (1 − p)].

where:

N: the population size within each stratum

Z: the confidence interval

P: the proportion of the target population within the overall population

D: the required degree of precision

Assuming a 95% confidence level, a precision of 5%, and a prevalence of mango producers of 20% within the agricultural population, and incorporating a finite population correction factor (Naing et al., 2006), the sample size was determined for each of the three strata. This enabled calculation of the total sample size, which was n = 443 for the project area (Korhogo 137 producers; Sinématiali 208 producers, and Ferké 98 producers) and n = 75 for the control area of Boundiali.

Data were collected through a formal survey utilizing a structured questionnaire, primarily centered on the producer module. Three enumerators conducted the data collection, with one assigned to each stratum within the project area. In the control area, two of the three enumerators were responsible for administering the survey.

Data analysis was performed using STATA software, version 13, with careful consideration of the survey’s structure, particularly the sampling weights. Estimates of means, totals, and proportions were generated using the SVY prefix in STATA to specify that the data originated from a survey. Each point estimate was accompanied by its standard error and corresponding confidence interval.

Assessing the impact of an intervention involves addressing the following question: would the observed situation have been different had the intervention not taken place? In essence, this requires comparing outcomes with and without the intervention. However, the fundamental challenge lies in the impossibility of directly observing the counterfactual scenario—what would have occurred in the absence of the intervention. Consequently, impact evaluation techniques are designed to estimate as accurately as possible the situation that would have prevailed without the intervention, while the situation with the intervention is empirically measured during the evaluation process.

A range of methodological approaches is available to analysts. The most basic is the before-after estimator, also known as the simple difference method. This approach calculates impact as the difference between measurements taken before and after the intervention. However, this method is unreliable because, without a control group, it is impossible to ascertain whether observed changes are attributable to the intervention itself or to other unrelated factors. To address this limitation, evaluators often employ the double difference method (difference-in-differences). This approach involves identifying a control area comparable to the intervention area. The impact is then estimated as the difference between the changes observed in the intervention group and those in the control group over time. This method is based on the theoretical assumption that, in the absence of the intervention, the beneficiary group would have followed the same trajectory as the control group—an assumption that may not always hold perfectly in practice. Nonetheless, the potential bias associated with the double difference method is generally less severe than that introduced by the simple difference approach. For this reason, the double difference method is widely preferred among analysts and is the recommended approach in the present study.

The findings of this study reveal that mango production is overwhelmingly dominated by men, with women constituting less than one percent of the sample ( Table 1 ). The average age of producers is under 50 years, indicating that the cohort is relatively young and within their most productive years ( Table 2 ). Notably, female producers, with a mean age of 43, are significantly younger than their male counterparts ( Table 3 ).

Marriage is the prevailing marital status among the producers, with over 97% reported as married, while singlehood and other forms of cohabitation account for less than 2% of the sample. These characteristics, particularly the extremely low percentage of cohabitants, suggest that Senoufo society is especially attentive to conventional forms of cohabitation ( Table 4 ).

Analysis of family structure among producers shows that just over half of the families are polygamous, while approximately 40% are monogamous, making monogamy the second most prevalent form of partnership. Single individuals of both genders comprise less than 3% of the sample. The marriages referenced are generally traditional, often accommodating polygamy. The near equivalence in the proportions of polygamous and monogamous couples points to a society in transition, gradually shifting from polygamous traditions toward monogamous modernity ( Table 5 ).

The analysis of producers’ social backgrounds reveals that over 99% are indigenous Senoufo, whereas migrants from other regions of Côte d’Ivoire constitute less than 1%. Notably, there are no foreign nationals among the producers, despite the geographical proximity and considerable presence of individuals from neighboring countries such as Burkina Faso and Mali, where mango cultivation is nonetheless relatively advanced ( Table 6 ).

In terms of producers’ origins, the data indicate that indigenous Senoufo producers are significantly older than their non-native counterparts. The age gap between these two social groups of producers is substantial, approaching a decade ( Table 7 ).

The average household size among mango producers is 13 individuals ( Table 9 ), and the household head typically possesses approximately twenty years of experience in mango cultivation ( Table 10 ).

Table 9. Average household size among mango producers.

Labor input is considered from a qualitative perspective. Due to constraints in time and data availability, the total quantity of labor devoted to a mango orchard is not estimated; instead, the focus is on analyzing the types of labor utilized by producers. The findings indicate that over 80% of mango producers rely on family labor. Approximately half of the producers employ hired labor, while a similar proportion do not engage in this form of labor ( Table 11 and Table 12 ). Furthermore, nearly three-quarters of producers predominantly depend on family labor rather than wage labor ( Table 13 ).

Regarding labor modalities, almost two-thirds of producers possess animal-drawn farming equipment, which serves as a reliable indicator of their involvement in cotton cultivation ( Table 14 ).

The average size of a mango orchard is approximately 6 hectares, with the total area under mango cultivation in the project region estimated at 14,573.51 hectares (see Table 15 and Table 16 ). Analysis of orchard size classes reveals that over two-thirds of mango orchards occupy no more than 5 hectares, while 20% are in the 5 to 10 hectare range; only about 13% of orchards cover 10 hectares or more (see Table 17 ).

The findings reveal that over 90% of orchards consist of no more than two plots, with these plots predominantly located within the same physical area (see Tables 18-20 ). Furthermore, over 90% of these plots are homogeneous in terms of varietal composition. Specifically, the exploratory survey indicates that the Kent variety is overwhelmingly predominant, while the Keith and Amélie varieties are present only in minor proportions.

The average age of mango orchards in the project area is approximately 22 years. Half of the existing orchards are between 20 and 40 years old, while one third fall within the 10 to 20 year age range. The oldest (40 to 60 years) and youngest (3 to 10 years) orchard age classes each represent less than 10% of the total orchard area (see Table 21-23 ).

Two major periods of mango orchard establishment are evident: the first spans from 1975 to 1995, and the second from 1995 to 2005. Collectively, these periods account for over 80% of the current orchards in the project area. In contrast, orchard establishment at both age extremes—40 to 60 years and 3 to 10 years—has been significantly lower, each constituting less than 10% of total orchards. Specifically, the proportion of orchards established within these two extreme age categories is approximately 8% for each.

An analysis combining orchard size and age reveals that, irrespective of the size class, the majority of mango orchards were established between 1975 and 2005. Specifically, in the 1 to 5 ha size class—which accounts for 65.15% of all orchards—over 53% were created during this period. In the 5 to 10 ha category, which represents 21.42% of the total orchard area, 84.45% of orchards were established between 1975 and 2005. For the largest orchards, those exceeding 10 ha and constituting 13.43% of the total orchard area, 90.53% were established within the same timeframe.

In contrast, during the two extreme periods, the proportion of total orchard area established—regardless of size class—never exceeds 10%.

The vast majority of mango producers own the plots they cultivate, with only approximately 5% lacking ownership of the land they farm ( Table 24 ). Inheritance constitutes the principal means of acquiring land, accounting for more than 93% of mango producers’ access to property ( Table 25 ). Land rental and purchase are extremely uncommon in the region, indicating that the land market remains at a nascent stage ( Table 26 ).

Land is perceived as a scarce productive resource, with over 95% of producers reporting that land availability in the project area ranges from rare to very rare ( Table 27 ).

Despite this scarcity, most mango plots are located on agricultural land rather than in urban domains ( Table 28 ), and more than two-thirds are situated on flat terrain. Nevertheless, the findings show that approximately 20% of plots are on sloped land, while 10% are located in lowland areas ( Table 29 ).

In spite of the pronounced scarcity of land in the project area, land disputes are infrequent; about 30% of surveyed producers consider land conflicts to be rare, and more than 60% regard the prevailing situation as normal, with no disputes arising ( Table 30 ).

This section demonstrates that the use of fertilizer among mango producers is minimal, with only approximately 3% of producers applying this input ( Table 31 ). In contrast, herbicides and insecticides are commonly used, with 70% and 86% of mango producers, respectively, applying these inputs ( Tables 32-33 ).

Irrigation practices are also extremely limited among mango producers, as less than 1% utilize irrigation ( Table 34 ).

Regarding the costs associated with chemical inputs, the largest expenditures are allocated, in descending order, to herbicides, followed by insecticides, and then fertilizers ( Table 35 ).

On a per-hectare basis, the average expenditure per producer is 638.40 FCFA for fertilizers, 4614.80 FCFA for herbicides, and 2,429.04 FCFA for insecticides. Given the price of 19,000 FCFA for 50 kg of NPK fertilizer, these values suggest an average fertilizer usage of approximately 1.68 kg/ha. Overall, the average mango producer spends around 48,000 FCFA on chemical inputs for their orchard, which corresponds to 7,682.26 FCFA per hectare ( Table 36 ).

With regard to the costs associated with orchard maintenance, the interventions considered primarily include weeding, pruning of mango trees, and other maintenance activities such as the construction or upkeep of fences. The findings indicate that, on average, producers allocate approximately 50,000 FCFA to various orchard maintenance tasks, equating to an average expenditure of 7965 FCFA per hectare (see Table 37 and Table 38 ).

On average, producers sold slightly more than 16,500 tons of mangoes to exporters ( Table 39 ). Of this regional total, over 85% of the mangoes collected for export originated from orchards aged between 10 and 40 years, with 36% and 49% of the total volume coming from orchards aged 10 - 20 years and 20 - 40 years, respectively. The oldest orchards and the most recently established ones contributed 5.82% and 4.74%, respectively, to the total volume of mangoes collected for export ( Table 40 ).

The data indicate that approximately 48% of the mangoes collected by harvesting technicians were sourced from orchards of at least 10 hectares. Orchards sized between 5 and 10 hectares accounted for about 25% of the total, while those between 1 and 5 hectares contributed roughly 28%. Therefore, large orchards—those of at least ten hectares—are the primary suppliers of mangoes destined for export ( Table 41 ).

Average yield was approximately 1.15 tons per hectare, which is significantly lower than the gross yield at the producer level: 3.5 tons per hectare for the Amélie variety, 5 tons for Keith, and 4 tons for Kent, the most prevalent variety in the project area. For Kent, the collected yield represents only 28.75% of the gross yield ( Table 42 ). Results in Table 43 reveal that yields are marginally higher in the oldest orchards compared to younger ones, although this difference is not statistically significant at the 5% level.

Yield analysis by orchard size shows that the highest yields were observed in orchards of 1 - 5 hectares, with yields in the other two size classes being of similar magnitude. Nonetheless, these differences were not statistically significant at the 5% level ( Table 44 ).

The overwhelming majority of producers are neither members of producer groups nor involved in any certification processes. Specifically, nearly three-quarters (73.01%) of producers lack affiliation with any group (see Table 45 ), while approximately four-fifths of producers do not hold any form of certification (see Table 46 ).

In terms of biotic constraints, diseases pose a production challenge for approximately 85% of producers ( Table 47 ), while pests—particularly fruit flies—represent a significant threat to production for nearly 97% of producers ( Table 48 ).

Regarding socio-economic constraints, the data presented in Table 49 and Table 50 show that the scarcity of agricultural labor constitutes a moderate to severe limitation for approximately two-thirds of producers. In contrast, the low purchase price is identified as an extremely severe constraint for nearly all mango producers.

An assessment of mango producers’ knowledge regarding flies and the damage they inflict reveals that over three-quarters of producers acknowledge their inability to identify mango flies ( Table 51 ). The overwhelming majority (over 97%) believe that flies cause significant damage to the fruit; more than 66% estimate that losses attributable to flies exceed 75% of total production, while approximately 20% assess these losses as falling between 50% and 75% ( Table 52 and Table 53 ). Only around 8.5% of producers report being aware of any endogenous solution ( Table 54 ).

Training constitutes the second essential step in combating fruit flies. The data presented in Table 55 show that fewer than 15% of producers state that they have received training in fruit fly control.

This study demonstrates that mango production in northern Côte d’Ivoire is predominantly undertaken by indigenous Sénoufo men, who are mostly married, possess limited formal education, and have, on average, approximately twenty years of experience ( Tables 1-10 ). Such a demographic profile aligns with findings by [7] , who indicate that traditional agricultural systems in West Africa are typically male-dominated, especially in the case of cash crops like mango.

The markedly low participation of women in this sector (less than 1%) stands in stark contrast to trends observed in other cash crop sectors [8] , thereby underscoring a potential opportunity to enhance women’s inclusion within agricultural value chains.

Production is predominantly dependent on family labor (87%), although nearly half of producers also employ hired workers. This situation indicates a semi-commercial production structure, wherein family labor remains dominant but proves inadequate, especially during peak demand periods. Such findings are consistent with those reported by [9] regarding fruit crop production in West Africa.

The average orchard size is 6 hectares, with most producers operating on small plots (≤5 ha). However, 13% of orchards—often those owned by exporters—exceed 10 hectares and account for over 50% of total exported output ( Table 41 ). This coexistence of small- and large-scale producers is a typical feature of agricultural value chains with significant export potential [10] .

The orchards have an average age of 22 years, with most being established between 1975 and 2005. The decline in new orchard creation observed since 2005 is attributed to the political crisis, competition from cashew cultivation, infestations by Bactrocera dorsalis, and increasing land pressure ( Table 23 ). This reduction in investment activity aligns with widely reported impacts of conflict and state withdrawal on the sustainability of agricultural value chains [11] .

Land tenure emerges as a major limiting factor: while 94% of producers own their land through inheritance ( Table 24 and Table 25 ), over 95% of survey respondents perceive land as highly scarce ( Table 27 ). This paradox—widespread ownership yet pronounced scarcity—underscores the rigidity of the rural land market, as previously documented by [12] . Such scarcity directly constrains opportunities for orchard expansion or renewal.

The findings reveal a low to moderate level of intensification: fertilizer application remains marginal (3.3%), whereas herbicides (70%) and insecticides (86%) are extensively employed (see Tables 31-33 ). This reliance on plant protection products is attributable to severe pest pressure, particularly from fruit flies (Bactrocera dorsalis), which represents a major challenge in the region [13] .

Furthermore, irrigation is nearly absent (<1%), significantly constraining productivity during drought periods. Despite this, existing literature on horticulture in the Sahelian context strongly advocates for the adoption of supplemental irrigation practices to ensure yield stability [14] .

The average export yield stands at 1.15 t/ha, representing approximately 30% of the gross yield observed for cultivated varieties (4 t/ha for Kent). This marked reduction is primarily attributable to the rigorous grading required by the phytosanitary standards of European markets. These findings corroborate those of [15] , who emphasize the direct effect of fruit fly infestations on export rejection rates.

Certification is associated with a modest increase in yield (1.24 t/ha compared to 1.14 t/ha); however, this effect does not reach statistical significance, in part due to the limited uptake of certification in the region (11%) ( Table 46 and Table 50 ). These results align with the observations of [16] , who demonstrate that the effects of certification are contingent upon producers’ organizational structures and the availability of institutional support.

Producers identify biotic constraints (such as fruit flies and parasites) as the most significant, followed by abiotic constraints (including strong winds and excessive rainfall), and finally economic constraints, particularly low purchase prices and a lack of organizational structure ( Tables 52-55 ). This prioritization of constraints typifies the challenges commonly encountered in tropical export crop production across sub-Saharan Africa [17] (Affognon et al., 2015).