Mango Postharvest Technologies: An Observational Study of the Yieldwise Initiative in Kenya

: Several studies have evaluated the effects of postharvest technologies on postharvest loss (PHL) incurred at a single stage of a food value chain. However, very few studies have assessed the effect of multiple technologies on PHL incurred at various stages of a food value chain. This study evaluated the effect of ﬁve technologies (harvesting tools, cold stores, plastic crates, fruit ﬂy traps, and ground tarps) promoted by the Rockefeller Foundation Yieldwise Initiative (YWI) in Kenya on PHL incurred at three mango value chain stages (harvest, transportation, and point of sale). After extensive screening of the YWI data, the Kruskal–Wallis statistical test was used to compare each YWI promoted technology to smallholder farmers (SHF) traditional practices. Results indicated that plastic crates used to transport or store mangos and fruit ﬂy traps used to attract and kill fruit ﬂies were statistically signiﬁcant ( p < 0.05) in reducing PHL at the point of sale. Meanwhile, no statistical evidence of PHL reduction was observed from SHF using harvesting tools, cold stores, and ground tarps. Cold stores were the least adopted of the promoted technologies due to their high costs of implementation and utilization. While this study asserts that increased technology adoption is associated with PHL reduction, further research is needed to identify additional factors that favor technologies’ efﬁcacy in reducing PHL in similar food value chains. While this study asserts that increased technology adoption is necessary to obtaining better PHL reduction efﬁcacy, further research is needed to identify additional factors of importance that favor technologies’ efﬁcacy in reducing PHL in similar food value chains.


Introduction
Rising incomes in low-income countries are driving changes in dietary patterns and increasing the demand for safe and nutritious food [1]. However, to equate future demand and supply of safe and healthy agricultural food, global food production will need to increase at a rate of 1.3 percent every year [2]. Sustainably achieving such a growth rate will require increasing plant-based food production. Such an effort will promote long-term food security without sacrificing nutrition [3] and will provide increased employment opportunities for farm workers [4].
The two commonly documented approaches for increasing plant-based food production are agricultural intensification and cropland expansion [5]. While both have contributed to global food security substantially, several limitations have also been reported. For example, the former has been challenging to achieve in geographic areas affected by climate change, especially as it pertains to increasing crop yield [6]. Meanwhile, the latter constitutes a potential threat to biodiversity by driving habitat loss. Additionally, cropland expansion impacts carbon storage through the loss of biomass and soil carbon [7].
Given these limitations, numerous studies have suggested postharvest loss (PHL) reduction as an essential and complementary approach to meeting the increasing demand for safe and nutritious food [8]. PHL can be defined as a measurable reduction in agricultural products that arise from changes these products undergo during postharvest handling [9]. Therefore, PHL reduction efforts, especially in sub-Saharan Africa (SSA), could be a catalyst for increasing profit for food value chain actors while at the same time improving food security [10]. Given the importance of PHL reduction, several PHL mitigation studies have been initiated over the last decade, focusing on improving food security in SSA, which remains the most food-insecure region in the world [11].
For example, notable PHL mitigation studies in SSA include introducing the Purdue Improved Crop Storage (PICS) hermetic bags, which prevent storage losses due to insects in maize and other grains without chemical pesticides [12]. The commercialization of this technology, funded by the Bill and Melinda Gates Foundation, led at least five other manufacturers to introduce hermetic storage bag technology products [13]. In 2016, the Rockefeller Foundation launched the Yieldwise Initiative (YWI), intending to provide smallholder farmers (SHF) access to markets, technologies, training, and financing [14] to reduce PHL of mangos in Kenya, maize in Tanzania, and tomatoes in Nigeria. More recently, the Consortium for Innovation in Postharvest Loss and Food Waste Reduction launched as a collaborative effort between the Foundation for Food and Agriculture Research (FFAR), the Rockefeller Foundation, Iowa State University (ISU), and several other academic and research institutions around the world (reducePHL.com (accessed on 2 June 2021)) to address social, economic, and environmental impacts from food loss and waste.
Over time several additional PHL mitigation projects have emerged [15], with a focus on either quantifying PHL by stages of a food value chain [16,17] or comparing the effect of postharvest interventions on PHL incurred at a single stage of a food value chain. However, relatively few PHL mitigation projects have compared the effect of several postharvest technologies on PHL incurred at several stages of a food value chain. Therefore, this study analyzed the YWI dataset generated within the Kenyan mango value chain to evaluate the effect of five YWI promoted technologies (harvesting tools, cold stores, plastic crates, fruit fly traps, and ground tarps) on PHL incurred at three value chain stages (harvest, transportation, and point of sale).
Over the past decades, mango farming in Kenya has expanded considerably, involving several value chain actors such as non-governmental organizations, farmer cooperative groups, aggregation centers, financial institutions, mango processors, and others [18]. Additionally, annual mango production in Kenya is estimated at 1,024,500 metric tons, with approximately 80% being sold to local markets [18]. Thus, mango farming is considered a major income earner for many SHF households in Kenya [19]. However, mango production is accompanied by major PHL estimated at 40-50%, which are mainly the result of a lack of suitable technologies for the postharvest handling and processing into a wide range of value-added mango products [18]. Therefore, comparing YWI promoted technologies and identifying the value chain stage at which they are most effective, is a key step in reducing PHL along the entire value chain and improving SHF livelihoods.

Data Collection
Following the launch of the YWI, the Rockefeller Foundation contracted Technoserve Kenya for implementation of the mango value chain study, whereby they conducted in-person surveys and collected field data from participating farmers and other value chain actors between June and July 2018 (The authors of this paper were neither involved in the survey design nor the data collection process.). Technoserve collected data from 920 SHF (row entries) who provided answers based on September 2017 to March 2018 mango harvesting season. For each respondent farmer, there were 697 recorded variables (column entries) grouped into 12 sections, including geography and socio-demographics, farm demographics, inputs and input costs, labor costs, production, production and PHL practices, harvesting, sales, grading and storage, training, top five sources of household income, and credit access. Finally, the YWI was performed in a quasi-experimental design. Its interventions were not randomly assigned to farmers, and farmers who benefited from the interventions were not randomly selected.

Data Review
Review of the mango dataset began by separating the dependent variables from the independent variables, also referred to as factors in this study. Thus, all numerical variables within the dataset are expressed in the unit of mango fruit, such as mangos consumed, mangos sold, and mangos losses in different ways, and were designated as potential dependent variables. Twenty-five (25) such potential dependent variables were determined from the dataset's 697 variables (total). The remaining 672 variables were designated as factors that potentially affect the dependent variables.

Independent Variables
The 672 potential factors were sorted by removing factors with one or more missing entries, except for the "production and PHL practices" factor. Following the removal of factors with missing entries, the resulting dataset was reduced to 61 factors.
Then, factors containing the respondent farmers' identification information, such as name, contact information, and survey starting and ending times were removed. Additionally, all factors containing "true and false" entries were removed from the dataset. Furthermore, several numerical factors were positively correlated, such as the "total number of mango trees" and "number of productive mango trees" owned by a farmer. In such cases, one (number of productive mango trees) of the two was removed to avoid collinearity [20].
Finally, a listwise deletion of rows within the factor "production and PHL practices" was performed. As mentioned in the first step, this factor was the only one that was not entirely removed from the dataset despite missing entries. The reason being that Technoserve experts suggested the "fruit fly traps," a subset of the "production and PHL practices" factor, played a crucial role in reducing insect infestations of mangos before harvest. Hence, by retaining this factor in the dataset, the importance of "fruit fly traps" in reducing insect infestations of mangos before harvest could be compared to its importance in preserving quality and reducing loss after harvest. The listwise deletion of rows was applied to remove any randomly missing entries of this factor. Although the listwise deletion of rows is a commonly used technique for handling missing data [21], it was only applied to the "production and PHL practices" factor and not to the entire dataset. Using such an approach to the entire raw dataset would have resulted in a 100% loss of information due to multiple missing entries.
The final dataset of factors consisted of nine sections and 21 factors (Table 1), where 19 factors were categorical (each containing at least two subsets), and two were numerical. Therefore, harvest methods, type of storage used after harvest, type of package for sale, and production PHL practices are the four identified factors that contain various technology subsets as specified in Table 1. Their effect on mango PHL will be evaluated in this study. Additionally, certain factors and subsets were renamed to provide more clarity, and some subsets were combined into fewer to facilitate the evaluation of their effect on PHL.
Following factor review and summarization, the four factors that contained postharvest technologies are listed in Table 2, along with their subsets, subset descriptors, and descriptions.

Dependent Variables
The 25 potential dependent variables were also sorted to identify the various types of mango losses along the value chain. The first step consisted of removing variables or columns with at least one missing entry. The second step consisted of identifying all mango PHL along the value chain. Though all 25 potential dependent variables were numerical data representing quantities of mango fruit sold, given to family, used as payment-in-kind, consumed by farmers, and lost along the value chain, not all were PHL variables. PHL variables are the hotspots of loss that form the entire PHL [22]. Therefore, in this study, mango losses that occurred during harvest and losses that occurred after harvest were the only types of losses considered to be PHL variables. Numerical factors consist of numerical values estimated by each respondent farmer. Column (d) renames subsets and combines them into fewer categories to facilitate subsequent analysis. Subset descriptors with the superscript YWI are identified as technologies promoted by the YWI. Column (e) indicates the number of respondent farmers belonging to each subset. For each factor, respondent farmers who reported more than one subset were assigned the subset Other **.

(a) Sections (b) Factors (c) Subsets of Factors (d) Subset Descriptors (e) # Observations (Respondent Farmers)
A. Geography and socio-demographics Following the selection of mango PHL variables, the resulting dependent variables consisted of nine types of mango PHL (Table 3) from the raw dataset's initial 25 potential dependent variables. The nine types of mango PHL were subsequently grouped based on the stages of the value chain at which they occurred ( Table 3).
The third step consisted of identifying and removing outliers [23] from dependent variables. To identify outliers, mango gross production per farmer was calculated for each farmer. The calculation consisted of summing all variables that contributed to mango gross production, including mangos sold, given to family, used as payment-in-kind, consumed by farmers, and all PHL variables shown in Table 3. It was then observed that the calculated mango gross production distribution was skewed with outliers. Hence, removing the rows containing mango gross production outliers resulted in eliminating outliers from PHL distributions at each value chain stage.
The last step consisted of expressing mango PHL at all three value chain stages as percentages of gross production ( Figure 1) for all 753 respondent farmers. Table 2. Summarizing and describing factors containing postharvest technology subsets: Column (a) shows the four factors containing postharvest technologies. Column (b) shows the subset descriptors, which are renamed subsets; these were determined to reduce the raw data into fewer categories, to facilitate subsequent analysis. Column (c) shows the subsets of each factor as initially recorded in the raw data. Column (d) describes the purpose of each subset. The superscripts YWI in Columns (b) and (c) refer to technologies that the YWI promoted.    The PHL data summarized in Figure 1 were subsequently combined with the facto listed in Table 1. This combination resulted in creating the YWI mango dataset (summ rized in Table 4) that formed the basis for the analysis and results presented in this stud Table 4. Summary of all factors, subsets of factors, respondent farmers, and the seven types of mango PHL: Column (a)  The PHL data summarized in Figure 1 were subsequently combined with the factors listed in Table 1. This combination resulted in creating the YWI mango dataset (summarized in Table 4) that formed the basis for the analysis and results presented in this study. Table 4. Summary of all factors, subsets of factors, respondent farmers, and the seven types of mango PHL: Column (a) lists all 21 factors including the 19 categorical C , two numerical N , and four containing postharvest technologies T . Column (b) expands each factor into subsets that were previously referred to as subset descriptors in Table 1. The superscript YWI is used to identify technologies promoted by the YWI. "Other" refers to the combination of multiple subsets as reported by respondent farmers. Column (c) indicates the number of respondent farmers belonging to each subset. Column (d) encompasses mango PHL at harvest, during transportation, at point of sale, and as a total of all three value chain stages. PHL averages cannot be categorized by numerical factors, hence the n/a notation.   21. total trees N # of trees 753 n/a n/a n/a n/a In addition to summarizing the YWI mango dataset in Table 4, each stage's PHL was expressed as a proportion of the total PHL ( Figure 2) by dividing each stage's average by the average PHL of the entire value chain. Furthermore, an online interactive mango PHL dashboard was created (https://phldashboard.shinyapps.io/phldashboard/ (accessed on 2 June 2021)) to explore average mango PHL as a function of each factor in Table 4 Column (a) and as a function of a selected combination of factors.

Statistical Analysis
Identification of the four factors containing postharvest technology subsets (Table 2) and the subsequent quantification of mango losses associated with each subset (Table 4) provided a basis for comparing PHL averages per subset and quantifying the effect size among postharvest technology subsets. However, to ensure that the PHL averages are significantly different among subsets or technologies, a preliminary analysis of the subsets' data was conducted to identify an appropriate statistical tool for comparing means. The initial analysis consisted of verifying the main mathematical assumptions of normality, homogeneity of variance, and independence [24] required to use parametric statistical tools.

Statistical Analysis
Identification of the four factors containing postharvest technology subsets ( and the subsequent quantification of mango losses associated with each subset ( provided a basis for comparing PHL averages per subset and quantifying the ef among postharvest technology subsets. However, to ensure that the PHL aver significantly different among subsets or technologies, a preliminary analysis of sets' data was conducted to identify an appropriate statistical tool for comparing The initial analysis consisted of verifying the main mathematical assumptions of ity, homogeneity of variance, and independence [24] required to use parametric s tools. The assumption of normality was considered violated as the distributions of subset were skewed, and the Shapiro-Wilk normality test results indicated skewed distributions were significantly different (p < 0.05) from a normal dist curve. However, the assumption of homogeneity of variance was not violated as L test results indicated a significant (p > 0.05) homogeneity of variance among subs four factors. Similarly, the assumption of independence was not considered vio PHL distributions per subset were identically distributed to the right for all four Also, observations within each subset were assumed to be independent, althou could be a sampling bias owing to a lack of randomization during the YWI farme tion process.
Consequently, the Kruskal-Wallis statistical test was identified as a suit proach for evaluating the effect of the YWI promoted technologies on mango curred at the three stages of the value chain. The Kruskal-Wallis test is the nonpa analog of a one-way ANOVA, which does not make assumptions about norma and is robust when data contain outlying observations [24]. When the Kruskal-W showed significance, it was followed by a Dunn test with Benjamini-Hochberg ment.
In addition to performing the statistical tests mentioned above, the size of th tion or increase in PHL was also calculated when PHL differences showed signifi < 0.05). The method used for calculating the effect size of the Kruskal-Wallis test The assumption of normality was considered violated as the distributions of PHL per subset were skewed, and the Shapiro-Wilk normality test results indicated that the skewed distributions were significantly different (p < 0.05) from a normal distribution curve. However, the assumption of homogeneity of variance was not violated as Levene's test results indicated a significant (p > 0.05) homogeneity of variance among subsets of all four factors. Similarly, the assumption of independence was not considered violated as PHL distributions per subset were identically distributed to the right for all four factors. Also, observations within each subset were assumed to be independent, although there could be a sampling bias owing to a lack of randomization during the YWI farmers selection process.
Consequently, the Kruskal-Wallis statistical test was identified as a suitable approach for evaluating the effect of the YWI promoted technologies on mango PHL incurred at the three stages of the value chain. The Kruskal-Wallis test is the nonparametric analog of a one-way ANOVA, which does not make assumptions about normality [25] and is robust when data contain outlying observations [24]. When the Kruskal-Wallis test showed significance, it was followed by a Dunn test with Benjamini-Hochberg adjustment.
In addition to performing the statistical tests mentioned above, the size of the reduction or increase in PHL was also calculated when PHL differences showed significance (p < 0.05). The method used for calculating the effect size of the Kruskal-Wallis test was the Epsilonsquared method [26]. Interpretation of the Epsilon-squared effect size was made using the measures of association rules [27]. However, since Epsilon-squared is a squared variable, the upper and lower bound of each bin mentioned were squared [27], yielding the following effect size rule: 0. Lastly, knowing that interventions within the YWI were not randomly attributed to farmers and that farmers who benefited from the interventions were not randomly selected, causal inferences from statistical analysis results to a larger population of SHF can be somewhat speculative. However, thinking of the p-values as approximate p-values for permutation tests will lead to concluding that observed evidence of differences in the results is valid, more so than can be explained by chance [24].

Harvesting Tools
Results indicate a PHL reduction at the point of sale from SHF using harvesting tools over traditional harvesting practices. However, this reduction was not statistically significant (p < 0.05) (Figure 3). Additionally, no PHL reduction was detected during harvest and transportation from SHF using harvesting tools over traditional harvesting practices ( Figure 3). Furthermore, a moderate PHL increase (p < 0.05, Epsilon-squared = 0.096) during transportation was detected due to SHF combining traditional harvesting practices with harvesting tools (Figure 3). 0.64 = strong.
Lastly, knowing that interventions within the YWI were not randomly attributed to farmers and that farmers who benefited from the interventions were not randomly selected, causal inferences from statistical analysis results to a larger population of SHF can be somewhat speculative. However, thinking of the p-values as approximate p-values for permutation tests will lead to concluding that observed evidence of differences in the results is valid, more so than can be explained by chance [24].

Harvesting Tools
Results indicate a PHL reduction at the point of sale from SHF using harvesting tools over traditional harvesting practices. However, this reduction was not statistically significant (p < 0.05) (Figure 3). Additionally, no PHL reduction was detected during harvest and transportation from SHF using harvesting tools over traditional harvesting practices ( Figure 3). Furthermore, a moderate PHL increase (p < 0.05, Epsilon-squared = 0.096) during transportation was detected due to SHF combining traditional harvesting practices with harvesting tools (Figure 3).

Cold Stores
Results indicate a PHL reduction at the point of sale from SHF using cold stores over traditional storage practices. However, this reduction was not statistically significant (p < 0.05). Additionally, no PHL reduction was detected during transportation and at the point of sales owing to SHF using cold stores over alternative traditional storage practices ( Figure 4). Moreover, a weak PHL increase (p < 0.05, Epsilon-squared = 0.01) during harvest was detected due to SHF using cold stores (Figure 4).

M A N G O P O S T -H A R V E S T L O S S P E R M E T H O D S O F H A R V E S T harvesting tools_YWI (n=49)
traditional practices (n=544) other (n=160)

Cold Stores
Results indicate a PHL reduction at the point of sale from SHF using cold stores over traditional storage practices. However, this reduction was not statistically significant (p < 0.05). Additionally, no PHL reduction was detected during transportation and at the point of sales owing to SHF using cold stores over alternative traditional storage practices ( Figure 4). Moreover, a weak PHL increase (p < 0.05, Epsilon-squared = 0.01) during harvest was detected due to SHF using cold stores (Figure 4).

Plastic Crates
Plastic crates were statistically significant (p < 0.05) in reducing PHL incurred at the point of sale (Figure 5), although the effect size of the reduction was weak (Epsilonsquared = 0.017). Additionally, PHL reductions were detected during harvest and transportation due to SHF using plastic crates over traditional packaging practices. However, these reductions were not statistically significant (p < 0.05) ( Figure 5).

Figure 4.
Comparing cold stores to alternative storage types after harvest. Values with different letters are significantly different at p < 0.05 from the Kruskal-Wallis analysis, Dunn test, and Benjamini-Hochberg adjustment. E^2 = Epsilonsquared value for effect size. YWI refers to the technology that the Yieldwise Initiative promoted. (n) refers to the number of farmers who reported using a given practice or technology. 'Other' refers to practices that combined both YWI promoted technologies and traditional practices.

Plastic Crates
Plastic crates were statistically significant (p < 0.05) in reducing PHL incurred at the point of sale (Figure 5), although the effect size of the reduction was weak (Epsilonsquared = 0.017). Additionally, PHL reductions were detected during harvest and transportation due to SHF using plastic crates over traditional packaging practices. However, these reductions were not statistically significant (p < 0.05) ( Figure 5).

M A N G O P O S T -H A R V E S T L O S S P E R T Y P E S O F S T O R A G E A F T E R H A R V E S T cold store_YWI (n=18)
traditional practices (n=686) other (n=49)

M A N G O P O S T -H A R V E S T L O S S P E R T Y P E S O F P A C K A G I N G F O R S A L E plastic crates_YWI (n=320)
traditional practices (n=119) other (n=314)

Plastic Crates
Plastic crates were statistically significant (p < 0.05) in reducing PHL incurred at the point of sale (Figure 5), although the effect size of the reduction was weak (Epsilonsquared = 0.017). Additionally, PHL reductions were detected during harvest and transportation due to SHF using plastic crates over traditional packaging practices. However, these reductions were not statistically significant (p < 0.05) ( Figure 5).

M A N G O P O S T -H A R V E S T L O S S P E R T Y P E S O F S T O R A G E A F T E R H A R V E S T cold store_YWI (n=18)
traditional practices (n=686) other (n=49)

M A N G O P O S T -H A R V E S T L O S S P E R T Y P E S O F P A C K A G I N G F O R S A L E plastic crates_YWI (n=320)
traditional practices (n=119) other (n=314)

Fruit Fly Traps and Ground Tarps
Fruit fly traps were statistically significant (p < 0.05) in reducing PHL incurred at the point of sale (Figure 6), although the effect size of the reduction was weak (Epsilonsquared = 0.017). Additionally, PHL reduction was detected during transportation due to SHF using fruit fly traps over traditional production practices. However, this reduction was not statistically significant (p < 0.05). Moreover, no PHL reduction was detected during harvest from SHF using fruit fly traps over traditional production practices ( Figure 6). Meanwhile, moderate PHL increases during harvest (p < 0.05, Epsilon-squared = 0.04) and during transportation (p < 0.05, Epsilon-squared = 0.06), and a weak PHL increase (p < 0.05, Epsilon-squared = 0.017) at the point of sale were detected from SHF using ground tarps over any other harvest practice ( Figure 6). reported using a given practice or technology. 'Other' refers to practices that combined both YWI promoted technologies and traditional practices.

Fruit Fly Traps and Ground Tarps
Fruit fly traps were statistically significant (p < 0.05) in reducing PHL incurred at the point of sale (Figure 6), although the effect size of the reduction was weak (Epsilonsquared = 0.017). Additionally, PHL reduction was detected during transportation due to SHF using fruit fly traps over traditional production practices. However, this reduction was not statistically significant (p < 0.05). Moreover, no PHL reduction was detected during harvest from SHF using fruit fly traps over traditional production practices ( Figure 6). Meanwhile, moderate PHL increases during harvest (p < 0.05, Epsilon-squared = 0.04) and during transportation (p < 0.05, Epsilon-squared = 0.06), and a weak PHL increase (p < 0.05, Epsilon-squared = 0.017) at the point of sale were detected from SHF using ground tarps over any other harvest practice ( Figure 6). Figure 6. Comparing fruit fly traps to alternative production practices. Values with different letters are significantly different at p < 0.05 from the Kruskal-Wallis analysis, Dunn test, and Benjamini-Hochberg adjustment. E^2 = Epsilon-squared value for effect size. YWI refers to the technology that the Yieldwise Initiative promoted. (n) refers to the number of farmers who reported using a given practice or technology. 'Other' refers to practices that combined both YWI promoted technologies and traditional practices.

Discussion
While mango SHF have reported seeing a PHL reduction due to using harvesting tools [28,29], traditional harvesting practices such as handpicking can also reduce damage caused during harvest [30][31][32]. Hence, increasing the adoption of correct mango handpicking practices could be effective, if not more effective, than harvesting tools (Figure 3).
Cold stores utilized by SHF (photovoltaic-powered coolers, charcoal evaporative coolers, and brick evaporative coolers) effectively preserve mangos [33,34]. However, they are costly for individual farmers to own. Hence, most mango cold stores are owned by farmers' cooperatives [33], requiring farmers to inspect mangos during harvest and only store fruits that can be well preserved in the cold stores. Therefore, PHL increase during harvest from SHF using cold stores (Figure 4) can be attributed to large quantities of poor fruit quality set aside during the inspection process before storage.
Packaging mangos in plastic crates instead of sacks allows adequate packaging and storage of mangos [35] needed to preserve quality and provide greater wholesale value

fruit fly traps_YWI (n=125)
traditional practices (n=310) other (n=203) tarp_YWI (n=115) Figure 6. Comparing fruit fly traps to alternative production practices. Values with different letters are significantly different at p < 0.05 from the Kruskal-Wallis analysis, Dunn test, and Benjamini-Hochberg adjustment. Eˆ2 = Epsilon-squared value for effect size. YWI refers to the technology that the Yieldwise Initiative promoted. (n) refers to the number of farmers who reported using a given practice or technology. 'Other' refers to practices that combined both YWI promoted technologies and traditional practices.

Discussion
While mango SHF have reported seeing a PHL reduction due to using harvesting tools [28,29], traditional harvesting practices such as handpicking can also reduce damage caused during harvest [30][31][32]. Hence, increasing the adoption of correct mango handpicking practices could be effective, if not more effective, than harvesting tools (Figure 3).
Cold stores utilized by SHF (photovoltaic-powered coolers, charcoal evaporative coolers, and brick evaporative coolers) effectively preserve mangos [33,34]. However, they are costly for individual farmers to own. Hence, most mango cold stores are owned by farmers' cooperatives [33], requiring farmers to inspect mangos during harvest and only store fruits that can be well preserved in the cold stores. Therefore, PHL increase during harvest from SHF using cold stores (Figure 4) can be attributed to large quantities of poor fruit quality set aside during the inspection process before storage.
Packaging mangos in plastic crates instead of sacks allows adequate packaging and storage of mangos [35] needed to preserve quality and provide greater wholesale value for the fruit [36,37]. Packaging mangos in crates can also reduce damage caused to the fruit during transportation (Figure 5), and by extension, reduces PHL at the point of sale ( Figure 5). Furthermore, plastic crates had the highest adoption (n = 320, Figure 5) of all the YWI promoted technologies as SHF and value chain actors saw value in using them.
Fruit fly traps were statistically significant in reducing PHL at the point of sale as their adoption was relatively higher (n = 125, Figure 6) than the other YWI promoted technologies, except for plastic crates. Two major factors were reported related to slowing the adoption of fruit fly traps. First, farmers' beliefs that fruit fly traps attract fruit flies from other farms caused the farmers to remove traps, leaving mangos susceptible to infestation and diminishing fruit fly traps' efficacy over traditional production practices [35]. Second, although farmers reported having fruit fly trap containers, without adequate financing, they could not refill the fruit fly trap containers with bait refills frequently enough for the traps to be effective [35]. Thus, overcoming these challenges could result in higher adoption of fruit fly traps.
Although essential, increased adoption and access to the technologies can be difficult to achieve. Cold stores, for example, are too expensive for SHF to own or utilize, especially without access to affordable credit [28,29]. Hence their adoption within the YWI was relatively low (n = 18, Figure 4). On the other hand, technologies easily accessible to farmers, such as plastic crates and fruit fly traps [38], had a relatively higher adoption rate. Hence, providing SHF easier access to affordable credit through innovative financing [35] or lower discount rates [39] could be an essential and initial step toward enabling increased adoption of preferred technologies. Alternatively, facilitating access to postharvest technologies through innovative subsidy programs could also increase the adoption of preferred technologies [40,41].
Lastly, discussions with Kenyan SHF revealed that buyers mainly do the harvesting and thus due to the informal and often hierarchal relationships between the two groups, farmers cannot intervene with the harvesting. Therefore, they do not have a say about whether or not ground tarps are used, increasing the chances of experiencing PHL during harvesting and, by extension, several other PHL types along the value chain ( Figure 6). Moreover, training and promotion of ground tarps delivered through the YWI may lose their impact over time, and refresher training will be necessary [36].

Conclusions
This study quantitatively compared postharvest technologies and their effects on mango PHL in Kenya via the Rockefeller Foundation's YWI. Five YWI promoted technologies were compared to Kenyan SHF's traditional practices at three value chain stages. Subsequently, the following conclusions were inferred from analyzing the YWI mango dataset: Efforts to reduce PHL in the mango value chain should prioritize adopting plastic crates and fruit fly traps. These technologies were statistically significant in reducing PHL incurred at the point of sale. In addition to preserving quality, plastic crates and fruit fly traps can be easily accessed and adopted by SHF compared to harvesting tools, cold stores, and ground tarps.
Harvesting tools as a YWI promoted technology and handpicking as a traditional practice to harvest mangos are similar in that both require careful handling of the fruit when picking. Therefore, PHL reduction from SHF using harvesting tools was not statistically significant because handpicking can effectively reduce mango PHL when done correctly. Further research is needed to determine factors other than increased adoption that increase the effectiveness of harvesting tools in reducing PHL.
PHL reduction from SHF using cold stores was not statistically significant. While several factors can contribute to this lack of statistical significance, this study posits that the low adoption of cold stores among SHF is due to their high cost of ownership or utilization.
The benefits of ground tarps should be further investigated because SHF are not always involved in the harvest and do not have a say about whether or not ground tarps are used, resulting in increased PHL. Additionally, training and promotion of technologies delivered through the YWI may lose their impact over time, and refresher training is recommended.
While this study asserts that increased technology adoption is necessary to obtaining better PHL reduction efficacy, further research is needed to identify additional factors of importance that favor technologies' efficacy in reducing PHL in similar food value chains.