*2.4. Data Collection*

Primary data were collected from the 174 randomly selected banana farmers using a semistructured questionnaire. The questionnaire was pretested for reliability and suitability and then modified for clarity and sequencing of questions based on the pretest experiences and results. Data were collected through face-to-face interviews with the selected banana farmers. Data were collected on seed security factors, farmer competencies and farmer characteristics related to uptake of the TC banana planting materials.

### Measures

Availability was measured based on statements (e.g., "TC is available in sufficient quantities"). Accessibility was measured based on 3 items (e.g., "The TC nursery operator provides information on how to plant TC plantlet"), acceptability was measured based on 3 items (e.g., "TC banana plantlets is of desirable banana varieties"), and adaptability was measured based on 2 items (e.g., "TC banana plantlets thrive in all soil types"). To measure farmer competencies, 8 statements were used. Knowledge was measured with 5 items (e.g., "I have sufficient technical knowledge to grow TC banana plants"), and attitude was measured with 3 items (e.g., "I have enough passion for growing TC banana plants"). The social influence variable had 3 items. A sample item reads as follows: "If I am informed about TC by a community leader, then I can use it as banana seed". Each item was measured on a five-point rating scale from 1 = least and 5 = highest.

In addition, the questionnaire included demographic characteristics of the farmers: These included age in years, the highest level of formal education in years, the gender of the farmer, farm size in acres, experience in banana farming in years, access to agricultural extension services, membership to farmer groups, receipt of agricultural information, land apportioned for crop production and banana cultivation.

### *2.5. Analytical Framework*

Preceding the regression analysis, principal component analysis (PCA) was carried out for data reduction and extraction of variables. Based on the criterion of eigenvalues being greater than 1 [53] in the PCA, underlying dimensions among the farmers' perceptions of the TC plantlets' attributes were generated.

This study breaks the banana farmer decision-making into two stages. In the first stage, the banana farmer's decision is either uptake of the banana TC technology or not. In the second stage, the farmer's decision is on how much of the TC plant material to use (uptake intensity), which in this study is represented by the size of land allocated to the banana TC plantlets. Data were analyzed using Cragg's double-hurdle model [54] consisting of a two-stage regression in which the first stage is a probit model to analyze factors influencing a binary decision on uptake of technology and the second stage is a truncated model which analyzes factors that affect uptake intensity. Farmer uptake of banana TC plantlets was analyzed in the following equation:

$$Y = \beta\_0 + \beta\_i [\text{Seed\\_Sec}]\_i + \beta\_j [\text{Farm\\_Envt}]\_j + \varepsilon \tag{1}$$

where *Y* represents farmer's decision-making for the uptake of banana TC plantlets uptake = 1; no uptake = 0) and in the second stage, *Y* equals the uptake intensity. [*Seed*\_*Sec*] is a vector of plant security factors, which includes perceived acceptability, accessibility, adaptability and availability of the plantlets. [*Farm*\_*Envt*] comprises three components: (1) social influence, (2) farmer competencies, and (3) farmer socioeconomic factors or characteristics. β<sup>0</sup> is the constant while β*<sup>i</sup>* represent the various coefficients of the security factors ranging 1–4. β*<sup>j</sup>* are the coefficients of the farmer environment factors ranging from 1 to 9 and ε is the error term. The a priori hypothesized signs of the coefficients are shown in Table 2.

**Table 2.** Apriori signs of explanatory variables used in the study.


#### **3. Results and Discussion**

Table 3 presents a summary of key descriptive statistics of study respondents. Results show an equal number of male and female farmers were interviewed. About 33% of respondents have adopted the TC banana planting materials. The average age of banana farmers in the study communities was 43 years, with an average of eight (8) years of formal education. The average farm size for all farmers was 4.3 acres, with an average of about 2.6 acres allocated to crops and about 1.1 acres to banana production.


**Table 3.** Description and summary statistics of respondent characteristics.

Standard deviation (SD) is in parenthesis.

Results of the Kaiser–Meyer–Olkin (KMO) measure (0.815) and Bartlett's test of sphericity (χ = 3804.143; *p* < 0.001) (Table 4) indicate sampling adequacy and suitability of the data for factor analysis [55]. Principal component results show that seven (7) extracted factors explained 76.2% of the total variance in the principal components. Specifically, as presented in Table 4, the variance extracted ranged from 32.4% (for farmer knowledge) to 4.4% (TC plant adaptability). In addition, the factor loadings for the extracted variables ranged from 0.508 to 0.925, and thus, convergent validity was confirmed [56]. Lastly, the Cronbach's alpha values (Table 4) range from 0.600 to 0.992, signifying adequacy of internal consistency, and thus, confirmation of measurement validity [57] in this study.


**4.**LoadingsofperceptionandattitudefactorsforuptakeofTCbananaplantingmaterials(*<sup>n</sup>*=174).


**Table 4.** *Cont.*

#### *3.1. Factors Associated with the Uptake Decision for Farmer Use of TC Banana Plantlets*

Results of the first stage analysis (probit regression) show that farmer perceived acceptability (β = 0.74; *p* < 0.01) has a positive and significant influence on farmer decisions for the uptake of the banana TC planting material (Table 5). Similarly, perceived adaptability (β = 0.69; *p* < 0.01) and perceived availability (β = 1.04; *p* < 0.01) had a positive and significant influence on uptake. These findings are in agreement with other studies regarding the influence of security factors on the uptake of improved plant materials [15,20,21]. In terms of marginal effects, TC plant acceptability (0.062) implies that, on average, a 1% increment in farmer acceptance of banana TC plants increases the probability of farmer uptake by 6.2%. This is likely because farmers tend to prefer introduced varieties that are comparable to their local varieties with regard to desirable attributes. Similar results are also reported by Mulugo et al. [33] and Akankwasa et al. [12] for the case of Uganda.


**Table 5.** Factors that influence farmers' decisions regarding adoption of tissue culture banana planting materials: results of the probit model.

\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001; dependent variable: uptake measured as a binary.

For farmer perceived adaptability of TC plantlets (marginal effect = 0.058), an additional increase in adaptability is associated with a probability of a 5.8% increase in uptake. Notably, the importance of TC plant adaptability to drought and poor soil fertility is emphasized [5,34]. Lastly, the average marginal effect of TC plant availability (0.087) implies that an additional TC nursery is associated with a likelihood of an 8.7% increase in farmer uptake of TC plantlets. Related studies like Okechukwu and Kumar [27] on availing disease-resistant varieties in Africa and Kromann et al. [23] on the provision of quality seed in Ecuador have reported sufficient and timely availability of quality plant material to increase farmer uptake of plant technologies.

Farmer knowledge about TC plant cultivation (β = 2.16; *p* < 0.001) has a positive and significant effect on uptake. Similarly, farmer attitude (β = 1.10; *p* < 0.01) posits a positive and significant effect on uptake. These results concur with the theory that farmer competencies predict uptake [19,41,42]. In terms of marginal effects, the result on farmer knowledge about TC plants (0.180) implies that, on average, an increment in farmer skill and technical knowledge on how to grow TC plantlets increases the likelihood of farmer uptake by 18%. This finding validates the study by Atieno and Schulte-Geldermann [24] on public–private partnerships in the multiplication of plant materials in Kenya. Specifically, knowledge sharing and training by extension agents increase farmer uptake of new technology.

Similarly, our findings substantiate previous research about farmer attitudes towards the uptake of introduced agricultural technologies [58,59]. Particularly, Mekoya et al. [58] found Ethiopian farmers' positive attitude towards multipurpose fodder trees (for their feed value and contribution to soil conservation) enhanced farmer uptake of the agroforestry technology.

Social influence (β = 0.81; *p* < 0.05) also had a positive and significant influence on farmer decisions for uptake of the TC banana plants [19,60]. The average marginal effect (0.068) infers that having an influential person inform farmers about the importance of banana TC plants increases the likelihood of farmer uptake by 6.8%, suggesting that locally institutionalized mechanisms need to be promoted. Similar results are also reported by Mulugo et al. [17] and Wauters et al. [61] for Uganda and Belgium, respectively. Specific for Uganda, using faith-based leaders, political and community leaders urge farmers to use TC banana plantlets is a significant predictor of farmer intentions to use TC planting materials.

Contrary to most studies [12,62,63], the model estimates show that the gender of the farmer (β = −1.38; *p* < 0.05) has a significant effect on the uptake decision, indicating that women were more likely to use TC banana planting materials compared to men. A possible explanation for women's interest in TC banana planting material could be attributed to the distinct role that women play in ensuring food security for their families. Food security in central Uganda is culturally viewed as a women's role, and women must ensure that their homesteads are food secure (Sanya et al. [64]). The study finding is in line with Sanya et al. [64], who found female farmers to be 12.4% more likely to adopt hybrid banana varieties compared to male farmers. Nevertheless, with recent developments by the government being skewed to banana-based processing enterprises, it is likely that men will increasingly grow bananas to boost their sources of income.

The amount of land allocated to crop production (β = 0.34; *p* < 0.05) has a positive influence on the likelihood of farmer uptake of the TC banana technology. Based on the average marginal effect (0.029), increasing the land size by one acre increases the likelihood that farmers would try the TC planting materials by 2.9%. A possible explanation for this finding could be that farmers with larger farms may be more willing to take risks and devote portions of their land to growing bananas using TC plantlets, compared with those with smaller land areas. This is in line with most adoption studies [50,65,66] that found that farmers with larger farm sizes have more land to allocate to new agricultural technologies.

#### *3.2. Factors Associated with Intensity of Farmer Uptake of TC Banana Plantlets*

Results of second stage analysis (truncated regression) show that farmer perceived acceptability (β = 0.39; *p* < 0.05) and accessibility (β = 0.39; *p* < 0.01) of TC plants posit a positive and significant influence on farmer decisions (Table 6). Essentially, the result specifies that if farmers perceive that the varieties being promoted through TC match their preferred food attributes, then they are more likely to expand their banana plantations using such a technology. Previous studies [12,33] attest to this finding.

Similarly, farmer accessibility (marginal effect = 0.058) indicates that farmers with fairly priced plant material coupled with information on how to grow it promotes plantation expansion by 5.8%. This finding is in tandem with studies [15,23,24] confirming accessibility in terms of affordability and awareness creation to be crucial for use and extent of uptake of introduced plant technologies. Specifically, access to quality potato planting material in Kenya increased farmer uptake of the planting materials by 30–40%, leading to increased yields (5184 t/year of potatoes) and profits (\$777,600) [24].



\* *p*< 0.05, \*\**p* < 0.01, \*\*\**p* < 0.001; dependent variable: proportion of land planted with banana TC seed.

Farmer knowledge about TC plant materials (β = 1.46; *p* < 0.001) and farmer attitude towards TC technology (β = 0.52; *p* < 0.01) both postulate a positive and significant influence on farmer decisions to allocate more of their land to TC banana production. The findings concur with previous research about farmer knowledge on introduced agricultural technologies to influence uptake and extent of use of improved crop varieties [19,24,41,42]. Similarly, farmer attitude result resonates with findings by [67] in Kenya, showing farmer attitudes to be key indicators in predicting uptake of aquaculture technologies among smallholder fish farmers.

Surprisingly, social influence (β = −0.19; *p* < 0.05) had a negative influence on farmer decisions to apportion more of their land to TC banana plantlets. This varied from the results of earlier research [47,68], which reported positive effects of social influence on farmer uptake of agricultural technologies. This probably is attributable to the cost implication of expanding banana cultivation using TC plant materials, inevitably requiring farmers to purchase plantlets, inputs and follow through with the recommended cultivation instructions that often are labor-intensive [14]. As such, it may be important that a related study on social influence is designed in a different context to confirm the actual impact of social influence on farmer intensity of use of agricultural technologies.

Although not a significant predictor of uptake, farm size (β = 0.12; *p* < 0.05) had a positive influence on uptake intensity of TC banana planting materials.

#### **4. Conclusions**

This study shows that farmers' decisions for the uptake of banana TC plants are positively and significantly influenced by seed security factors. From a practical perspective, the study contributes to results that show the importance of developers in the seed system in focusing on farmer desired crop attributes. Further, the study emphasizes the need for more involvement of extension services and research institutions in education about cultivation and promotion of TC planting materials in the banana farming communities. This involvement could incorporate the use of community role models since social influence plays a pivotal role in increasing uptake. We recommend seed security factors (acceptability, accessibility, adaptability and availability), social influence and farmer competence (knowledge, skill and attitude) as variables to be considered in programs aimed at increasing farmer uptake for seed system technologies.

**Author Contributions:** L.M., F.B.K., P.K., B.A.O. and E.M.K. conceived and designed the study. L.M., F.B.K. and P.K. developed the data collection tools. L.M., F.B.K. and P.K. participated in field research. L.M. and F.B.K. analyzed and interpreted the data, E.M.K., and L.M. acquired funding. All authors participated in writing, reviewing and editing the article. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the CGIAR Research Program (CRP) Roots Tubers and Bananas (RTB) through the "RTB Seed Systems Project" grant number 96202-500-A1050-B100461-C100078-UGAND and the Regional Universities Forum for Capacity Building in Africa (RUFORUM) with funding from the Carnegie Corporation of New York grant number RU/2016/Carnegie/DRG/009.

**Acknowledgments:** We are grateful to the banana farmers in the study communities who participated in this study. Our appreciation goes to Emmanuel Ngolobe, Eva Tereka, Ruth Nakintu, Ssebunya Mohammed Ali and Ssenyonjo Michael for their participation in data collection. Justine Kyobe, Sserunkuma Margaret, Nakafu Margaret Kyasa (Mukono district) and Kiggundu (Luweero district) are appreciated for their role in mobilizing the respondents. This research was undertaken as part of the CGIAR Research Program on Roots, Tubers and Bananas (RTB).

**Conflicts of Interest:** The authors declare no conflict of interest.
