The application of ICTs in sorghum production
Our investigation identified the widespread application of ICTs throughout the farming life cycle (Fig. 2). The survey results revealed that most household heads (68.91%) utilized ICTs across all three stages of farming: pre-cultivation, cultivation, and post-harvest. A smaller portion (15.97%) employed ICTs in two stages, primarily pre-cultivation and cultivation. Additionally, 6.72% of respondents relied on ICTs solely during the pre-cultivation stage, while 4.2% utilized them exclusively during the combined stages of cultivation with harvesting, and 4.2% for post-harvest.
Application of ICTs in the Sorghum production cycle. Source: Authors’ analysis based on survey data.
Focus group discussion (FGD) participants described they are using ICTs in various ways, including direct calling, the 8028-hotline service, and text messaging via mobile phones. Additionally, they mentioned accessing information through radio and television broadcasts, as well as attending audio-visual programs offered by Pico devices. The participants stated:
“We use direct calling, text SMS, and the 8028-hotline service by using mobile phones, and we also attend radio, TV, and Pico programs.”
This participant statement shows a multi-faceted strategy for obtaining information in the farming cycle. They use both traditional as well as more contemporary forms of communication.
Mobile phone communication: Direct telephone call and text SMS are common and popular channels of interpersonal communication and getting personalized information.
8028 Hotline Service: This would suggest a single, perhaps special service for accessing a certain information or aid. The“8028”number would likely refer to a local or regional service.
Mass Media (Radio and Television): They are the traditional broadcast media, with the ability to address a massive audience with wide information and focused programs.
Pico Programs (Audio-Visual): This implies the use of Pico devices, which are usually small, portable projectors. These devices are used in an effort to distribute audio-visual content which are developed with in the native community and indigenous knowledge, perhaps for educational or informational purposes, in a communal setting.
The use of such a diverse range of methods suggests an effort to reach a wide segment of the population through channels that are both accessible and familiar to them. It also implies that different types of information or engagement might be delivered through these various platforms.
Then, we pick the points from the findings ICT application into 3 main stages of farming as follows in Fig. 3.
Outlined ICT applications in the farming cycle. Source: Authors’ findings based on survey data.
ICT usage within Socio-economic, infrastructural, and perception variables
The ANOVA is used to compare means across different groups. In this case (Radio, Mobile, TV, and Pico) across Age, Income, and Education. There are significant differences in age among mobile, Pico, Radio, and TV users. There are significant differences in age and income among mobile, Pico, Radio, and TV users (Table 1). Let’s see one by one:
There is a significant negative difference between the ages of those who primarily use mobile devices and those who primarily use radio. This suggests that mobile users tend to be younger. There are significant differences in age between Pico users and Mobile users, with Mobile users being younger. The result shows a significant positive difference in income between TV and radio users, suggesting that TV users have higher incomes. There is a significant difference in income between Pico and TV users, indicating that Pico users tend to have lower incomes. Although there is an overall significant difference in education, there are non-significant differences in education levels between Pico, Mobile, Radio, and TV users.
Table 2 presents the results of a Chi-Square test, which shows a significant association between categorical variables. In this case, we’re examining the relationship between various factors and the choice of ICT (Radio, Mobile, TV, Pico).
There is a significant difference between males and females in the choice of Pico, Mobile, Radio, and TV. The test reveals a significant difference between those who have and don’t have electric power access in the choice of Pico, Mobile, Radio, and TV. In addition, there is a significant difference between those who have and don’t have network access in the choice of Pico, Mobile, Radio, and TV.
There is a significant difference between those who perceived and didn’t perceive the relative advantage of technology in terms of tool choice. Additionally, there is a significant difference between those who got and didn’t get hedonic motivation from technology in the choice of Pico, Mobile, Radio, and TV. Furthermore, there is a significant difference between those who chose tools that fit with them and didn’t in the choice of Pico, Mobile, Radio, and TV. It also shows a significant difference between those who chose first info quality and didn’t in the choice of Pico, Mobile, Radio, and TV.
There is a significant association between social influence and ICT choice. There is a significant difference between those who have received peer assistance and those who haven’t in the utilization of Pico, Mobile, Radio, and TV. The test shows a significant difference between those who perceived and didn’t perceive the lower price value in the choice of Pico, Mobile, Radio, and TV.
Factors determining farmers’ ICT preference
Our study investigated the factors influencing farmers’ preferences for specific Information and Communication Technologies (ICTs) in their agricultural practices. The dependent variable was the chosen tools class, with radio designated as the reference category for comparison with other options. This choice aligns with Martin’s19 suggestion to compare categories based on logical or significant differences, in this case, the distinct audio-only nature of radio compared to the audio-visual capabilities of other ICTs.
To ensure the model’s robustness, we assessed multicollinearity using the Variance Inflation Factor (VIF). All VIF values were below 10, indicating no problematic multicollinearity. Additionally, the Hausman test confirmed the validity of the Independence of Irrelevant Alternatives (IIA) assumption, signifying that the dependent variables were mutually exclusive. The model’s goodness-of-fit was further supported by a highly significant likelihood ratio test (p < 0.00) and a pseudo-R-squared value of 0.6985, suggesting that the independent variables explained approximately 69.85% of the variation in ICT preference.
While parameter estimates in the Multinomial Logit (MNL) model reveal the direction of the effect of independent variables on the dependent variable, they don’t directly represent the magnitude of change or probabilities. Therefore, marginal effects were employed to estimate the expected change in the probability of choosing a specific ICT class due to a unit change in an independent variable14. Our analysis identified 11 out of 14 hypothesized independent variables as having significant influences on ICT preference, as detailed in Table 3. The positive or negative signs of these significant variables indicate whether the likelihood of preferring a particular tool class increases or decreases relative to the reference category (radio) with a unit change in the corresponding variable.
The Multinomial Logit (MNL) analysis revealed that gender significantly influenced farmers’ decisions to choose Pico devices, with a negative effect at the 1% significance level. The marginal effect indicated an approximate 2.1% increase in the probability of female farmers choosing Pico compared to male farmers. Age emerged as another significant factor, negatively impacting the choice of mobile phones at the 1% significance level. A one-year increase in age was associated with a 5.3% decrease in the probability of utilizing mobile phones.
Education level also exerted a negative influence on the preference for Pico devices, with a statistically significant effect at the 1% level. Each additional year of formal education was associated with a 0.14% decrease in the likelihood of using Pico. Farm income emerged as a positive predictor for both mobile phone and TV utilization with a marginal effect of 0.0000154 and 5.96e-06, respectively. This means when farm income increases by one unit the probability of using Mobile increases by 0.00154% and TV by 0.000596%.
Network access played a crucial role in mobile phone usage, with a significant positive effect. The marginal effect indicated that access to a network significantly increased the likelihood of using a mobile phone by approximately 65.9%. Interestingly, electricity access positively impacted the choice of mobile phones, TV, and Pico projectors compared to Radio. This suggests that access to electricity significantly increased the probability of using mobile phones, TVs, and Pico by about 34.3%, 52.5%, and 0.06%, respectively.
Perceived relative advantage positively influenced the choice of both mobile phones and TVs. Farmers who perceived a tool as having a higher relative advantage were more likely to choose it, with marginal effects indicating a 67.9% and 0.57% increase in the probability of using mobile phones and TVs, respectively. Hedonic motivation, reflecting the desire for enjoyment and relaxation, also exerted positive effects on the selection of Mobile phones, TVs, and Pico devices. The marginal effect revealed that when farmers sought more relaxation, the probability of choosing mobile phones, TVs, and Pico devices increased by 44.6%, 15%, and 0.2%, respectively.
Compatibility perception emerged as a significant predictor of mobile phone usage. The marginal effect indicated that a positive perception of compatibility with a tool corresponded to a 78.2% increase in the likelihood of utilizing a mobile phone. Furthermore, information quality positively influenced the choice of mobile phones. Farmers seeking high-quality information were more likely to choose mobile phones, with the marginal effect suggesting a 44.3% increase in the probability of doing so. Finally, the perceived price value of a tool positively impacted the choice of Pico devices. When farmers considered the price-value proposition to be favorable, the probability of using Pico increased by approximately 0.5%.
These findigs outlined and Summarized key Marginal effects and their practical implications in Table 4 below:
link