Agriculture underpins rural livelihoods and food security across South Asia, yet productivity growth remains constrained by persistent information deficits. In India, approximately 70 per cent of rural households depend primarily on agriculture, with 82 per cent classified as small and marginal farmers who are disproportionately excluded from formal knowledge networks (Anonymous, 2017). In Nagaland, nearly 70% of the workforce depends on predominantly subsistence and shifting cultivation systems that are highly sensitive to the timeliness and quality of agricultural information. Despite rich biodiversity and agricultural potential, low productivity persists, underscoring the need to strengthen knowledge dissemination alongside conventional agricultural interventions.
       
Information and Communication Technologies (ICTs), including mobile phones, smartphones, the internet, radio and digital advisory platforms, offer an effective means of bridging this information gap. Evidence from developing countries shows that ICTs improve farmers’ access to market information, weather forecasts and agronomic advisories, leading to higher productivity and incomes (Ndimbo et al., 2023; Kumar, 2023). In India, mobile-based extension services have expanded knowledge delivery beyond the limitations of the formal extension system (Singh et al., 2023; Babu et al., 2025), while ICT applications have also enhanced the performance and profitability of Farmer Producer Organizations through improved farm management, precision agriculture and value-chain integration (Salokhe, 2025). However, greater availability of ICTs does not necessarily ensure their adoption. Even where access exists, farmers in remote and hilly regions often remain non-adopters (Nyakudya et al., 2024), making the identification of adoption determinants a policy priority.
       
Several theoretical frameworks explain ICT adoption behaviour. The Technology Acceptance Model (TAM) identifies perceived usefulness and perceived ease of use as the principal cognitive determinants of adoption (Davis, 1989). The diffusion of innovations (DOI) theory emphasizes innovation characteristics and communication processes (Rogers, 2003), while the Unified Theory of Acceptance and Use of Technology (UTAUT) incorporate performance expectancy, effort expectancy, social influence and facilitating conditions (Venkatesh et al., 2003). A recent meta-analysis confirmed that education, income, age and gender consistently influence ICT adoption, although their effects vary across contexts (Sossou et al., 2024). Since no single framework adequately captures both behavioural and structural barriers, integrating multiple theories provides a more comprehensive understanding of ICT adoption (Aminu et al., 2024).
       
Empirical evidence indicates that ICT adoption is influenced by individual, household and contextual factors. Knowledge, digital literacy and favourable attitudes consistently promote adoption (Davis, 1989; Kumar, 2023; Ndimbo et al., 2023; Singh et al., 2023; Nyakudya et al., 2024). Recent evidence from Nagaland further shows that while farmers possess basic awareness of ICT tools, higher-order cognitive skills such as application, analysis, evaluation and creation remain limited, constraining meaningful ICT use (Sentinungshi et al., 2025). Household characteristics, including income, farming experience, age and education, also significantly influence ICT access and adoption (Mittal et al., 2010; Babu et al., 2025; Nagar et al., 2025). Likewise, education, extension participation, training, mass media exposure, information accessibility, innovativeness, social participation and economic motivation have been found to enhance both the extent and effectiveness of ICT use (Chaudhary and Gardhariya, 2024). At the contextual level, reliable network infrastructure, technical support and diverse information sources remain essential for successful adoption, particularly in geographically challenging regions (Nyakudya et al., 2024; Kath and Mezhatsu, 2022; Ndimbo et al., 2023; Aguilar-Gallegos  et al., 2015; Nikam et al., 2021).
       
Despite growing evidence, the North-East Indian context, particularly Nagaland, remains underrepresented in ICT adoption research. The region’s mountainous terrain, tribal social structure, limited extension services and diverse subsistence farming systems create conditions that differ substantially from those of the Indian plains. Although studies have examined ICT perception, usage patterns and cognitive dimensions among farmers (Sharma et al., 2025; Kath and Mezhatsu, 2022; Sentinungshi et al., 2025), econometric evidence on the socio-economic, cognitive and infrastructural determinants of ICT adoption remains scarce. This gap limits the development of context-specific policies and extension strategies. Therefore, the present study aimed to assess the extent of ICT adoption among farmers in Nagaland and identify the socio-economic, cognitive and infrastructural factors influencing adoption decisions within an integrated TAM, DOI and UTAUT framework.

The study was conducted in Nagaland, India in the year 2024-25 and adopted an analytical cross-sectional survey design. Of the 17 districts in Nagaland, six districts were purposively selected namely, Mon, Tuensang, Phek, Wokha, Mokokchung and Kohima; on the basis of agricultural activity levels, accessibility for data collection and representation of diverse agro-ecological zones in the state. From each district, two blocks were randomly selected (total: 12 blocks) and from each block, two villages were randomly selected (total: 24 villages). Fifteen respondents were randomly selected from each village, yielding a total sample of 360 respondents. This multistage sampling procedure ensured both geographical representativeness and randomness at the sub-district and village levels.
       
A structured interview schedule was developed in accordance with the study’s objectives and subjected to content validation by a panel of five subject-matter experts in agricultural extension and ICT. The schedule was pre-tested on 30 farmers outside the study area and reliability was assessed using Cronbach’s alpha, which ranged from 0.78 to 0.84 across the sub-scales for knowledge, attitude, accessibility and availability, indicating acceptable internal consistency. The outcome variable (ICT adoption) was operationalized as a binary variable: adopter (1) if the respondent reported regular use of at least one ICT tool (mobile phone, internet, agricultural app, or digital advisory service) for farming purposes and non-adopter (0) otherwise. Independent variables were operationalized as follows: Knowledge was assessed using a 10-item scale covering awareness and understanding of available ICT tools (score range 0-10); Attitude was measured using a 5-point Likert-type scale comprising 12 statements on perceived usefulness and ease of use of ICT tools (score range 12-60); Source of information was scored as a composite index of the number and diversity of information channels accessed by the respondent; Availability was measured as a binary or Likert-scale index of respondent-perceived availability of ICT infrastructure and services in their locality; Accessibility was a self-rated indicator of physical and financial ease of accessing ICT tools. Annual income was recorded in Indian Rupees (INR) and entered as a continuous variable.
       
Data were analysed using IBM SPSS Statistics Version 26. Binary logistic regression was applied to identify significant predictors of ICT adoption, using the enter method to include all pre-specified predictors simultaneously. The logistic regression model is expressed as:
 

ln(p / 1-p) = β₀ + β₁X₁ + β₂X₂ + … + βkXk

 
Where,
ln(p/1-p) = Log-odds of adoption.
β₀ = Intercept.
βi = Regression coefficient for the ith predictor.
Xi = Value of the ith predictor.
p = Probability of adoption.
       
Regression coefficients were estimated using the maximum likelihood method and odds ratios (OR) with 95% confidence intervals (CI) were calculated to interpret effect magnitude and direction.
       
Model fit was evaluated using the Hosmer-Lemeshow goodness-of-fit test (χ2 = 7.43, df = 8, p = 0.49), indicating adequate fit and Nagelkerke R² = 0.71, suggesting that the model explains approximately 71% of the variance in adoption. Overall classification accuracy was 82.5%. Multicollinearity was examined using variance inflation factors (VIF), with all predictors yielding VIF values below 3.2, well below the conventional threshold of 10, indicating no problematic multicollinearity.

Adoption of ICT tools by the respondents
 
Table 1 shows the distribution of respondents according to their ICT adoption status. Of the 360 respondents, 195 (54.2%) were adopters and 165 (45.8%) were non-adopters, with a mean adoption score of 0.54 (SD = 0.49). The moderate adoption rate indicates that while ICT tools are gaining traction among farming communities in Nagaland, a substantial proportion have yet to integrate them into their practices. This level of adoption is broadly comparable to findings from other parts of North-East India, where moderate adoption is commonly reported owing to infrastructural and socio-economic constraints (Kath et al., 2022; Sharma et al., 2025).


 
Factors influencing adoption of ICT tools by the respondents
 
Table 2 presents the full logistic regression results, including coefficients, standard errors, Wald statistics, p-values, odds ratios and 95% CIs for all predictors in the model. The model demonstrated good overall fit (Hosmer-Lemeshow χ2 = 7.43, p = 0.49; Nagelkerke R² = 0.71; overall accuracy = 82.5%). Seven predictors were statistically significant: annual income, farming experience, source of information, knowledge, attitude, availability and education.


       
Annual income (B = 0.000, OR = 1.000, 95% CI: 1.000-1.000, p = 0.016) was significantly associated with ICT adoption. The OR of 1.000 reflects a scaling artifact: annual income was entered in absolute INR units and the effect per one-rupee increment is negligibly small. To illustrate the practical magnitude, a unit increase of INR 10,000 in annual income yields an estimated 10.5% increase in the odds of adoption (OR = 1.105), suggesting a cumulative effect across realistic income differences. This underscores that financially better-off farmers have greater capacity to purchase devices, pay for data services and sustain recurring ICT costs, consistent with findings from other agricultural ICT adoption studies (Mittal et al., 2010).
       
Farming experience (B = 0.089, OR = 1.093, 95% CI: 1.017-1.175, p = 0.015) showed a significant positive effect. More experienced farmers tend to develop better judgment about the practical utility of new tools and are often more motivated to adopt efficiency-enhancing technologies. This is consistent with the DOI framework, in which early adopters and early majority tend to be those with greater accumulated knowledge and stronger problem-solving orientation (Rogers, 2003; Aguilar et al., 2015).
       
Source of information (B = 3.150, OR = 23.34, 95% CI: 4.66-116.88, p<0.001) was a highly significant predictor. The large OR reflects the composite nature of the information source variable, which captures both the diversity and richness of channels accessed (extension agents, radio, peers, mobile advisories). Farmers with access to multiple, diverse sources are substantially more likely to encounter ICT tools, assess them positively and adopt them. Wide confidence intervals are acknowledged and are attributable to the moderately small sub-group sizes for higher information-exposure categories; findings should be interpreted with appropriate caution. This aligns with the DOI theory’s emphasis on communication channels as key drivers of innovation spread (Rogers, 2003; Lemma and Tesfaye, 2016; Odiaka, 2015).
       
Knowledge (B = 5.025, OR = 152.17, 95% CI: 19.51-1186.88, p<0.001) registered the strongest effect among all predictors. While the very large OR and wide CI warrant caution in interpretation, the direction and significance are consistent across model specifications. The large magnitude likely reflects the binary or near-binary nature of the knowledge scale in the sample farmers who crossed a threshold of functional ICT literacy were dramatically more likely to adopt. Potential quasi-separation between knowledge and adoption was investigated and ruled out; the variable was retained given its strong theoretical grounding in TAM (perceived ease of use and usefulness) and empirical support across the ICT adoption literature (Davis, 1989; Pandey, 2017). Digital literacy and functional understanding of ICT tools are arguably prerequisites for adoption, which explains the dominant effect size.
       
Attitude (B = 1.596, OR = 4.933, 95% CI: 2.244-10.847, p<0.001) was significant and consistent with TAM’s core premise that positive attitudes toward technology, rooted in perceived usefulness, drive adoption intention and behaviour (Davis, 1989; Diaz et al., 2021). Farmers with favourable attitudes toward ICT tools were nearly five times more likely to adopt them.
       
Availability (B = 2.874, OR = 17.71, 95% CI: 2.77-113.31, p = 0.002) significantly increased the odds of adoption. This finding reflects the UTAUT’s facilitating conditions construct: even motivated and knowledgeable farmers cannot adopt technologies without reliable network coverage, affordable devices and technical support (Venkatesh et al., 2003). In Nagaland’s hilly terrain, uneven mobile network coverage remains a structural barrier that individual farmers cannot independently overcome, highlighting the critical role of rural digital infrastructure investment.
       
Education showed a significant overall effect (Wald χ2 = 28.68, p<0.001). Compared to the illiterate reference category, those educated up to primary level showed OR = 2.14 (p = 0.09), middle school OR = 3.87 (p = 0.03), high school OR = 6.21 (p = 0.01) and graduate and above OR = 11.43 (p<0.001), indicating a dose-response relationship between education and adoption. This is consistent with the general finding that formal education enhances digital literacy, information processing capacity and confidence in using technology-based platforms (Nikam et al., 2021).
       
Non-significant predictors (p>0.05) included age, sex, secondary occupation, family type, family size, social participation, total landholding and accessibility. The non-significance of age is noteworthy: it suggests that ICT adoption in this context is not exclusively a phenomenon of younger farmers. Older farmers who are knowledgeable and have positive attitudes are equally likely to adopt, consistent with findings by Diaz et al., (2021) from Philippine bamboo farmers. The non-significance of sex suggests that gender per se is not a barrier in this sample, though this result may be influenced by sampling composition; gender-sensitive programming remains advisable given broader evidence on gender digital divides (Arvila et al., 2018). The non-significance of social participation and landholding suggests that adoption in Nagaland is more strongly driven by individual cognitive and infrastructural factors than by social or economic scale dimensions.
       
The findings of this study broadly align with TAM and DOI frameworks and with comparable empirical evidence from developing-country agricultural contexts. The dominant role of knowledge underscores that functional ICT literacy not merely physical access to devices is the most critical lever for driving adoption. This is consistent with Pandey (2017) and reinforces the position that supply-side interventions (providing devices and connectivity) must be accompanied by demand-side capacity building to be effective.
       
The large odds ratios observed for knowledge and source of information, while statistically significant and theoretically meaningful, should be interpreted with awareness of their wide confidence intervals. These wide intervals reflect sub-group sample size constraints inherent in a single-state cross-sectional survey and suggest that future studies with larger samples and multi-state designs are needed to obtain more precise estimates. The potential endogeneity between knowledge and adoption is acknowledged: farmers who adopt ICT tools may, in turn, develop greater knowledge through use. This bidirectional relationship is a recognized challenge in cross-sectional adoption studies (Davis, 1989) and longitudinal designs would be better suited to establish causal directionality.
       
The significant effect of information source diversity is consistent with Diaz et al., (2021), who reported that social influence and peer exposure were significant drivers of willingness to adopt mobile applications among bamboo farmers in the Philippines. Similarly, Arvila et al., (2018) demonstrated among Maasai farmers that socio-cultural norms and social network structures critically mediate technology adoption decisions, underscoring that adoption is a socially embedded process. The role of availability further echoes findings on facilitating conditions from the UTAUT literature, reinforcing that infrastructure investment is a necessary though not sufficient condition for adoption (Venkatesh et al., 2003).
       
Critically, Nagaland’s context adds several dimensions not fully captured by generic adoption frameworks. The state’s mountainous geography, tribal governance structures, strong community-based decision-making and historically limited formal extension infrastructure all shape the adoption environment in distinctive ways. Policies that leverage existing community institutions village councils (Gaon Bura system), self-help groups and farmers’ clubs as delivery channels for ICT training and information may prove more effective than top-down extension approaches. Localized digital content in Nagamese and tribal languages would further lower entry barriers for less-educated farmers.
 
Limitations
 
This study has several limitations that should be acknowledged. First, the cross-sectional design precludes causal inference; the observed associations between knowledge, attitude and adoption may reflect reverse causality or unmeasured confounding. Second, the six study districts were purposively selected, which may limit the generalizability of findings to all 17 districts of Nagaland. Third, very large odds ratios with wide confidence intervals for knowledge and source of information suggest potential instability in parameter estimates, likely due to moderate sub-group sizes; these estimates should be interpreted with caution. Fourth, self-reported adoption and knowledge scores are subject to social desirability bias. Future studies should employ longitudinal designs, probability-based district sampling and objective measurement of ICT knowledge and use to address these limitations.

This study assessed the extent and determinants of ICT adoption among farmers in Nagaland using binary logistic regression on a cross-sectional sample of 360 respondents. Fifty-four per cent of farmers were identified as adopters. Knowledge of ICT tools, diversity of information sources, availability of ICT infrastructure, positive attitude, farming experience, annual income and education were identified as significant positive predictors of adoption. Non-significant predictors included age, sex, landholding, family size and social participation.
       
The findings, interpreted through TAM and DOI frameworks, suggest that ICT adoption in Nagaland is primarily driven by cognitive and infrastructural factors rather than demographic characteristics. Policy implications include: investment in targeted digital literacy and ICT training programmes for farmers, delivered through community-based institutions and in local languages; accelerated rural digital infrastructure development, particularly mobile network expansion and affordable internet access; strengthening of agricultural extension services to serve as information multipliers and promotion of positive ICT attitudes through demonstration events, farmer field schools and peer-learning platforms. Future research should address the causal dimensions of knowledge-adoption relationships through longitudinal designs and explore the differential impact of specific ICT tool types on productivity and income outcomes.

The present study was supported by my supervisor, National fellowship for ST Students and Nagaland University for which I extend my heartfelt gratitude.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not necessarily represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but do not accept any liability for any direct or indirect losses resulting from the use of this content.
 
Informed consent
 
The study involved human respondents (farmers) and their informed consent was obtained prior to data collection. All ethical guidelines for research involving human participants were followed. Ethical approval for the study was obtained from the School of agricultural Sciences, Nagaland University prior to commencement of data collection.

The authors declare that there are no conflicts of interest regarding the publication of this article. No funding or sponsorship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.