3.1. Descriptive Statistics Results
The descriptive results are given in
Table 3.
The mean value for the TTMA question is 2.27, indicating that the average farmer in our sample belongs in the contemplation stage (TTMA = 2). In our sample of 200 farmers, 16% currently use a drone on their farm (Drone; TTMA = 4).
On average, farmers in our sample cultivated 672 ha of arable land, which is far above the Hungarian average of 20 ha [
45] of arable land (
Table 4). The average farmer in our sample is 53 years old, similar to the Hungarian average of 58 years [
45]. Concerning education, 49% of the farmers in our sample hold a university degree, which lies above the Hungarian average in agriculture of 3.4% [
46]. 6% of the participating farmers are female (Gender), which is significantly lower than the Hungarian average of 29% [
47]. The share of full-time farmers (55%) is slightly lower than the Hungarian average of 63% [
46]. Drone technology literacy reached 3.58 points on the 5-point Likert scale. Job relevance of drones for farm operations reached 2.71 points on the 5-point Likert scale. The attitude of confidence reached 4.03 points on the 5-point Likert scale. Since this scale is tautologically negatively generated, a decreasing value implies a higher measure of confidence. With respect to agricultural higher education, 45% of the farmers in our sample hold an agricultural university degree, which is higher than the Hungarian average of 9% [
47]. The reason for these differences lie in the fact that precision technology (especially drones) is typically used in large scale farms run by those with a higher education degree in agriculture.
Taking into account the data in
Figure 2, as well as expert opinions, we argue that the following levels adequately characterise the levels of application of precision technology elements in Hungary:
Basics of PF (43–94%)
At this level, we have taken into account the elements without which the application of PT (precision technology) is unthinkable. In addition to the availability of GPS, RTK and broadband internet, automatic steering machines have been included here because this is what justifies the purchase of RTK.
Data collection phase (external or internal sources, 25–32%)
Time series of at least three years for several crops and up-to-date information on remote sensing are essential for a reliable basis for decisions. Survey responses are unlikely to include precision tillage, which precedes the actual cultivation of the crop. The network of agricultural weather stations covers the whole country and provides farmers with real-time data, but they tend to use only the data that are most relevant to them. To optimise machine operation, data packages have been developed to facilitate this, which could play a vital role in disseminating relevant Green Deal standards (pesticides, fuel saving). Another limitation of the application is the availability of appropriate technology for measuring yields of only the most important arable crops.
Use of database to extract information (5–23%)
Smart data generation and its professional use is the next level of technology and a value creator. Precision crop protection is also increasingly linked to other technological elements (e.g., coupled nutrient supply). Although split fertilisation is widely used in conventional technologies (e.g., spots where the machines turn, no sowing on ditch banks—we believe this explains a large proportion of the positive responses), it is significantly different from differential nutrient application in precision technology (e.g., differential fertilisation of saline soil patches), which is a significantly higher and less frequently used technological element. The measurement of soil moisture is important information available in the database of all soil weather stations. Still, practical experience shows that only a minority of farmers use that information, which also requires the use of specialised experts for its development.
The future (4–17%)
This category includes only drones, apart from robotics. The use of drones is already of considerable importance (17% of the data received for this use) in the field (especially in the differentiated treatment of soil and plant patches with different conditions and infestations). However, their use in imaging is still negligible. Drone technology capable of performing both functions has already emerged at the experimental level, so we consider the inclusion of drones in this category to be justified. In Hungary, aerial crop protection, for which drones can play an vital role, is experiencing a revival, since, in addition to the precision treatment of infested patches, they are also able to work at night, when wind conditions are more favourable than during the day.
The responses of Hungarian farmers using precision technology provide a good basis for categorising precision technology elements. Still, due to gaps in interpretation, the results obtained are over-represented for drones and split fertilization, and under-represented for remote sensing systems.
3.2. Results of the Ordinal Logistic Regression
The ordinal logit model helps explain the relationship between the outlined perceptions regarding the technical barriers to drone adoption and the attitudes towards the use of drones. The model also draws attention to on-farm characteristics that authenticate potential early adopters. The ordinal dependent variable reflects the adoption stages regarding the intention to use drones, namely precontemplation, contemplation, preparation and action. The results of the ordinal logistic regression are provided in
Table 5. This table contains the OR, the standard errors, the significance levels and the 95% confidence intervals. A likelihood ratio test was significant (LR χ
2 (9) = 175.93;
p < 0.001), indicating that one or more coefficients significantly differ from zero. The log-likelihood value is −346.59. Other model fit criteria imply quite an acceptable model fit with McFadden Pseudo-R
2 and Nagelkerke Pseudo-R
2 (0.34–0.63) and a significant chi-squared value (
p < 0.001).
Predicted probabilities and marginal effects for each category of the TTMA variable are given in
Table 6. Predicted probabilities show that half of the farmers of the sample belong with 43% probability in the contemplation stage (TTMA = 2). There is a change in the sign for all variables between the contemplation stage (TTMA = 2) and the preparation stage (TTMA = 3) of the model. These results are similar to Michels et al. (2020) [
23]. This indicates that variables with a statistically significant effect make a difference between farmers with no or only overall interest in drones and farmers with concrete plans to use or who already use a drone. The result is a unique Hungarian case study on TTMA, which can be complemented by research in several technological, socio-economic and other country-specific contexts.
Table 5 shows that the highest exponential beta (Odds Ratio) was observed for the variable of higher agricultural education. It can be concluded that someone with higher education in agriculture is 6.33 times more likely to (intend to) use a drone than someone without such education. In second place was the question “the use of drones is important for my job” (OR: 3.67). In this case, the exponential beta means that if someone rated their answer to this question one category higher (agreed more with the question), they were 3.67 times more likely to have an intention to use drones. Among German farmers, Paustian–Theuvsen (2016) [
29] could not detect a significant effect of higher education (OR = 1.28,
p = 0.46). When we interpreted education uniformly as higher education, similar to the German method, we obtained the same result (OR = 6.33,
p = 0.024) as Michels et al. (2020) [
23] (OR = 0.45,
p = 0.326). However, when we separately examined the effect of higher education in agriculture on openness to drone use, we found the positive effect reported earlier. According to our results, those with tertiary education in agriculture were 6.33 times more likely to decide to use a drone than those without such education. According to Paustian–Theuvsen (2016) [
29], the explanation for this finding could be that agricultural education has an impact on openness to precision farming, and we were able to demonstrate this in the case of Hungarian farmers.
The same was observed (but to a lesser extent) for full-time farmers, as the exponential beta (Odds Ratio) was 3.34. This implied that farmers who produce full-time were much more likely to use a drone on their farm than those who only produce part-time, i.e., the odds of using a drone increased by a factor of 3.34 for full-time farmers. For Hungarian farmers, we find similar correlations to those formulated for German farmers by Michels et al. (2020) [
23].
Significant differences were also found for the question “learning to use drones is not a problem for me”, where the exponential beta (Odds ratio) was 1.54, and for the question “I don’t think I would use drones because using them seems too complicated for me”. For the latter, the exponential beta (Odds Ratio) was 1.47. In these cases, the exponential betas mean that if someone rated their answers to the questions one category higher (agreed more with the questions), those farmers were 1.54 and 1.47 times more likely to want to use drones.
This area of study is crucial because, as an Australian study (Higgins et al., 2017.) [
48] pointed out, even though farmers adopt technology learned through their partners, this technology adoption is often accompanied by negative emotions. Also, the presence or absence of advisors may be a factor affecting diffusion [
33].
The effect of age was the opposite of the variables presented so far, as the exponential beta was less than 1 (Odds Ratio: 0.97), meaning that an increase in age decreases the probability of using a drone. Our results (OR = 0.97,
p = 0.029) are similar to those of large-scale farmers in Germany (OR = 0.97,
p = 0.06 in Michels et al., 2020 [
23]), showing that openness to adopting new technologies, in this case drones, decreases with age. This result is in line with findings in the international literature on the adoption of innovative technologies [
19,
37]), which may be due to the fact that older farmers may have less experience with digital technologies (smartphone, computer), a shorter time horizon available to them, and a tendency to stick to habits [
20], as well as a greater reliance on their practical experience with new technologies [
29].
Among German farmers, Michels et al., 2020 [
23] were able to show a clear relationship based on gender (OR = 4.18,
p < 0.01) for openness to drone use, i.e., male farmers are 4.18 times more likely to accept the use of drones than female farmers. Due to the low number of respondents in the sample, we were not able to detect any association with women among Hungarian farmers. The German results are also in line with those of Zhang et al. (2019) [
25]. In the context of German farmers, the authors note that their results are also noteworthy because the European Commission (2017) data show a steady increase in the proportion of farms headed by female farmers [
23]. We could not show any relevant gender differences among the Hungarian farmers, because the proportion of women among the precision farmers was very low.