Multiple linear regressions on determinants of ginger production in yeki district, Sheka Zone, South West Ethiopia

Ginger is an important crop that is produced worldwide for both spice and medicine. Ethiopia is one of the sub-Saharan countries which cultivate and export ginger to other countries. Even though ginger is an essential spice in the country, constraint faced during production reduces its output. Thus, this study aimed to identify the determinants of ginger production in the case of Yeki woreda. Primary data were collected using a self-administrative questionnaire administered on 110 ginger producers selected using a simple random sampling technique. Furthermore, data analysis methods descriptive statistics and multiple linear regressions were applied. The mean value of ginger yield was 51.74 quintals. The empirical result discovered that fertilizer use (p<0.001), bacteria disease (p<0.0340), education levels (p<0.0001, <0.0001, <0.0009, <0.0034), farm size (p<0.0025), farmer experience (p<0.0003), and weeds (p<0.0018) were signifi cant predictor variables in determining ginger production. Moreover, the result revealed that 85.00% of the discrepancy of ginger production explained by the independent variables included in the multiple linear regression model. Generally, fertilizer use, farm size in a hectare, the experience of farmers, weeds effect, diseases like bacteria wilt, and education level of farmers were the signifi cant factors of ginger output. Therefore, the study recommends the implementation and improvement of ginger production at the producer level by considering the use of fertilizer, farm size, herbicides, education level of farmers, and control bacterial wilt disease. Thus, developmental institutions, agricultural extensions, and governments are advisable to improve the yield of ginger production via controlling the signifi cant determinants. Research Article Multiple linear regressions on determinants of ginger production in yeki district, Sheka Zone, South West Ethiopia Alemu Bekele Eticha* Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi, Ethiopia Received: 14 August, 2020 Accepted: 31 August, 2020 Published: 02 September, 2020 *Corresponding author: Alemu Bekele Eticha, Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi, Ethiopia, E-mail:


Introduction
In developing countries like Ethiopia, the agricultural sector plays a vital role in economic growth and industrial development. The importance of agriculture in the economy of Ethiopia evidenced in percent of GDP shared in 2019 was 33.88%. Moreover, 90% of the poor gain their livelihood from the agriculture sector. However, Ethiopian agriculture is characterized by low productivity due to technical and socio-economic factors. Regularly the farmers with the same resources are producing different per hectare output because of several unknown determinants in developing countries like Ethiopia [1].
The largest ginger producer countries were India, China, Jamaica, Nigeria, Sierra Leone, Thailand, and Australia [2].
Ethiopia is also another moderate producer of ginger since the 13th century [3]. Ethiopia is the major supplier of ginger to India and international trade [4,5]. The major ginger growing areas in Ethiopia were the south nation nationality people's regional state and the Oromia state [6].
According to the Ministry of Agriculture and Rural Development, 99 % of the ginger production of Ethiopia was from the Southern Nations, Nationalities, and Peoples Regional State, while about 1 % was from the Oromia National Regional State [6]. The Southern Nations, Nationalities and Peoples Regional State of Ethiopia, is endowed with verity spices including "Kororima" (Aframomumkororima), Turmeric pepper (Piper nigrum), Cinnamon (Cinnamomum Verum), and ginger (Zingiber Offi cinale) more than any other regions of the country [4,7].
Ginger is the most important spice that has a vital role in the agricultural economy of the country. It contributes to raise the socio-economic status of the rural people and to earn foreign exchange currency and environmental protection.
Ginger is an incredibly important crop produced in different countries for both spice and medicine. Mostly ginger plant needs infertile sandy soil and a warmly humid warm climate [8]. The substance analysis of ginger shows that it contains over 400 different compounds that give many functions [9]. Ginger has an indispensable ingredient used for food processing throughout the world because of pleasant aroma biter taste and carminative property. Ginger has been used as species and medicine in countries like India, China, and England since ancient times. Ginger widely hunted for culinary purposes in gingerbread, biscuits, cakes, puddings, soups, and pickles. Also, it has been in the production of ginger oil and some alcoholic drinks like Ginger brandy, Ginger wine, Ginger beer, and Ginger ales. Anti-infl ammatory, anti-oxidant, antiplatelet, hypotensive, and hypolipidemic were among animal drugs made from ginger. Traditionally infections like high cholesterol levels, infl ammation, asthma, migraine, morning sickness, cancer, and vomiting in humans are treated by ginger ingredients [9][10][11][12][13]. Ginger is the most imperative spice crop in Ethiopia. Pregnant women use herbal remedies like garlic, ginger, and eucalyptus to treat pregnancy-related problems during pregnancy in Ethiopia [10].
Thus, Ethiopian ginger must compete with other countries in terms of quality, quantity, and price.
The farmers of Yeki district produce ginger traditionally did not change their living standards because of earning low yield of ginger due to many factors, of which ginger bacterial wilt disease was reported [6]. Despite this, there are so many factors that infl uence ginger production. The study from Tanzania that fi tted a regression analysis and point out that farmer's education level, the use of fertilizer, land size under ginger production and frequency of contacting extension services had a signifi cant contribution to ginger farming [2]. The study from Nigeria reported that age, income, seed, fertilizer, and agrochemicals were signifi cant variables [14].
Another study from Nigeria indicated that, if all farmers effi ciently use the available resources, the resulting increase in ginger output offset, increasing the farmers' income [15].
Thus, this fi nding has the main objective in identifying factors of ginger production in Yeki woreda by using multiple linear regression method.

Study area
The study takes place on farmers of Yeki Woreda, Sheka Based on geographical location, Latitude: 7° 14' 60.00" and Longitude: 35° 24' 59.99" E. It is situated at the altitude of 1200 meters above sea level representing a low land altitude and characterized by hot humid with an average annual rainfall of 1559 mm and a mean maximum and minimum temperature of 30.23 ºC and 16.09 ºC, respectively [15]. The main economic activities of the people in this area are producing crops and fruits such as maize, wheat, coffee, banana, inset, ginger, avocado, mango, papaya, sugarcane, tomatoes, khat, potato, turmeric, and cabbage. From these ginger and coffee are used as a cash crop; wheat, maize, inset, and others used for fi nancials as well as household consumption purposes. The farmers in the woreda produce ginger for fi nancial income. They supplied it to local merchants. The local merchants also buy the total quantity of output from farmers and further send it to those merchants that found in Addis Ababa ( Figure 1).

Study design and study population
A cross-sectional study design applied to ginger producer farmers who found in Yeki woreda. The study populations are all ginger farmers who produce it in 2018.

Sampling methods
A simple random sampling was in data collection within self-administers questionnaires on selected farmers in Yeki woreda. Accordingly, 110 ginger producer farmers were selected randomly in the district. Additionally, the secondary data on backgrounds gathered from Yeki Woreda agricultural offi ce.

The study variables
The dependent variable was the amount of ginger produced in the study area before June 2018. The independent variables were: sex, age, number of labor, work time, Family member, price in birr, Fertilizer applied in quintal, Farm size in a hectare, weeds, experience of farmer and diseases like bacteria Table 1.

Multiple linear regression
Linear regression analysis is a statistical method used to estimate the relationship between one or more independent variables and a single dependent variable. A Multiple Linear Regression method was applied to determine signifi cant factors from potential explanatory variables. The general form of a multiple linear regression model is given by: Where Y = Ginger produced in quintal;  o is the intercept and  1,  2…,  k are coeffi cients of the variable X 1 ,X 2,…, X k are independent variables and ε i error term [15].

Estimation of model parameters
The Maximum Likelihood was to estimate the parameters of the regression model. Analysis of variance used to test the general signifi cance of the model. The model was signifi cant when at least one predictor variable is signifi cant to the dependent variable.

Model adequacy checking
The model adequacy diagnostics checked via residual analysis or residual plots. These include: Linearity: Linearity indicates that the relationship between the dependent and independent variables should be linear in the parameter.
Normality: Normality of random error tested with a plot of residual against the cumulative probability or quantilequantile plot.
Homoscedasticity: Plotting the standardized residuals against time order is to examine the variance of the error term is constant.

Multi collinearity:
The decision on multicollinearity based on variance infl ection factor (VIF). If the value of VIF is less than ten, the collinearity was tolerable, but if it is more than 10, VIF is a risk.

Descriptive results
This section represents the results and characteristics of the farmers. Of the total of 110, 75(68.2%) of them were males and 35(31.8%) of the respondents were females.
Similarly, percentages 34.0%, 38.2%, 7.8%, 2.7%, and 17.3% were education levels of illiterates, primary, High school, preparatory school, and above the preparatory school. In the case of the use of fertilizer, 66.4% of farmers used fertilizer on ginger, and 36.4% of respondents not used fertilizer on ginger.
Also, 75.5 percent of farmers reported that ginger production was affected by the bacterial disease, while 88.2% stated that ginger was affected by a variety of weeds.
From Table 2, the average ginger produced in the area was

Results on regression analysis
The regression analysis result shows that the general regression model was signifi cant to study ginger as indicated by analysis of variance (Table 3)    The regression analysis on specifi c predictor variables was given in Table 4 and show that ginger production is infl uenced by farm size, fertilizer use, farmer experience, education level of the farmer, diseases like bacteria and weeds because each of them has p-value less than 0.05. In the next part the interpretations of each coeffi cient of the regression model. An intercept =16.35 was the average ginger produced while the model was not considered any variable. Land area is one of the most important and scarce in agricultural production. The farm size has a positive impact on ginger production. Thus, the land area coeffi cient = 8.78 is the change in the ginger yield in quintal for a one-hectare increase of land size by fi xed all other variables constant. It implies that when the farm area increased by one hectare, then the ginger yield is improved by 8.77quital.
The experience of the farmer coeffi cient = 0.4953 is the change in ginger production when all other factors constant. This indicates that when the experience of a farmer improved by one year, the ginger production is increased by 0.495 quintals.
The education level dummy is signifi cant to ginger production. The educational literature of the farmer is confi rmed that important feature that determines the readiness of the producer to accept new ideas and innovations. More educated farmers more expected to adopt new technologies to increase their ginger production. The coeffi cient of illiterate = 6.539 implies that the ginger production was decreased by 6.539 quintals when the farmer was illiterate than the above preparatory education level. Similarly, the coeffi cient of primary school = 8.735 indicates that the ginger production was decreased by 8.735 quintals when the farmer education level was primary school than that education level above the preparatory school when every other predictor variables were kept constant. In the same way, the coeffi cient of high school =6.429 points out that ginger production decreased by 6.429 quintals for a primary school than above preparatory school when all other variables keep constant. The estimator of preparatory school =7.541 indicates that the ginger production decreases by 7.51 quintal for the preparatory school education level than that of above the preparatory school when all other variables kept constant.
The coeffi cient=-2.879 indicates that the change in ginger production when the farmer report any bacteria disease. This coeffi cient implies that the ginger yield was averagely decreased by 2.879 quintals when the ginger was infl uenced by any bacterial and when all other variables kept constant. The estimator =-3.88 indicates that the change in ginger production when the ginger is affected by weed. Also, this means that the average ginger production decreased by 3.88 quintals compared with unaffected with weeds when all other variables kept constant. The coeffi cient of fertilizer =20.45039 was the change in ginger production when fertilizer used increased by one kg. The ginger production increased by 20.45039 quintals, compared with none fertilizer users, and when all other predictor variables were constant in the regression model. Lastly, all infl uenced factors are very signifi cant in the production of ginger in the study area.

Model diagnostics
The Normal Probability plot checks the relationship between ginger production and predictor variables. Figure   2, demonstrates that the p-p plot graph indicates that all observations are lie approach to the straight line fi tted and approximately confi rm linearity and normality postulate.  (Table 5).
with [21]. This means there was a strong positive relationship between farm size and ginger production. Even though different scholars reported different results for their study, this study identifi ed more variables in ginger production. The farmers also suggested farming all available farmland and sharing their experience on ginger farming. According to [2], the result from regression analysis showed that farmer's education level, the use of fertilizer, land size under ginger production, and frequency of contacting extension services had a signifi cant contribution to ginger farming. In line with this study confi rmed the education level of farmers, the use of fertilizer, and farm area of ginger were statistically signifi cant. It revealed that ginger yield no infl uenced by social-economic factors such as age and sex [2]. However, the experience of farmers on producing ginger, a bacterial disease of ginger, and weeds effect were the additionally identifi ed factors of ginger output under this paper. The study from Tepi agricultural research center reported that bacterial disease wilt was a problem on ginger production in the area [6]. The effect of weed was confi rmed offset, decreasing the yield of ginger [18,19]. Another scholar from Nigeria examined the socio-economic factors that affect ginger production (Zingier Offi cinale) farming technologies using multiple linear regressions point out that educational level and credit capital infl uenced the ginger production farming innovation [20]. In that fi nding, education is related to the ability to produce and how to improve the production of ginger [20]. In contrast an education was not factor according to the study from Nepal but it stated as it has positive correlation with production [21]. The study from Nigeria based on the sample size of 100 farmers, by using the stochastic frontier production function showed that fertilizer was a signifi cant determinant of farm output [19]. In the same way fertilizer was signifi cant variable ginger production while labor had no enough evidence to support.

Conclusion
Thus, fertilizer was confi rmed with others. Also farm size was also a very signifi cant variable in ginger production similar