Contributing factors of maize production using multiple linear regressions in mizan-aman district, bench-shako zone, Southwest of Ethiopia

Agriculture is the backbone of the Ethiopian economy. Maize is one of the plants which mostly used for indigenous food like bread, injera, qolo, nifi ro and soup in Ethiopia. Despite farmers of Mizan-Aman produce maize, they challenged in a low amount of the production that comes from unknown causes. Therefore, this study was conducted to access the production of smallholder farmers and to identify factors that infl uence production of maize. The study based on the information which collected from 105 randomly selected farmers, using self-administrative questionnaire. The multiple linear regressions were used and identifi ed factors of maize production. In addition, data diagnostics analysis was used to determine the goodness of the regression model on the maize yield. According to the result, the mean production of maize in study was 33.72 quintal per hectare. Additionally, the regression result revealed that productions of maize were signifi cantly infl uenced by the education level of producers, fertilizer use, and variety maize seeds that used since each of them have p-value 0.04, 0.000 and 0.042 respectively. Thus, the study suggested, Mizan-Aman developmental agricultural workers must help the maize producer farmers on education level, fertilizer use and seed of maize to improve their production. Furthermore, education level of the farmers will be improved by give training on farming and farming related technologies. Research Article Contributing factors of maize production using multiple linear regressions in mizanaman district, bench-shako zone, Southwest of Ethiopia Alemu Bekele Eticha* Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi, Ethiopia Received: 10 October, 2020 Accepted: 03 November, 2020 Published: 05 November, 2020 *Corresponding author: Alemu Bekele Eticha, Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi, Ethiopia, Email:


Introduction
The agriculture sector plays an important role in economic growth of developing countries. Agriculture is considered as a backbone of the Ethiopian economy. Agriculture contributes about 34 percent of the GDP in 2019 and 90 percent of the population is dependent directly or indirectly on agriculture in Ethiopia [1]. The crop maize (Zea mays) originated in Mexico, 7000 years ago from a wild grass, and Native Americans transformed maize into a better source of food [2]. Recently, maize is grown throughout the world, United States, China, and Brazil being the top three maize-producing countries in the world [2]. In addition, maize is the major common diet over 85 percent of the population in Kenya [3]. It is believed that the maize was comes from Kenya to Ethiopia. In the recent history, maize has emerged as a leading cereal crop in Ethiopia [4]. Ethiopia is the second highest maize producer in Sub-Saharan Africa next to Nigeria [5,6] Maize production of Ethiopia increased from 939 thousand tonnes in 1970 to 8500 thousand tonnes in 2019, growing at annual rate 7.64% [7].
Similarly, the Maize production increased from 8350 thousand tonnes in 2018 to 8500 thousand tonnes in 2019 with growing annul rate 1.8% [7]. It has been reported that a lot of households now grow maize than any other cereal in Ethiopia [8]. Maize is an agricultural product in most of the Ethiopia regions like the Amhara, the Oromia and the south nation and nation and national states.
Maize can be processed into a variety of food and industrial products, including starch, sweeteners, oil, beverages, glue, Citation: Eticha AB (2020) Contributing factors of maize production using multiple linear regressions in mizan-aman district, bench-shako zone, Southwest of Ethiopia. Open J Plant Sci 5(1): 040-045. DOI: https://dx.doi.org/10.17352/ojps.000023 industrial alcohol, and fuel ethanol. Approximately 40% of the maize production in the United States is used for fuel ethanol [12]. As the ethanol industry absorbs a larger share of the maize crop, higher prices for maize will intensify demand competition and could affect maize prices for animal and human consumption. Low production costs, along with the high consumption of maize and micronutrient defi ciencies are common public health problems, make this staple food an ideal food vehicle for reinforcement [2,13].
Different studies clear out that variety of seeds and resource management practices to achieve higher yield, income and minimize risks [8]. The identifi ed factors that affect maize production were education of farmer [14], fertilizer use [4], land-labor ratio, use of fertilizer, use of pesticide, manure and household size [15], age of farmer [16] and farm size [17].
However in this study aimed to identify more maize production determinants. Provision of agricultural technology might not assure its success in reducing the maize yield gap unless successfully used by maize growers [5].
It is known that Ethiopia has suffi cient input resources of agriculture but the majority of the populations are still under severe risk of poverty. Thus the country still imports different cereal crops for food purpose. Then the best optimal solution for the country may be to identify determinants of the production and productivity in the country. The cropping system in former Bench-Maji and Bench-Shako zone is predominantly undertaken by the use of traditional farming system. The farmers believed that large quantity of yield is obtained expanding land seize and the farming is characterized by the low production per hector [14,16]. Consequently, the annual income from farming, especially maize is too low which is not suffi cient for households to change their live standard. The study in Mizan-Aman only identifi ed a single signifi cant factor education level of the farmers while education is multilateral [14]. However, this study aimed to identify more factors that infl uence maize yield in Mizan-Aman. Accordingly, variety of seed was added to the predetermine variables in this study. The multiple linear regressions method applied to identify several independent variables these depend on maize production, since multiple linear regressions were used to identify more independent variables [17,18].

Study area
This study conducted in Mizan-Aman district farmers which are found in Bench-Shako zone, south nation and national people's regional state, south west Ethiopia. It is located 562 km away from Addis Ababa. The major economic activities of the people in this area are producing crops such as maize (Zea Mays), wheat, coffee, banana, Enset /false banana (Ensete ventricosum) and ginger (Zingiber offi cinale). From these wheat and maize are mostly used for food consumption, while coffee and ginger are for cash. Mizan-Aman district has well suitable climate and rich in natural resources [14]. The maize produced in the area used for food consumption and supplied to local merchants.

Data collection
The primary data was directly collected from randomly selected farmers by administrative questionnaire. In general, out of total 589 farmers, 105 sample farmers were selected in the study area. The sample selected by simple random technique because it is simple and honest method.

Study variable
The interest variable of the fi nding was yield of maize production in quintal per hectar in Mizan-Aman district. The predictor variables were: education status of farmer, age of farmer, working hours per day, fertilizer use, farm size (in hectare), seed type, soil fertility, weeding effect, and use of irrigation.

Multiple linear regressions
Multiple linear regression model (MLRM) is a statistical techniques used to identify the linear relationship of dependent and independent variable [18]. The general linear model for MLRM in which response is related to a set of independent variable ( ) Xi is given:

Model adequacy checking
The model checked whether the typical assumptions of multiple linear regression method are fulfi lled. Analysis of residual is an effective way to discover several type of model adequacy. Also, normal probability plot is used to test linearity and normality assumptions. Thus, multicollinearity, linearity, normality and others are tested in the study. Test of presence of multicollinarity refers to the existence of high (perfect) linear relationship among regressors. Therefore the variance infl ation factor (VIF) was applied in this study. The general rule of thumb indicates multicollinarity problem for corresponding variable if the value of VIF greater than 10.

Descriptive of socio-demographic
The descriptive statistics of socio-demographic factors displayed in Table 1

Multiple linear regression analysis
Overall regression model analysis: The analysis of variance result displayed on Table 3 revealed that the overall regression model is signifi cant since p-value of regression model is 0.000 which is less than signifi cance value  value =0.05. Therefore, at least one of the independent variable has a signifi cant impact on the linear regression of maize production. Thus, the general multiple linear regression models are statistically signifi cant for the maize production data since it has p-value=0.000. Table   4, the estimated parameters for multiple linear regression models were displayed. It can be observed that education level was statistically signifi cant effect on the production of maize (p-value=0.048) at 95% confi dence interval. In addition modifi ed seeds use was statistically signifi cant effect to the production of maize (p-value=0.00) and fertilizer use on maize production was statistically signifi cant to production of maize

Model adequacy diagnostics
The model goodness of test is proved with R square. Hence, the value of R 2 = 0.6895 shows that 68.95% of the variability of the maize yield is explained by regressors included in the model.

Mullticolinearity:
If VIF statistic greater than 10 it is severe to muticollinerity. Accordingly, each of the independent variable has variance infl ation factor (VIF) value less than 10, which indicates that there is no co-linearity severe multicollinearity (see Table 5).     Homoscedasticity: Homoscedasticity checked using plot between the standard predictor values and standard residuals.
As displayed using Figure 3, the scatter plot of standardized predicted residual versus the response variable we can see that the residual shows there is no systematic random pattern. It indicates that the error term has constant variance. Therefore, it satisfi es constant variance assumption approximately.

Discussion
The main objective of this study is to identify the infl uential factors of maize production in Mizan-Aman district, Southwest Ethiopia, since the improvement of maize production must start at the farmer level. A Multiple Linear Regression is used to analysis the maize data because the maize yield per hectare is quantitative variable and there are several independent variables. Moreover, the assumption in multiple linear regressions was satisfi ed with the data [17]. The simple linear regression method is applied on maize production in Bench-Maji zone [14].
The multiple linear regression identifi ed factors of maize production in Sheka zone. The results of that study revealed that fertilizer, farm size and labor were signifi cant variables [17]. Similarly, in this study fertilizer and education level were major determinants of maize production in Mizan-    Aman district, southwest Ethiopia. However, seed effect was identifi ed in this study only.
The fertilizer had a signifi cant effect on maize plant height, lodging percentage and grain yield [19]. Similarly, several scholars suggested that fertilizer has vital role in corn production [4,15,17]. In line with other scholars this paper confi rmed the importance of fertilizer on maize production in Mizan-Aman district. The yield of maize and fertilizer has strong positive association since it increases the fertility of the soil. According to CSA report, about 23% of fertilizer used for maize production. Infrastructures such as transportation like road and fi nancial constraints were reported related to commercial fertilizer use problem on maize farm [4].
Traditionally, Ethiopian farmers have used organic fertilizers for agricultural production [8].
Additionally, the education level of the farmers is another important factor of maize production. In line, the education issue was reported by several scholars [14,20]. The education level was also reported in coffee production factor in Mizan-Aman district [9]. It is suggested that education is used to selected adopted technologies [20,21].
Lastly, seeds are another factor of maize production which identifi ed in this study result. The improved seeds used for maize production signifi cantly increased. Similar to this result, it suggested that many farm households are aware of the complementary nature of improved maize seed and chemical [21]. The common varieties of seed are local seed and hybrid seed. Lastly, if the farmers used modifi ed seed, improved their knowledge on farming and apply either biological or chemical fertilizer, they improve their crop production including maize.

Conclusions
This study revealed that, education level, fertilizers and seeds are factors that infl uence the grain yield of maize. At the producer level, the important determinants which can improve maize production are use of modifi ed maize seed, use of fertilizer, and education level of farmers. It is usual that, identifi cation of the important determinants of maize may not improve the maize production unless the producers incorporate it in their productivity. Thus the agricultural sectors in Mizan-Aman, and developmental agricultural workers in each district should help the farmers regarding identifi ed factors.
Additionally, the farmers have to incorporate all identifi ed factors to improve the maize production.