• For Contributors +
• Journal Search +
Journal Search Engine
ISSN : 1225-8504(Print)
ISSN : 2287-8165(Online)
Journal of the Korean Society of International Agricultue Vol.31 No.3 pp.214-225
DOI : https://doi.org/10.12719/KSIA.2019.31.3.214

# Factors Influencing Consumer Preference for Rice in the North West Region of Cameroon

Bobah Brillant Sisang**, Sang-man Kim*, Jin Lee*, Yu-cong Sun*, Jong-In Lee*†
*Department of International Cooperation Major in Global Agricultural Economics, Kangwon National University, South Korea
**Department of Agricultural and Resource Economics, Kangwon National University, South Korea
Corresponding author (Phone) +82-33-250-8668 (E-mail) leejongin@kangwon.ac.kr
April 29, 2019 September 18, 2019 September 23, 2019

## Abstract

Cameroon is a rice-importing country due to rapid population increase and urbanization growth. As agriculture becomes more market oriented so too are consumers preference changing. Therefore, this study specifically examines the factors influencing consumers preference for rice in Cameroon with a case study of Mezam and Ngoketunjia Divisions of the North West Regions. With the use of well-structured questionnaires, 300 rice consumers were sampled across the major market areas of the divisions. The multinomial logit model was used to assess why households in the study area consume the different categories of rice in the market, while factor analysis and linear regression were used to identify the rice quality or market attributes that had an influence on consumers decisions. The study revealed that rice consumers were in three categories, those who ate only local rice, and those who ate only imported rice, and those who ate both combinations of local and imported rice. The major socioeconomic factors that significantly influences consumers preference were income level, educational level and working status of rice household consumers. The rice quality or market related attributes with significant influence on consumers preference were taste, aroma, texture, and degree of whiteness of the grains after milling. Market attributes like origin and availability of rice were also significant factors, which influences consumer preference. Grouping these factors into two groups identifies the production or rice quality factors to have a general significant influence on rice consumption in the region. There are therefore needs to improve on the production and post-harvest chain. Those will greatly improve the quality and competitiveness of local rice in the Region.

# 카메룬 북서부 지역의 쌀에 대한 소비자 선호도 영향 요인

보 하 브릴란트 시상**, 김 상만*, 이 진*, 손 우총*, 이 종인*†
*강원대학교 대학원 국제협력학과
**강원대학교 농업자원경제학과

## INTRODUCTION

Rice is the fastest growing staple food source in most African countries, providing the bulk of dietary energy to the growing population. It is positioned as the 5th most prominent source of energy in diet responsible for about 9 percent of caloric intake (FAOSTAT, 2012). In Cameroon, rice has become a major stable crop in recent years although the crop has been cultivated only for several years. Its consumption in various forms has drastically increased and has been gradually taking over from traditional crops, mainly roots, tubers, and cereals like maize and millet. It has become a staple food for both rural and urban dwellers unlike it was the case a few decades ago when rice was considered to be eaten only on special ceremonial days or frequently consumed by mostly the rich or urban dwellers.

There continue to exist an increasing demand for rice as most consumer’s preferences have changed over the past years. According to the Coalition for the Development of Rice in Africa (CARD) 2017/2018 rice sector outlook, Cameroon national consumption volume stood at 1,040 metric tons of rice, of which local productions accounted for just 183,000 tons while the wide gap was filled by imported rice accounting for 857,000 tons. The widening gap between national consumption volume and national production volume can be attributed to both production constraints and the rapid change in consumer’s behavior and preference towards rice. Industrialization, urbanization, population growth, income growth and accordingly change in consumer preferences have over the past few decades greatly increased the demand for rice in most of the Sub-Saharan Africa (SSA) countries (WARDA, 2008). As income rises in low income countries, this translates into high demand however, as income grows, consumers' tastes and preferences change.

Consumers determine the survival of new crop varieties in any given reference area by choosing whether to consume them and in what proportion (Dalton, 2004). Thus the successes of research on rice over the years has largely focused on new varietal releases, production and cultivation. Little attention has focused on demand-driven or market research by investigating consumer preferences and willingness to pay for locally produced rice. Any new rice varieties will largely depends on whether it is accepted and consume in significant quantities by the target population. Local rice producers are struggling with issues of market access for local rice and this can be answered through producing rice varieties, which meet consumer needs and preferences (Adu-Kwarteng et al., 2003). Consumers’ preference and acceptances govern the traders’ acceptance and during cultivation farmers consider the preference pattern of these two groups. Knowing the rice qualities acceptable to the market will enable researchers to develop rice cultivar with acceptable end-user attributes. Until now most rice producers sell all the rice they produced because demand exceeds supply. However, liberalized markets have raised the need to understand what the market wants so that what is produced will be sold. Hence, focusing as well on consumer’s preferences becomes very important in the rice sub-sector. It’s in the view of this that it has become essential to study and understand consumer preference for rice in the market.

The issues discussed above raise a number of research questions especially concerning consumer preference for rice in the North West (NW) Region. The following pertinent research questions guided this study: (1) What are the most important attributes that consumers consider when purchasing rice? (2) What are consumers’ perceptions towards local rice? (3) What are the roles played by consumers’ socio-demographic preference for each of the attributes?

Preferences determine the consumers choice of a product as reflected by their willingness to pay for the particular product. Even if the current situation has revealed that rice demand exceeds local supply hence local producers can depend on the domestic market, paying attention to rice attributes is still important for improving the competitiveness of local rice varieties in the liberalized market as well as for future rice market (local and export). As income level increase and the market for rice expands, more people will be willing to pay premium prices for specific rice attributes. Moreover, knowledge of consumer preferences for rice attributes enables actors such as farmers, researchers, processors, and traders to design appropriate strategies for incorporating or retaining such attributes during breeding, production, processing and marketing of rice. It is therefore within the above-mentioned context that this study seeks to identify the preference considered by consumers when they make decision to purchase rice in the North West Region of Cameroon. The general objective of the study was to identify the determinants of consumer preference for rice in the NW markets: (1) Establish the socioeconomic factors that determines consumer preference for rice available in the markets. (2) Identify the qualitative and marketing factors that influences consumer preference for rice.

## MODEL AND DATA

### 1. Data collection:

This study used a cross sectional research design whereby primary data was collected with the used of structured questionnaires administered to respondents at one point in time. 300 rice consumers across the North West Region were randomly sampled within 2 Divisions of the 7 Divisions that make up the region as seen on figure 1. Two Division was selected, Mezam Division, as the most populated and cosmopolitans’ division and home to the regional headquarters. The second division was the Ngoketunjia Division with a unique characteristics of a major rice production basin in the region and country.

In the Mezam Division, the sampling areas/markets were the Commercial Avenue, the Nkwen market and the Bambili market area. Meanwhile for the Ngoketunjia Division, the Ndop Central Sub Division market area was selected.

One section of the questionnaire was designed to gather information on the socioeconomic characteristics or variables of the respondents including gender, age, marital status, household size, income level, educational level, household expenditure for food and rice. This information was included to gain insights on the distinctive characteristics of the consumers that influences their rice purchasing behavior. The second section covered information on consumer purchasing behavior for rice. This section looked at consumption frequency, quantity, and forms of rice consumption preferred by the consumers. The section equally had variables of rice attributes for preference such taste, aroma, packaging, price, grain size, stickiness, degree of whiteness, cleanliness, origin, % of broken grains, availability, food safety. The last section of the questionnaire focused on the knowledge, awareness and patronage of the different existing rice varieties or brands present in the NW markets by consumers. The sample of 300 consumers randomly selected in this survey was distributed as shown in table 1.

### 2. Description of study area

This study was carried out in the North West region of Cameroon specifically in two out of the seven Divisions that make up the Region as shown on Figure 1 below. The Region is the third most populated Region with more than 2 million inhabitants. It is believed to be one of the largest market for both domestic and local rice in the country as a result of its host to a large production basin in the country.

Mezam divisions as one of the principal study site is host to the regional headquarters, the Bamenda cosmopolitan city and it is the most populated town in the Region. It is located on Latitude 5°56‟0”N and Longitude 10°10‟0” E coordinate (WGS84). With about 600,000 inhabitants, it covers an area of 1,745 km2.

The consumers surveyed came from the market areas across the town spanning from the commercial Avenue market to Nkwen market area and then to the Bambili market area. These areas all have a significant rice marketing outlets such as supermarkets, shops and community market space from which most consumers do buy both local and imported rice brands.

The division lies between latitude 5º 15N and 6º 10N and longitude 10º 15E and 10º 40E (DDARDN, 2013) and is house to about 200,000 inhabitants across an area of 1,126 km2. Also known as the home of ‘Ndop Rice’, the Division is host to the Upper Nun Valley Development Authority (UNVDA), a major state corporation established in 1972 by the state to promote and empower small scale farmers especially rice farmers who made up a significant portion of the population. It has three Sub-Divisions, the Ndop Central, Babessi and Balikumbat Sub Divisions, across where rice is produced. The consumers surveyed in this study were within the Ndop Central Sub-Division, the Bamunka main market area. This market area has a wide brand of both local and imported rice brands from which consumers choose what to consume.

### 3. Theoretical Models

The data collected was analyzed and processed with the use of SPSS V20. Descriptive statistics and empirical results gotten.

The tools used for analyzing the study data were the descriptive and the multinomial logit analysis to analyze the socioeconomic factors that influences rice consumption (objective 1). Factor analysis and linear regression was used to identify the factors of rice attributes that affects consumers’ consumption preference (objective 2). The socioeconomic characteristics of the respondents were analyzed using descriptive statistics and results were presented using the measures of central tendency. The multinomial logit tool was used to examine those socioeconomic factors that influence consumers’ preference for the imported rice only, local rice only, or a combination of both the imported and local rice.

#### 1) The Multinomial Logit Model

The multinomial logit model was used to assess why households in the study area consume the different categories of rice in the market. The model was chosen based on survey data which revealed that household rice consumption (dependent variable) was found to be a categorical variable which can take three categories or levels. These categories were assigned numbers 0, 1 and 2. 0 was used to indicate the combination for (local and imported rice) consumer, 1 for those who consume only imported rice and 2 was used to indicate the local rice consumers’ group. The local rice consumer group was taken as the reference group. The multinomial logit model was therefore used to identify the variables that make households belong to categories 0 for local and imported rice consumer group; 1 for imported rice consumer group and 2 for the local rice consumer group.

The probability that the ith household belongs to the jth rice consumer group Pij reduces to:

$P i j = e β j x i ∑ k = j e β j x i$
(1)

The model makes the choice of probabilities on individual characteristics of agents. Following Maddala (1983) and Babcock et al., (1995), the basic model is written as;

$P i j = e β j x i ∑ k = 0 e β j x i$
(2)

Where: i = 1, 2,..., n variables;

k = 0, 1,..., j groups and,

βj = vector of parameters that relates Xj to the probability of being in group j;

where there are j+1 groups.

In this survey, the Xi variables range from X1 – X8. For the socioeconomic /demographic characteristics of the sampled rice consumers, represented as below:

X1=gender, X2=Age, X3=Marital status, X4=Household size, X5=Educational Level, X6=Working status, X7= Employer, X8=Income level.

#### 2) Normalization of the model

As required, the summation of the probability for the three categorical groups in the model must equal to unity. This calls for normalization of the equation model. The common rule is to set one of the parameters vectors equal to zero (Kimhi, 1994). Hence, for k number of choices only k–1 distinct parameter are identified and estimated.

Based on Equation (2), the probability of being in the reference group: the local rice consumer group with parameter vectors equal zero is

(3)

Similarly, the probability of being in each of the other j groups is

(4)

Dividing equation (3) by (4) gives

$P i j P i o = e β j x i$
(5)

This denotes the relative probability of each group to the probability of the reference group. Hence, the estimated coefficients for each group reflect the effect Xi has on the likelihood of the consumer’s household belonging to that alternative group relative to the reference group. The logarithm of the odd ratio in the equation to base e gives the estimating equation.

#### 3) Factor Analysis and Linear regression to obtain objective 2

Factor analysis is a multivariate technique in which, most commonly employed factor analytic procedures in marketing applications are principal and common factor analysis. The major objective to employ this analysis is to determine the variables which influence the consumer’s buying decision for different rice brands and varieties.

Principal component analysis can accommodate a large number of variables and reduce the information to a convenient size. The inter-relationship among a set of many interrelated variables are examined and represented in terms of a few underlying factors or dimensions that explains the correlation among a set of variables. This assumes that the observed variables are linear combinations of some underlying source variables, which are known as factors.

The factor analysis program uses the correlation matrix as input to identify interrelations between variables. Using those correlations one can see what information and hypotheses can be obtained. Factor loadings provide the correlation between the variable and the underlying dimension. The product of corresponding factor loadings can obtain the correlation between any two variables.

Since the objective of the factor analysis is to represent each of the variables as linear combination of the smaller set of factors, we can express this as

X1 = λ11F1 + λ12F2 +…+ λ1mFm + e1;

X2 = λ21F1 + λ22F2 +…+ λ2mFm + e2;

...

Xn = λn1F1 + λn2F2 +…+ λnmFm + en;

where X1 to Xn = Standardized scores;

F1 - Fn = Standardized factor scores;

e1- en = Error variance.

The maximum number of factors possible is equal to the number of variables.

However, small number of factors by themselves may be sufficient for retaining most of the information on the original variables.

Using the factor scores in linear regression gives the equation as below:

Consumer Preference = a+b1FS1 +b2FS2+…. + bkFSk+e, where, a = regression constant;

b1, ..., bk = regression coefficients;

FS1, ..., FSk = factor scores;

e = random error term.

## ESTIMATION AND RESULTS

It covers the results gotten form the 300 consumers surveyed in this research and analyzed as stated in the previous chapter. Results from the socioeconomic and demographic characteristics are presented in the first section while the results from the rice attributes that determines or guides consumer preference is presented in the second section and then the other import variables surveyed explained in the last part.

### 1. Descriptive statistics

From the surveyed results, table 2 indicates the descriptive statistics of the 300 consumers. It was described though three parts: demographic characteristics of respondents, socioeconomic characteristics of consumers, consumer rice purchasing and consumption behavior.

The part of demographic characteristics of consumers indicates that of the 300 consumers, 65% of them were women while 35% were male. This indicates that most of the respondents who bought rice were women from the surveyed market areas. This observation was good for the findings given that most food consumption decisions are made by woman. For the sampled age distribution, it is observed that 41% of those surveyed were within the age range of 21-30 years as the most populated age class. The age range from 31-40 had 21.7% of the respondents, the range from 41-50 had 15.0%, the 51-60 years had 7.3%. Meanwhile these last categories had the least from the surveyed population above 70 years of age, 2.7%. A proportion of 51.0% of those surveyed were married while 37.7% of them were still single and those who had lost their wife or husband constituted just 11.3%. Of the five ranges of household composition surveyed, the highest household size range was that of 2-5 members with 43% while those with above 10 family members made up 5.3%.

From the surveyed result of socioeconomic characteristics of consumers, it was presented that most of the respondents had some university studies amounting to 45.7% of the sampled population and just 1.3% had not been to school or received any form of official education. The overall high educational level of the respondents in this survey suggest that they could read and write. This characteristic is important as literate consumers can read rice packaging labels or other useful information which can influence their choice of a desired product. The working status of the respondents were as follows, 23.3% said they worked full time, 30.7% were studying at various levels. Those who had part time jobs constituted 27.3%. Those who were retired made up 7.7% of the population while 11.0% of the respondents were jobless. Government employees within the sample was 15.3%. Most of the respondents were self-employed accounting to 31% of the sampled population. The private sector employed 25.0% of the workers while 28.7% said they were involved in other sectors not mentioned. Within the monthly income level ranges surveyed, a good number of the respondents fell within the 1,000-50,000 range amounting to 56.7% of the population. 23.3% fell within the 50,000-100,000 income range, while the last category of above 300,000 incomes is 0.7%.

This surveyed of consumer rice purchasing and consumption behavior revealed that a majority of the households ate rice 2 to 4 times every week, representing 61.7% of the sampled population, those who ate once a week constituted 32.3%, those who ate once a month made up 2.3% and those who said they ate rice every day were 3.7%. The purchasing behavior of the sampled respondents varied. The largest group purchase rice once a month representing 46%. This group was followed by those who bought once a week representing 23.0% of the sampled population. The least groups were those who bought rice once in 3 months and once in 4 months, representing 4.3% and 3.7% respectively. For quantity purchased on monthly bases, 32.3% of the respondents buy 25 kg of rice monthly. This group is the least those who buy above 50kg made up the least group representing 8.0% of the consumers. The finding indicated that 41.3% of the consumers bought rice from local small stores or shops around their neighborhood. This group was followed by those who said they often bought rice directly from local rice millers or retailers, representing 25.3% of the population. Those who bought from local markets made up 22.0% and the least group was those who bought from supermarkets representing 11.3% of the consumers. There just 13.0% of the consumers ate only imported rice. Those who eat just local only represented 33.7% while those who consumed a combination of both local and imported rice represented 53.3% of the sampled population.

### 2. Empirical Results for objective one

As earlier explained, multinomial logit regression model was used in this study to identify the socioeconomic or demographic factor that could influence the choice consumers make when making rice purchasing or consuming decisions.

The variables chosen were as follows;

Dependent variable: Y=Choice Consumed, (in three non-ordered categories).

Independent variables: X1=Gender, X2=Age, X3=Marital status, X4=Household size, X5=Educational Level, X6=Working status, X7=Employer, X8=Income Level.

#### 1) Durbin-Watson test

In statistics, the Durbin–Watson statistic is a test statistic used to detect the presence of autocorrelation at lag 1 in the residuals (prediction errors) from a egression analysis. The value of DW is 0<D<4, with statistical significance as follows: When residuals and independent variables are independent of each other, the closer D=2 or DW is closer to 2, the greater the assurance of non-autocorrelation; When the residuals of two adjacent points are positively correlated, D<2 and the value of DW is closer to 0, there is the stronger the positive autocorrelation; when the residuals of two adjacent points are negatively correlated, D>2 and the value of DW is closer to 4, there is a stronger the negative autocorrelation.

As table showed that, the value of Durbin-Watson is 1.723 closer to 2 which has the greater assurance of nonautocorrelation. 3

#### 2) Multi-collinearity test

The desired variables were subjected to a multi-collinearity test to be observe if any collinearity existed with the socioeconomic variables and if the data set was suitable for the multinomial logit regression model chosen.

The tolerance value and the Variance Inflation Factor (VIF) values were of the selected variables subjected for analysis were all within an acceptable range. The reciprocal of tolerance, the larger VIF, the more severe collinearity. The empirical method shows that when 0<VIF<10, there is no multi-collinearity. When 10≤VIF<100, strong multi-collinearity exists. When VIF ≥ 100, there is severe multi-collinearity. As shown on table 4, the VIF values were of the selected variables subjected for analysis were shown that there were no multi-collinearity.

#### 3) The multinomial logit regression model

The results of table 5 indicates that data set was significant to measure for the socioeconomic factors that had an influence of the choice of rice consumers. The likelihood ratio test for the model was 84.716, which was significant at 5% confidence interval. This clearly showed that the rice consumers groups were heterogeneous. The P-value being P<0.000, was an indication that the model fits or better still that at least one of the coefficient in the model was not equal to zero.

As shown on table 6 below, the data subjected to analysis identified that the socioeconomic factors that signifi-cantly (at 1%) influence the decision rice consumers took were their Educational Level (P-value of .002). The various working status of the consumers with a P value of .041 and the income levels of the consumers, with a P-value of .010. This results suggest that the how educated the consumers were influence their decisions to purchase or consume either the imported rice or local rice or both. This survey as earlier presented showed that 96% of the respondents had some level of education with just 4% not haven’t received any formal education. Therefore, an educated consumer can identify the type or quality of rice he wishes to purchase or consume from its packaging labels and this could influence their preference on which type of rice to buy.

The working status influences the decision consumers made towards rice consumption. This means or will suggest that the diverse economic activities of the respondents played a significant role on the choice of what kind of rice to purchase or consume. This is further explained by the income level equally having a significant influence of the consumption of rice in the study area. Different working status accords different income levels to respondents and their income significantly affects their consumption decisions. Income level has a positive impact on rice consumption.

These significant results showed in table 6 agree with the findings of Kassali et al., (2010) and Ogundele.O (2014) who found that among other socioeconomic characteristics, income level, and the educational level of rice consumers had significant influence on household food rice consumption in Ghana and Nigeria respectively.

On the other hand, the results indicated that socioeconomic characteristics such as, the gender of the consumer, the household size and the employer of the consumers did not significantly (P>0.05) have an influence on the decisions taken by the various households were consuming rice.

### 3. Empirical Results for objective two

Table 7 indicates that the Kaiser-Meyer-Olkin measure of Adequacy for the data set in this analysis model was 0.649, which attested that the data set was suitable to be subjected to such a model for data reduction. The Bartlett’s Test of Sphericity was significant (P<.000) for the data extracted.

As presented on table 8, the various variables extraction through the principal component analysis method gave the following extraction loading for each variable.

The extraction factor loading shows that 6 factors were loaded with an initial Eigenvalues above 1. These 6 factors explained 62.23% of the total variance that existed among the data set variables. With the purpose of factor analysis being to obtained a number of reduced components or factors. These 6 factors could therefore represent the various factors or groups under which the various rice attributes or characteristics that consumers considered when purchasing or consuming rice falls. The attributes or characteristics are Availability, Origin, Taste, Whiteness, Price, Swell ability, Ease of cooking, % broken grains, Cleanliness, Stickiness, Packaging, Food safety, Aroma, Grainsize, Texture which are represented by from component 1 to component 15.

After the extraction or identification of the preferential factors in the first step, the next step was to subject the factors to rotated component matrix analysis. This was done with the goal of obtaining the covariance and correlations of the rotated factor loadings. The results obtained is presented on table 9. Values below 0.50 were omitted from the loading and just those above 0.50 were retained to identify the loading under each factor.

#### 2) Using the identified factor scores in linear regression

After obtaining the factor scores from the factor loading, these factor scores were saved as variables and analyzed with linear regression to identify the factors or attributes, which had a significant influence on the decision of rice consumers. This is better illustrated as below for the linear regression analysis of the factor scores.

Dependent variable: Consumer Preference

Independent Variables: Factor score 1, Factor score 2, Factor score 3, Factor score 4, Factor score 5, Factor score 6.

Using the factor scores in linear regression was significant at 5% attesting that the variables were fit to be analyzed using the model. Table 10 presented that this model tested significant (P<.001) at 5% confidence interval therefore it was appropriate to use for the analysis of the factor scores.

The dependent variable was the rice consumers’ preference while the 6 extracted factors were the independent variables. Recalling that these factors were extracted as table 11 stated.

As shown on table 12, After analysis, Factor 1, Factor 5 and Factor 6 were statistically significant at 5% (With Pvalues as stated below) having an influence on the rice consumption preference (Factor 1 at P<.000, Factor 5 at P<.027, Factor 6 at P<.028).

This implies the rice consumption preference in the study area is significantly influence primarily by the taste of the local or imported rice brand, the origin of the rice, the availability of a given rice brand or variety, and the degree of whiteness of the grains after milling or polishing. This is resonated by the work of Gideon et al., (2014) and the findings of Denis et al., (2017) who revealed that the rice quality attributes as taste, grain size and whiteness and availability of local rice had a strong significant influence on the choice of consumers in Ghana.

Consumers in the North West Region especially within the study area are well exposed to different varieties or brands owing to the historic local rice nature of the Region serving as the second largest basin of rice production in the country.

The second significant factor is the texture of rice after cooking. Consumers are believed to be strongly attached to the texture of rice after cooking and hence justifies the influence that the attribute has on their preference or choice.

The last significant group of attributes is factor 5, which had the grain size and Aroma of rice. Consumers in the study have a long preference for long grains of rice and thus are sensitive to another other rice grain size present in the market.

On the other hand, attributes such as % of broken grains after milling, the cleanliness of rice from foreign particles, the stickiness of the grains after cooking, the price, swelling ability and ease of cooking were found not to significantly have an influence on the preference of rice consumers on their choice of rice brands or varieties purchased or consumed.

The above observations can be explained by the fact that as rice becomes a staple food in the area coupled with its historic rice production aspects, the consumers seem not to be influenced by the price, broken grains percentage, cleanliness, and ease of cooking nor swelling ability of the different rice brands or varieties.

## CONCLUSIONS

This research was focused on identifying the factors that influences consumers’ preference for rice in the Mezam and Ngoketunjia Divisions of the North West Region of Cameroon. The survey randomly sampled 300 rice consumers from the zone and through a structured questionnaire, their consumption behavior was surveyed and the data collected and analyzed.

From the survey, the following major findings were revealed: (1) A majority of the rice consumers sampled were women and mostly of the youthful age group. This can be justified by the key role women play in making household food consumption decision and dynamism of the youthful population of the study areas. These consumers had some basic educational background and perhaps could explained why the socioeconomic variable such as educational level had a significant influence on their rice consumption decision making. These factors had both a negative and positive influence on the preference of consumers depending on the category of rice preferred by the consumer. (2) The diverse economic activities of the consumers and their different income levels all had a significant influence on their choice of rice eaten, frequency and brand of the desired product consumed. This indicates that there could be a correlation between both variables even though that wasn’t the focused of the analysis. As the degree of urbanization increases in the study area, consumption preference changes, more rice consumers’ response or consumption behavior varies. (3) There exist numerous attributes of rice from the different chains of production, processing or marketing, that can influence the rice market and consumer decisions. However, this study has revealed that consumers in the North West region take in consideration attributes from two different categories. Firstly, the production category, where rice quality attributes like taste, aroma, grain size and texture were revealed to have a significant influence on consumers’ preference. Taste of the rice depending on their attachments to particular taste. The influence of these quality attributes can be clearly understood as the study areas are known to have a strong historical and cultural attachments to locally produced rice. The second category is the processing and market related attributes such as the origin of the rice, the availability of the rice and degree of whiteness of grains after processing. Those who prefer just local rice pay attention to the source of the rice before purchase. Availability of desired brands in the market or local market outlets equally influences the frequency, quantity and brand consumed.

At a time when the national demands for rice in Cameroon is on the rise, as the ‘golden grains’ slowly but surely replaces the traditional cereals and tubers that once used to be the staple food for most households, Consumers are becoming more market oriented responsive and hence the necessity of re-orienting agriculture (the rice sector) to be market oriented to better meet the needs of consumers. The Cameroon rice sector if well exploited could boost economic growth and enhance rice self-sufficiency and combat hunger and malnutrition.

The rice market has become very competitive and more complex fueled by market liberalization and the local rice sector in Cameroon especially the North West Region, will need to readjust from production to marketing strategies to meet consumers’ preference or demands. Therefore, understanding the driving factors of consumers’ preference is very important to improve and promote the quality and marketability of local rice in the region. Hence, understanding the determinants or driving factors that guides consumers’ choice and behavior is very important to improve on the quality and marketing characteristics of rice. An understanding of the production, processing or marketing related attributes of importance to consumers is helpful to promote the production, marketability and competitiveness of local rice in the North West.

There are some limitations of the research. The first limitation is that it was carried out just in two of the seven Divisions that make of the Region. As such, it will be useful in the future to do a balance representative study to include the other Division of the Regions like the Menchum and others who too are a large basin of the rice production map of the region. The second limitation is that the work deals on stated preference in its questionnaire design, it could be more comprehensive to do a choice experiment model that will the permit the consumers to make choices based on already constituted groups of factors or attributes base on their desired utility maximization from each of such bundles.

Even if the results identifies that the impact of taste origin, availability, degree of whiteness, aroma, grain size and texture are proven to be significant on the decisions of consumers, it does not mean that they are the only attributes that influence consumers preference. However, this result is based on the model chosen for this research, multinomial logit regression and factor analysis based on the choice of the stated preference design of the work. Doing a conjoint analysis in a choice experiment model in future studies will be useful to equally identify the choice combinations of consumers in the study area.

## 적 요

카메룬은 인구증가와 도시화가 진전됨에 따라 쌀을 수입에 의존하고 있는 국가이다. 농업이 시장중심으로 전환되면서 소 비자들의 선호 역시 달라지고 있다. 본 연구에서는 카메룬 북 서부의 Mezam과 Ngo-ketunjia 지역 소비자의 쌀 선호도에 영향을 미치는 요인을 조사하였다.

특히 카메룬에서 소비자 선호에 영향을 주는 요인들에 대해 Mezam와 Ngoketunjia Divisions of the North West Regions 라는 사례연구와 함께 살펴보았으며 구조화된 설문지를 통해 주요 시장 지역들에 걸쳐 300명의 쌀 소비자들을 대상으로 조 사하였다. 사회 경제적 요소들, 유도된 소비자 행동이나 선호 를 확인하기 위해서 다항식 기호 논리학 회귀 분석을 사용하 였고 소비자 선택에 영향을 미치는 쌀의 품질이나 시장 속성 을 확인하기 위해 선형회귀분석을 사용하였다. 조사 분석 결 과 소비자들은 오직 지역 쌀만 먹는 부류, 오직 수입된 쌀만 먹는 부류, 지역 쌀과 수입쌀을 모두 먹는 부류의 3가지 구매 형태를 나타냈다. 쌀 선호도에 영향을 미치는 주요 사회경제적 요소로는 소비자의 소득, 교육, 직업 수준인 것으로 나타났다.

소비자의 쌀 선호에 영향을 미치는 품질 요소는 맛, 향, 질 감, 도정 후 낟알 색의 흰 정도였고, 시장 속성에 관해 소비 자의 쌀 선호에 영향을 미치는 요소는 쌀의 출처와 구입 가능 성이었다. 이러한 요소들을 2가지 차원으로 나누어 볼 때, 생 산성과 쌀 품질에 관한 차원은 카메룬의 쌀 소비에 있어서 중 요한 영향을 미친다고 확인 할 수 있었다.

## Figure

Map of the selected study sites

## Table

Sample size per site selected

Descriptive statistics of the sample households

Summary model fitting

Coefficients of socioeconomic variables

Model Fitting Information

Likelihood Ratio Tests of socioeconomic variables

KMO and Bartlett's Test

Factor identification and extraction

Rotated Component Matrix of extracted factors

ANOVA test of the model

Extracted factors and variables

Coefficients of factor scores after linear regression analysis

## Reference

1. Adu-Kwarteng, E. , Ellis, W.O. , Oduro, I. and Manful, J.T. 2003. Rice grain quality: a comparison of local varieties with new varieties under study in Ghana. Food Control 14:507-514.
2. Babcock, B.A. , Chaherli, N.M. , and Lakshminarayan, P.G. 1995. Programme participation and farm-level adoption ofconservation tillage: Estimates from a multinomial logit model. Working Paper 95-WP136, Centre for Agricultural and Rural Development, Iowa State University, Ames, Iowa.
3. Dalton, T.J. 2004. A household hedonic model of rice traits: economic values from farmers in West Africa, Agricultural Economics, 31:149-159.
4. Dennis, Sedem .2017. Households’ Preference for Local Rice in the Upper East Region, Ghana. Hindawi Advances in Agric.1812975, p.9.
5. FAO.2017. Rice Market Monitor, Trade and Markets Divisions of Food and Agriculture Organization of the United Nation, 20(1).
6. Danso-Abbeam, G. , M. Armed, and F. Baidon.2014. Determinantsof consumer preference for local rice in Tamale Metropolis, Ghana. International Journal of Education and Sosial Science, 1(2):114-122.
7. Kassali, R. , Kareem, R.O. , Oluwasola, O. and Ohaegbulam, O.M. 2010. Analysis of Demand for Rice in Lle Lfe, Osun State, Nigeria. Journal of Sustainable Development in Africa, 12(2):63-78.
8. Kimhi, A. 1994. Participation of Farm Owners in Farm and Offfarm Work including the Option of Full-Time Off-farm Work. J. Agric. Econ., 45(2):232-239.
9. Madadala GS. 1983. Limited-Dependent and Qualitative Variables in Econometrics, New York: Cambridge University Press, pp.401.
10. Ogundele.O. 2014. Factors influencing Consumers’ Preference for Local Rice in Nigeria. African Journal of Marketing Management. 6(4):49-55.
11. WARDA.2008. Africa Rice Center (WARDA), FAO and SAA (2008) NERICA® : the New Rice for Africa—A Compendium. In: Somado EA, Guei RG, Keya SO (eds). Africa Rice Center, Cotonou, Benin.