INVESTIGATING FACTORS AFFECTING FOOD SECURITY IN HEIPANG

COMMUNITY OF BARKIN LADI, PLATEAU STATE

Alade, Segun Peter1, Pegoelong, Mathew David2 and Kwol, Victoria Stephen2
1
Department of Statistics, Plateau State Polytechnic, Barkin Ladi
2
Department of Hospitality Management, Plateau State Polytechnic, Barkin Ladi

Abstract
This paper, an investigation into factors affecting food security in Heipang is aimed at
discriminating the community into those who are secured food wise and those who are not,
and to also determine the factors that are significant in the discrimination. The paper
considered some of the factors that FAO suggested in its definition of what it means to be
secured food wise. Some of these factors are; cultural food, accessibility to food,
affordability, nutritional value, variety, household size and adequacy in terms of quantity.
The data for the study was collected via questionnaire which was issued out to 100 household
heads in the community and a discriminant analysis was carried out using SPSS version 25.
The result showed that among the three predetermined groups, which are those who rarely
have food shortage, those have food shortage and those who often have food shortage showed
that there is no distinction among those who have food shortage and those who often have
food shortage, but there is a distinction between these two groups and those who rarely have
food shortage. The percentage of those who rarely have shortages is 69.6 while those with
food shortages is 30.4, this clearly shows that quite a large number of the community is
secured food wise, but this is in terms of quantity and not the quality of food consumed. The
factors that discriminated among the groups are same meal (variety), affordability, balance
diet (nutritional value) and household size. It is recommended that for there to be
sustainability food wise in the community, efforts must be geared towards creating variety of
food, that is both nutritious and affordable for the people.
Key words: Discriminant, food security, sustainability, food affordability, food accessibility

1
Introduction
Food security according to FAO means access by all people at all times to enough food for an
active, healthy life [CITATION Hon03 \l 1033 ] . Food is an important need for human life,
because it gives the human body energy and nutrients which help in growth and development
of the body. Any human body that lacks adequate and quality food would be able to add little
or no value to life and human endeavour. Food which is one of the basic needs of human
according to Abraham Maslow hierarchy of needs must be met before a man can actually
move to the next level of need.

Over the years, the world over Nigeria inclusive, there has been striving for man to be able to
provide food for its ever-increasing population, and various scholars have come up with
different measure to determine what it means for an individual or household to be satisfied
food wise. Most of these measures are tailored to find out whether there is adequate and
quality food available for the teeming population or not. Adequate measurement of food
security would help analyst, service providers and policy makers assess and assist in making
right decisions that would make food available. The problem of food and nutrition security in
Nigeria has not been adequately and critically analyzed despite various approaches at
addressing the challenges. Government has introduced several projects and programmes to
improve agriculture and boost food production in the country. However, the empirical records
of many of these programmes and projects are not impressive enough to bring about the
expected transformation of the sector (Ihimodu 2004). Today, the problem continues to exist
at an increasing pace as more than 900 million people around the world are still malnourished
(FAO 2010). According to Adebiyi (2012) Nigeria remains a net importing nation, spending
about N1.3 billion on importing of basic food items annually. The food security problem in
Nigeria is pathetic as more than 70 percent of populace live in households too poor to have
regular access to the food that they need for healthy and productive living with an increasing
high levels of malnutrition and poverty (Babatunde, Omotesho and Sholotan, 2007).

Food security relates directly to nutrition and health. Typically, food security is thought of as
being related to availability and access of foodstuffs. Yet, the threat to food security also lies
with urbanisation, income disparity, overpopulation, ecosystem degradation, animal health,
and food wholesomeness. Food wholesomeness is also an important aspect of nutrition.
Wholesomeness is monitored via food safety and food defense programmes, making them
critical components of a food security programme as well (Havas and Salman, 2011). There
are various variables that can be used to determine if an individual is secured food wise or

2
not. Some of these variables are availability of food at all times, means of access, its
nutritional value, adequacy in terms of quantity, quality and variety, and also its acceptability
within the given culture (FAO 2010). There are many factors in today’s global environment
that aggravate food security. It is true, we live in an age where we are growing and producing
more food than ever before. There is enough food to feed the world’s population, but it is not
distributed properly nor is all food culturally appropriate across the globe. Local food access
differs dramatically and the greatest difference exists between developed and developing
countries (FAO, 2019). The primary reason for this lack of justice is an income-related
difference between these populations (Hazell and Wood, 2008). It must be stated though, that
in every country of the world there is hunger, and this often falls along economic and social
lines. The underprivileged – be it individuals or countries – often have less. (Havas and
Salman, 2011). Other factors that affects food security are overpopulation, climate change
and urbanization. Areas of the world with the highest birth rates and population, where
demand often exceeds supply, also have the greatest levels of hunger and disease. Local
ecosystems provide the resources a population needs for food production, health,
environmental management and water. Examples include rangeland, fertile soil, nutrient
cycling, and wildlife for hunting, among others. The local ecosystem has a certain carrying
capacity, and once this is exceeded the ecosystem becomes stressed and begins to break down
(Ericksen, 2008). The result includes over-farmed soils, exposed and eroded grazing lands
and dried up or contaminated wells, all of which contribute to an underfed population.
Numerous are the factors that affects food security but this paper measures food security in
the Heipang metropolis of Barkin Nigeria based on the variables defined FAO.
Abu and Soom (2016) in a study carried out in Benue State Nigeria reported that majority of
the rural households (53.3%) and urban (62.2%) households were able to meet the
recommended calorie intake of 2500kcal per capita per day. And also that household heads
with 6-10 persons per household were majority accounting for 40% in rural areas and 46.7%
in urban areas. The result implies that families with small household size are more food
secure than those with large household size. This is because increase in members of
household added more responsibilities to household heads especially when many of the
family members depend totally on the household head. Again, as household size increases,
income per head declines and the less food secure the household becomes.

A similar study carried out in Kenya where households were classified into three groups of
food security status as poor, borderline and acceptable. This classification was done using

3
food consumption scores profile, these profiles are amount of milk, meat, Fish, vegetables,
sugar, oil and the likes that a household consumes within a specific period of time (Onoja,
Babasola and Ojiambo, 2018). This paper on the other hand looked at the variables in a
different light because rural dwellers in Nigeria do not have a standard for measurement of
the amount of some certain food consumed, so it would not be an efficient task to carry out
the study as it was done in the Kenya study. Therefore, the paper considered to use the
variables that made up the composition of the definition of food security by FAO, and these
variables are income, quality of food, adequacy and acceptability of food, availability of
food, access to the food and variety.

Materials and Methods

Discriminant Analysis
Discriminant analysis is a statistical technique is useful in the investigation of various aspects
of a multi-variate research problem. The multiphasic character of discriminant analysis is (a)
the establishment of significant group-differences, (b) the study and 'explanation' of these
differences, and finally (c) the utilization of multivariate information from the samples
studied in classifying a future individual known to belong to one of the groups represented.
Essentially these same three problems related to discriminatory analysis. Originally
developed in 1936 by R.A. Fisher, Discriminant Analysis is a classic method of classification
that has stood the test of time. Discriminant analysis often produces models whose accuracy
approaches (and occasionally exceeds) more complex modern methods. Discriminant
analysis can be used only for classification (i.e., with a categorical target variable), not for
regression. The target variable may have two or more categorical data. The objective of a
discriminant analysis is to classify objects, by a set of independent variables, into one of two
or more mutually exclusive and exhaustive categories (Johnson and Wichern, 2007).
Given a set of p independent variables, the technique attempt to derive a linear combination
of these variables which best separate or discriminates the groups. The functions are
generated from a sample of cases for which group membership is known; the functions can
then be applied to new cases with measurements for the predictor variables, but unknown
group membership.
The procedure automatically chooses a first function that will separate the groups as much as
possible, it then chooses the second function that is both uncorrelated with the first function
and provides as much further separation as possible. The procedure continues adding
functions in this way until reaching the maximum number of functions as determined by the

4
number of predictors and groups in the dependent variable. In two group discriminant
function, there is only one discriminant function. The discriminant score obtained from the
discriminant function is used to classify the dependent variable into one of the two or more
groups (Balogun, Akingbade and Oguntunde, 2015)
For classificatory discriminant analysis, Fisher’s Discriminant function is generally used.
Fisher’s idea was to transform the multivariate x to univariate observations y such that the y
’s derived from the populations were separated as much as possible. Fisher’s approach
assumes that the population are normal and also assumes the population covariance matrices
are equal because a pooled estimate of common covariance matrix is used.
A fixed linear combination of the x’s takes the values y 11 , y 12 , . .. , y 1 n for the observations
1

from the first population and the values y 21 , y 22 ,. . . , y 2 n for the observations from the second
2

population and so on. The separation of these sets of univariate y’s is assessed in terms of the
differences between the yexpressed in standard deviation units. That is,
| ý 1− ý 2|
separation= ,
sy
n1 n2

where
∑ ( y 1 j− ý 1 ) +∑ ( y 2 j− ý 2 )2
2

W = j=1 j=1
n 1+ n2−2
is the pooled estimate of the variance. The objective is to select the linear combination of the
x to achieve maximum separation of the sample means ý i. This result in the linear
' '
combination y= I^ x= ( x́1 −x́2 ) W pooled x which maximizes the ratio
−1

2
( Squared distance between sample mean of y ) ( ý 1− ý 2 )
=
( Sample variance of y ) S 2y
2
( ^I ' x́ 1− ^I ' x́ 2)
=
I^ ' S❑pooled I^ '
2 ' −1
The maximum of the above ratio is D =( x́ 1− x́ 2 ) W pooled ( x́ 1−x́ 2 ) ,the Mahalanobis distance .
If we assume the populations are multivariate normal with a common covariance matrix, then

n1 +n2− p−1 n1 n2
a test of H o: µ1¿ µ 2 versus H 1: µ1≠ µ2 is accomplished by referring
( n1+ n2−2 ) p n 1+ n2
D (
2
)
to an F-distribution with v1¿ p and v2¿ n1+n 2− p−1 degrees of freedom. If H o is rejected we
conclude the separation between the two population is significant (Johnson and Wichern,
2007).

5
The discriminant function is a weighted average of the values of the independent variables.
The weights are selected so that the resulting weighted average separates the observations
into the groups. High values of the average come from one group, low values of the average
come from another group. The problem reduces to one of finding the weights which, when
applied to the data, best discriminate among groups according to some criterion. The solution
reduces to finding the eigenvectors, V , of S−1 S S
W , S A . Where W and A are the sum of squares for

within groups and error respectively. The canonical coefficients are the elements of these
eigenvectors.

A goodness-of-fit parameter, Wilks’ lambda Λ, is given as follows (Todorov, and Filzmosor,

2007) :

|S W| m 1
Λ= =∏
|S A| j=1 1+ λ j

where λ j is the jth eigenvalue corresponding to the eigenvector described above and m is the
minimum of K−1 and p , K is the number of groups and pis the number of variables measured
on each observation .
The canonical correlation between the jth discriminant function and the independent
variables is given by:

λj
r cj =
√ 1+ λ j

The overall covariance matrix, T , is given by:

T= ( N 1−1 ) S T

The within-group covariance matrix, W , is given by:

1
W= ( N−K )S W

The among-group (or between-group) covariance matrix, A, is given by:

1
A= ( K−1 )S A

The linear discriminant functions are given by:

6
y= I^ ' x= ( x́1 −x́2 )' W −1
pooled x

The standardized canonical coefficients are given by:

vij √ wij

where vij are the elements of V and w ij are the elements of W .

The correlations between the independent variables and the canonical variates are given by:

p
1
Corr jk = ∑v w
√ w jj i=1 ik ji
For the purpose of this work, the discriminant analysis was used to classify the households of
Heipang into groups of those who often experience food shortage, those who rarely have
shortages and those who experience acute shortages of food. This classification was done on
the basis of the factors such as size of household, household income, accessibility to food,
affordability of food, quality of food (balanced diet) and nature of food (cultural).
Study Design
The study was conducted in Heipang District of Barkin-Ladi Plateau State. The community,
also known as Haifam is located on the geographical coordinates latitude 9.38 o North and
longitude 8.53o East. It is a type P populated place with most members of the community
being peasant farmers (www.googleweblight.com).
Two hundred (100) households were selected and questionnaire issued to the household head
or its representative. The test item issued to them contains questions on factors that determine
the security of a household food wise. Since most of the community members are Berom by
tribe, an interpreter was also used to ensure the that the test items were well understood and
appropriate response supplied. is mostly Berom.
Data obtained was analysed using descriptive statistics and also Discriminant Analysis with
the aid of SPSS version 25. The discriminant analysis was used to first determine if the pre-
determined populations were actually distinct and also to determine the variables that
distinguish them.
Results and Discussions
A total of 100 questionnaires were issued out to the head of households in the community. A
total of 92 responses were obtained, this represents 92% of the total questionnaires
administered. The predetermined groups for the analysis were, those who often experience

7
food shortage, those who very often experience shortage and those who rarely experience
food shortage.

Seven variables were entered to be able to discriminate between these groups, they are
Household size, Food balance, Cultural food, Food affordability, Food accessibility, Family
income and same meal.

Table 1: Test of Equality of means

Wilks' F P-value
Lambda
What is the size of your household? 0.895 5.215 0.007
How balanced is the food you consume in your household? 0.917 4.023 0.021
The food you consume are they cultural food? 0.784 12.251 0.000
How affordable is the food you consume? 0.952 2.251 0.111
How accessible is your reach to the food you consume? 0.986 0.640 0.530
Does your income adequate to cover the food consumption in 0.988 0.519 0.597
your household?
Is there a tendency to consume the same meal three times a day? 0.899 4.982 0.009

Table 1 shows test of equality of means, this test is usually assessed using the Wilks Lambda,
the statistic takes on values between 0 and 1, if it is 0 it means the variable completely
discriminate, but if it is 1, it does not discriminate at all. Here it is discovered that the most
discriminating variable between the two groups is the consumption of cultural food (0.784),
followed by household size (0.895), then tendency to consume same meal every day, and
lastly balanced nature of the food consumed. Variables that are not significant in
discriminating among the groups are affordability of the food, accessibility to the food and
worst in the line is income. The variables not significant, implies that these variables do not
discriminate between the three groups significantly. In order words they cannot be used to
differentiate among the groups. This does not mean to say they have no effect on food
security of a household but rather they do not discriminate the groups. For instance, about
71.1% of the 64 households who said they rarely have food shortages said that their income is
inadequate, which implies that income affects all categories of grouping. Income is not a
discriminating factor between the groups when it comes to Food shortages but rather incomes
looks more of a general problem affecting the entire population whether you have enough
food or not. Income, accessibility and affordability does not discriminate among the groups
but variety and quality of meal consumed does, this is also opined by Roos, Ruthven,
Lombard and McLachian (2013). Since income affects food security in the population under

8
study then one would not be too surprise that affordability also does, and since most of the
people living in the community are Berom, they have virtually equal accessibility to the food
they consume.

Table 2: Box's Test of Equality of Covariance Matrices

Box's M 52.093
F Approx. 1.435
p-value .066

Table 2 shows the result of the Box’s M test, the essence of Box’s M is test for equality of the
covariance matrix, for discriminant analysis to be carried out, the covariance matrices must
be equal. The p-value>0.05, which means there is no significant difference in the covariance
matrices and so thus satisfies one of the conditions for the use discriminant analysis model.

Table 3: Eigen values and Wilk’s Lambda

Function Eigen % of Canonical Wilks' Chi- p-value
value Variance Correlation Lambda square
1 0.763 92.7 0.658 0.535 53.775 0.000
2 0.060 7.3 0.238 0.943 5.024 0.541

Since there are three categories of grouping in the study, two discriminant functions would be
created, Table 3 shows the efficacy of the two functions created. Function one has an eigen
value of 0.763 which accounts for 92.7% of the total system variability and it has a lower
Wilk’s lambda value of 0.535 compared to function two. Also the chi-square test is
significant, all this implies that function one carries most of the information concerning the
system than function two and it does better than chance at separating the groups compared to
function two whose chi-square test is not significant. It means in this analysis we can ignore
function two.

Table 4: Standardized Canonical Discriminant Function Coefficients

Function
Variables
1 2
What is the size of your household? 0.538 0.266
How balanced is the food you consume in your household? .0488 0.695
The food you consume are they cultural food? -0.667 0.895
How affordable is the food you consume? -0.347 -0.583
How accessible is your reach to the food you consume? -0.297 0.305
How adequate is your income to cover the food consumption in your household? -0.053 0.037
Is there a tendency to consume the same meal three times a day? 0.552 0.008

9
Table 4 shows the coefficients of the two standardized canonical discriminant functions. We
would concentrate on function 1 since it had earlier been shown that function two is not
significant.The discriminant function D is given by :

D=0.538∗Household ¿ Food Balance−0.347∗Food Affordability +0.552∗Same meal−0.667∗Cultural food

It can be observed from the table that cultural food, household size, consumption of same
meal everyday and balanced food among others have the largest absolute value, the larger the
absolute value of the coefficient of a functions is, the more discriminating ability it has. The
least absolute value is for the variable income. The standardized canonical function also
buttresses the point earlier made from test of equality of means.

Table 5: Classification results

How often do you Predicted Group

experience food Membership
shortage in your Rarely Often Very
household? Often
Original Count Rarely 51 8 5 64
Often 0 14 1 15
Very Often 4 5 4 13
Percentage Rarely 79.7 12.5 7.8 100.0
Often .0 93.3 6.7 100.0
Very Often 30.8 38.5 30.8 100.0

Table 5 shows the classification result, a good number of the cases where correctly classified,
the average correctly classified among the three groups is 75%, this leaves an apparent error
rate of 25%. According to Gagne (2014), a model is efficient if it can correctly classify up to
seventy percent of the cases. It can therefore be said that the discriminant analysis is an
efficient model in classifying household food security in Heipang metropolis of Barkin Ladi
local government area.

Discussion
From the result obtained, 70% of the household rarely have food shortages, 16%, often have
food shortages while 14% very often have food shortages. Invariably most of the households
in the community rarely have food shortage. But most of those that rarely have food shortage
have household size of 5 and above, they account for about 56% which is 51 households in
the community. This seems contrary to reports that say larger households are usually not
secured food wise. Only about 46% who those who rarely have food shortage actually eats a
balanced meal, so although the larger homes don’t have food shortages the meal they

10
consume are rarely balanced. About 70% also they consume cultural food but a larger chunk
of this households have food shortages. 10% of the community says their income is adequate
and they can afford the food they consume. 50% of the populace says their income is not
adequate to afford them the food they consume. Only 12% of the community have access to
the food they consume and about 38% says they consume same meal everyday, which means
their food lack variety, surprisingly about half of this households are from those that rarely
have food shortages.
From the discriminant analysis, cultural food intake, household size, same meal everyday and
balanced food discriminate between the groups, it implies that these variables do not perform
equally in the three groups. Although the accessibility of the community to the food they
consume is not significant with about 60% having good access to the food they eat only about
28.2% could actually afford this food, also income and affordability are not significant, which
implies that these variables perform equally among the three groups. Income and
affordability affects every one of the three groups The correlation coefficient between food
balance and Household size was found to be 0.1896, this implies that as the household size
increases the chances of eating a balanced meal decreases. No surprising that although most
of the population said they rarely have food shortages about 54% of these household do not
eat balanced meal, and also this set of household have large sizes.

Conclusion
Although most of the residence of Heipang community don’t have food shortage, it can still
be observed that they lack so much in terms of the variety of food being consume and also its
quality, not meeting this requirement also means they are not secured food wise. In order to
achieve food security among residence of the Heipang, efforts must be tailored towards
educating the populace on creating variety of food and also improving on its quality.
Agencies responsible for agricultural products must endeavor to come up with different
variety of food that is quite affordable for the populace. And lastly the people must be
educated on limiting their household size to a number that they can afford to feed adequately
with quality and balanced meal. If the populace is encouraged to produce the local food they
consume, it would give them good access to the food and also remove the limited income
factor that is affecting the community in general.

11
References
Abu, G.A. & Soom, A. (2016). Analysis of factors affecting food security in rural and
urban farming households of Benue State, Nigeria. International Journal of food and
agricultural economics. Vol 4 (1) pp. 55-68

Adebiyi, B. (2012). Jonathan’s agricultural business plan, Nigeria Tribune, Wednesday

13 June, 2012
Babatunde, R.O, Omotesho, O.A. & Sholotan, O.S (2007). Socio-economic characteristics
and food security status of farming households in Kwara state, North-Central Nigeria.
Pakistan Journal of Nutrition 6(1): 49-59
Balogun, O.S., Akingbade, T.J., & Oguntunde, P.E. (2015) An assessment of the
performance of discriminant analysis and logistic regression methods in classification
of mode of delivery of an expectant mother, Mathematical Theory and Modeling vol.
5(5)

Ericksen, P.J. (2008) ‘What is the vulnerability of a food system to global environmental
change?’,Ecology and Society, Vol. 13, No. 2, p.14.

Food and Agriculture Organisation (2010). Food insecurity in the World: Addressing food
Insecurity in protracted crisis. Food and Agriculture Organisation of the United
Nations, Rome.
Food and Agriculture Organisation (2019). The state of food security and nutrition in the
World.

Gagne, F. (2014) Biochemical ecotoxicology: principles and methods. Elsevier, Academic

Press, pg. 209-229

Johnson, R. A. & Wichern D.W. (2007) Applied multivariate statistical analysis. 6th Edition

Havas A.K & Salman, M.D (2011) Food security: its components and challenges,
International Journal of Food Safety Nutrition and Public Health,
Hazell, P. & Wood, S. (2008) ‘Drivers of change in global agriculture’, Philosophical
Transactions of the Royal Society B: Biological Sciences, Vol. 363, No. 1491,
pp.495–515.
Honfoga, B.G. & Van Den Boon, J. G. M. (2003) food consumption patterns in Central West
Africa, 1961 to 2000, and challenges to combating malnutrition, Food and Nutrition
Bulletin, vol. 24(2), pp.167-182.

Ihimodu, I. L. (2004) Marketing of agricultural products and food security programme in

Nigeria. Paper presented at 13th annual congress of the Nigerian Rural Sociological
Association at LAUTECH, Ogbomoso Nogeria, Nov 25-28, 2003.
Onoja, A.A. Babasola, O.I., & Ojiambo, V. (2018) Application of discriminant analysis in
classification of food security status, International Journal of Scientific Engineering
Research, vol. 9(2)

12
Roos, J.A., Ruthven, G.A., Lombard M.J., & McLachian M.H., (2013) Food accessibility and
availability in the local food distribution system of low-income, urban community in
Worcester, in Western Cape province, South African Journal of Clinical Nutrition
volume 26(4).
Todorov, V. & Filzmosor, P, (2001) Robust statistics for one-way MANOVA,
Computational Statistics and Data Analysis. 54. 37-48

13