Content uploaded by Sarah Chawsheen
Author content
All content in this area was uploaded by Sarah Chawsheen on Oct 02, 2023
Content may be subject to copyright.
4
227
Vol.23, No.4, 2019
Using Factor Analysis for Expenditure Patterns of Peshmerga Households
in Erbil Governorate, with Focus on Food Items
ID No. 2395
(PP 227 - 244)
https://doi.org/10.21271/zjhs.23.4.13
Sarah Ahmed Hassan Chawsheen
College of Administration and Economic- Salahaddin University- Erbil
sarah.chawsheen@su.edu.krd
Received: 01/08/2018
Accepted: 15/05/2019
Published:19/09/2019
Abstract
Peshmerga is a military force in the Kurdistan Region of Iraq. This study elaborates on the expenditures
items of a sample of 385 Peshmerga households throughout two major sets of variables. The first set consists of
22 general expenditure patterns (variables) per month such as salary, years of service, military ranks, type of
healthcare; public sector or private sector, transportation expenses in Iraqi Dinars, children education expenses,
and so forth. And the second set consists of 22 food items (variables); bread expenses, meat expenses, sugar
expenses... etc. For each of either set, the researcher conducted an exploratory factor analysis by adopting the
method of Principal Component Analysis to reduce the number of variables to a fewer number of components.
As a result, 8 components extracted from the first set with the total variance explained the ratio of % 64.085, and
6 components from the second set with the total variance explained of %52.856.
Keywords: expenditure patterns, general household items, food items, factor analysis, principal component
analysis.
1 - Introduction
urdistan region of Iraq has been targeted by hostile countries for more than a century
make their highest benefit throughout trying to overpower and exploit its resources.
The last threat on this region was by ISIS attacks in the year 2014, where Peshmerga
Governorate of Erbil in the year 2015. Throughout highlighting the most significant factors
that may affect their living, by applying Factor Analysis over two sets of variables.
Factor analysis is used to find small number of original sizes, or factors, which can be
used to represent relationships among interrelated variables. There are two major types of
factor analysis exploratory analysis and Confirmatory analysis, each of them has many
methods to conduct. In this study, the researcher conducts a Principal Component Analysis
(PCA) method regarding Exploratory Factor Analysis. Too many variables influence any
househo
possibly will produce the similar data. So it is better to use processes that reduce the number
of the variables, for example, PCA with as fewer loss of data (Giri, 2004).
According to Deaton and Muellbauer (1980) assumption the utilities a household
derives from various materials at different levels of expenditure would be affected by
economic conditions. Therefore, consumption budget reduces in recession periods, consumers
cut expenditures disproportionately more in less essential categories, and larger shares for the
more essential. While in the economy's growth time frames, consumers are relied upon to
increment both aggregate utilization spending plan and offer dispensed to the less basic
classifications (Kamakura, and Du, 2011).
Peshmerga and their families have been gone through difficult life due to severe war
emphasis against ISIS forces, many of their children became orphans, many of
K
4
228
Vol.23, No.4, 2019
their wives became widows, as well as delaying and reducing of their monthly salaries due to
economic and financial crises in the region. All of these reasons made the expenditure on
households necessities affect their living adversely.
eholds expenditures in this study
diversified into two sets; general categories (items), and food categories, and each of them
consists of many variables as the researcher reveals them in the following sections.
1-1 The Significance of the study:
Peshmerga force is considered an important group in Kurdistan region society, the
significance of this study is to reveal the livelihood of their households, despite difficult
conditions that they were through particularly after ISIS attacks. By showing the most
significant expenditure patterns of their households and to provide an additional resource
for the related institutions in order to improve this living conditions in the future.
1-2 The Aims:
The aims of this study are to determine the significant factors that affect the
Expenditures of Peshmerga Households in the Governorate of Erbil in the year
2015. And showing the most significant variables in each factor throughout reducing the
magnitude number of general expenditure patterns (variables).
1-3 Hypotheses of the study:
1 - The amount of the variance of variables that is accounted for by the components
are having higher values for some crucial variables of the first set, such
as; salary, years of service, age. Also, other variables of the second set may take high
such as; prepared meals and vegetable expenses.
2 - The ratio of the total variance explained by the extracted factors, of all expenditure
patterns exceeds 60%, while for food items exceeds 50%.
3 - The correlation between the extracted components and both sets of the mentioned
variables or have higher values than the remaining variables.
4 - The resulting of extracted component score variables are representative and can
be used in place of, the original variables.
1-4 The Methodology:
This study adopts numerical analysis via on exploratory factor analysis. In order to
reduce the number of lots of variables that affect households expenditures to
a fewer number of components. Therefore, the method of principal components has been
chosen, including the most crucial tests concerning this method by the use of IBM SPSS
21 program for numerical calculations.
1-5 Data Descriptions:
A sample of 385 Peshmerga selected from a population of more
than 200,000 Peshmerga (Wikipedia, 2016), which their names recruited in KRG-
Ministry of Peshmerga. Peshmerga expenditures patterns consisted of two
main sets of variables; the first set is various patterns of household expenditures, and the
second set consisted of food items expenditures.
The first set of variables consists of;
Salary per Month in 1000 Iraqi Dinars (IDs), Years of Service, Military Ranks (1for
Peshmerga or private, 2 for private first class , 3 for corporal, 4 for sergeant, 5 for first
sergeant, 6 for non-commissioned officer, 7 first commissioned officer, 8 for lieutenant,
17 for major general), Age, Academic Certification (1 for non, 2 for the primary
and 3 for preparatory, 4 for college), Wife Income in 1000 IDs, House Rent in 1000 IDs,
delaying Salary (1 for yes, and 0 for no), Debts (0 for no, and 1for yes), Installment
4
229
Vol.23, No.4, 2019
per Month (0 for no, and 1 for yes), Extra Job income in 1000 IDs, Health Care type (1
for public and 0 for the private), Car Expenses in 1000 IDs, Transportation in 1000 IDs,
Children Education in 1000 IDs, Clothes in 1000 IDs, House Equipment in 1000 IDs,
Electricity in 1000 IDs, Water in 1000 IDs, Personal Expenses in 1000 IDs, War
Effects (1 for very bad, 2 for bad, and 3 no), and Food Expenses in 1000 IDs.
While the second set of variables entails the following food items expenditures in
1,000 IDs:
Bread, Edible Oil, Rice, Burghul, Wheat, Wheat Groats, Chicken Meat, Red Meat,
Fish Meat, Sugar, Flour, Vegetables, Fruits, Beans, Prepared Meals, Prepared Sweets,
Tea, Nuts, All items expenses except food, Salary, Wife income, and Extra job income.
2 - The Theoretical and Conceptual Frame of Factor Analysis:
Factor analysis is a statistical technique used to identify a fairly small number of
factors, which can be used to represent relationships among interrelated variables. Factor
analysis is the identification of core "factors," that might explain the dimensions associated
with data variability (Bartholomew et al., 2008).
The main objectives of using Factor Analysis, which construed by (Gorsuch, 2015) are:
1 -
2 - Describing data economically.
3 - Hypothesis testing.
4 - Data transformation.
5 - Exploratory uses.
6 - Treating with multicollinearity problems.
2-1 Types of factor analysis (FA):
There are two common types of factor analysis (FA): exploratory, confirmatory.
Exploratory factor analysis is utilized to estimate the hidden components influencing the
factors in an information structure without setting any predefined structure to the result, while
confirmatory factor analysis is executed for reconfirming the impacts and conceivable
connections existing between an arrangement of foreordained elements and factors (Dennis,
2006, p.52).
For a set of standardized variables like X1, X2P, their variances are each equal to
one and their covariance is the correlation coefficients. Xi is a standardized variable, i.e., xi =
(Xi
i. The correlation matrix must be intercorrelated, but no too powerfully, i.e. no
strong correlation between the variables. In SPSS the Bartlett's test of Sphericity used to
check for variables intercorrelation. The null hypothesis assumes the original correlation
matrix is an identity matrix. This test must be significant, i.e. there would be no relationships
between the variables. Also, we can check for Multicollinearity by SPSS throughout the
determinant of the correlation matrix, if the determinant is more than 0.00001, at that point,
there is no multicollinearity (Field, 2000, pp.445).
The objective of factor analysis is to represent the variables as a linear arrangement of
a smaller set of common factors plus a unique factor. We express this representation
mathematically as:
XP= LP1 F1+ LP2 F2Pm Fm + eP ------- (1)
Where the following assumptions are made:
1- m is the number of common factors (typically this number is much smaller than P, where
2- F1, F2m; are the common factors.
3- Lij is the coefficient of Fj in the linear combination describing xi. This is known as the
loading of the ith variable on the jth factor.
4- e1, e2P are unique factors, apiece associating to one of the original variables.
4
230
Vol.23, No.4, 2019
Since Xi is standardized, it is variance is equal to one and is composed of the
following two parts:
1- The Communality, i.e., the part of the variance that is due to the common factors.
2- The Specificity, i.e., the part of the variance that is due to the unique factor ei.
We can write the variance of Xi as: Var Xi = 1= hi2 + ui2 -------- (2)
Where; hi2 is the Communality, and ui2 is Specificity.
2-2 Methods of Factor Analysis:
Two main steps have to be achieved in order to do factor analysis:
- Estimating the factors loadingswhich called the initial solution.
- Rounding the factors to reveal their meanings and achieving the Final Solution.
There are many approaches for estimating factors loadings matrix in order to get the
initial solution, according to Child (2006)below are the main approaches:
1- Principal Component method exploratory.
2- Principal Axis method.
3- Image method.
4- Maximum Likelihood method Confirmatory.
5- Alpha method.
6- Unweighted Least Square.
7- Generalized least method.
8- The Centered method.
9- Rao method.
In this study, the method of Principal Component is applied due to its accuracy in
estimations.
2-3 Principal Components Analysis:
Principal components are uncorrelated, therefore, they are presenting an attractive
choice as factors. The greatest proportion of variance is explained by the first principal
component, therefore it is the most important.
To satisfy the assumption of unit variances of the factors, each principal component is
divided by its standard deviation. That is, the j common factor Fj is defined as;
Fj= C j/ ( Var C j)1/2 ------- (3), where C j is the j principal component.
variables Xi and the principal components C j should be reminded, (Afifi, et al., 2004, pp.
393-396) derive these relationships as:
C1= a11 x1+ a12 x21P xP
C2= a21 x1+ a22 x22P xP
.
.
.
CP= aP1 x1+ aP2 x2PP xP ------- (4)
It may be shown mathematically that this set of equations can be inverted to express
x1= a11 C1+ a21 C2P1 CP
x2= a21 C1+ a22 C2P2 CP
.
.
.
xP= a1P C1+ a2P C2PP CP -------- (5)
Note that the rows of the first set of equations becomes the columns of the second set
of equations. Now since Fj = Cj/ (VarCj)1/2, it follows that Cj = Fj ( Var Cj)1/2, and the ith
equation then could be expressed as:
X1= a1i F1 (Var C1)1/2 + a2i F2 (Var C2)1/2 Pi FP ( VarCP)1/2 ----------- (6)
4
231
Vol.23, No.4, 2019
This last equation is now modified in two ways:
1- Using the notation Lij = aji (Var Cj) 1/2 for the first m components.
2- Combining the last P-m terms and denote the result by ei. That is,
ei = am+1,i Fm+1 (Var Cm+1)1/2 Pi FP (Var CP)1/2 -------------- (7)
With these manipulations each variable xi is expressed now as
Xi= Li1 F1+ Li2 F2Lim Fm + ei ---------
Each variable is linked with the principal factor, this link is known as factor loading.
The principal component model has been transformed to produce the factor model. Note that
when the original variables are standardized, the factor loading (L ij) turns out to be the
correlation between Xi and Fj.
The sum of the squared factor loadings for all factors for a given variable is the
variance in that variable accounted for by all the factors, and this is called the Communality.
hi2 = Li12 + Li22 im2 ------- (9)
2-4 Factors Extraction Criteria
The number of factors to be extracted applying factor analysis is determined by
keeping the factors that record for the most variances. Suhr (2006) shows the criteria for
factor number determination are:
2-4-1 Kaiser's constraint:
According to Kaiser if the studied variables are having different measurement units,
common factors of eigenvalue greater than one should be extracted (Nunnally, 1978).
2-4-2 The Scree test of Cattell's (1966):
This plot looks somewhat like describe of an inclination. The piece past the elbow
relates to the debris, or Scree, that gathers. Cattell's rule bring for holding factors over the
elbow and removing those underneath it.
2-4-3 Interpretability criteria all through noting these inquiries:
A- No under 3 factors with essential loadings (> 0.30).
B- Factors that load on a factor share significant rank.
C- Variables that load on various components measure diverse models.
D- The rotated factor matrix shows a simple structure. There are relatively high loadings
on one factor or low loadings on others.
2-5 Rotation of Axes:
As we know the main purpose of factor analysis is to derive from the data easily
interpretable common factors. The initial extracted factors are habitually difficult to interpret,
but we can find new factors loadings, which are easier to interpret. These new factors called
the Rotated Factors, which are selected so that some of the loadings are very large near + - 1
and the remaining loadings are very small near zero. Then again, we would preferably wish,
for some random variable, it has a high loading on just a single factor. If so, it is simple to
give each factor an understanding emerging from the variables which is extremely correlated
(high loadings).
The orthogonal rotation techniques are: Equamax, Quartimax, Promax, and Varimax
which is adopted in this study. While other techniques such as Direct Quartimin, Promax, and
Direct Oblimin are applied upon Oblique rotations.
The Varimax rotation is attained by enlargement the total of the variances of the
squared factor loadings inside each factor. Harman (1976) says that these factor loadings are
balanced by dividing every single one of them by the Communality of the corresponding
variable, which saw as the systematization of Kaiser.
3- The Practical Part :
As mentioned previously the subject of this study is a
expenditures patterns in the Governorate of Erbil in the year 2015. The practical part showed
throughout two main sections, the first is about the descriptive statistics of expenditures of
4
232
Vol.23, No.4, 2019
Peshmerga households and the most important variables affecting them. While the second
section elaborates factor analysis of these households expenditures items. Throughout
highlighting the most significant factors that may affect their living during the course of
applying Factor Analysis over two sets of variables. The method of Principal Component is
used, and then interpret the results from an economic perspective for each set of the
outcomes.
3-1 Expenditures of Peshmerga Households and the Variables Affecting Them:
By implementing the software of SPSS 21 the descriptive statistics results of the first
set of variables are illustrated in table (1). The minimum salary of Peshmerga is 500 000 Iraqi
Dinars (ID) while the maximum is 3 950 000 ID, but the average salary is about 788 000 ID.
Their average expenditures on food items are 284 000 ID while clothes take about 88 000 ID
of their income. The maximum amount of money they might receive from an extra job despite
being Peshmerga is 1000 000 ID.
Table (1) Descriptive Statistics for the Expenditure Patterns of Peshmerga Households
(First Set of Variables)
Variables
N
Minimum
Maximum
Mean
Std. Deviation
Variance
Salary
385
500
3950
788.29
265.449
70462.982
Years of service
385
1
40
9.49
6.603
43.600
Military ranks
385
1
14
3.38
2.601
6.763
Age
385
0
78
34.52
8.510
72.417
Certification
385
1
4
2.09
0.874
0.763
Wife income
385
0
815
41.31
126.008
15878.104
House rent
385
0
600
124.17
184.832
34162.790
Delaying salary
385
0
1
0.95
0.211
0.045
Debts
385
0
1
0.72
0.449
0.201
Installment
385
0
2
0.68
0.474
0.225
Extra job income
385
0
1000
115.83
205.240
42123.328
Health care type
385
0
1
0.72
0.449
0.201
Car Expenses
385
0
500
68.74
68.632
4710.372
Transportation exp.
385
0
300
74.74
57.313
3284.785
Children education exp.
385
0
450
52.13
57.014
3250.574
Cloths exp.
385
0
530
87.95
61.246
3751.130
House Equipment exp.
385
0
575
69.72
67.897
4610.036
Electricity exp.
385
0
935
40.35
50.319
2532.019
Water supply exp.
385
0
30
5.84
3.531
12.470
Personal exp.
385
0
400
39.37
37.224
1385.646
War effect
385
1
3
1.38
0.546
0.298
Total of food exp.
385
101
769
283.84
91.586
8388.040
Source: calculated by the researcher using a random sample data of 385 survey forms/ Ministry of Peshmerga- KRG /2015.
In table (2) notice that on average nearly 23000 ID spent providing bread monthly,
nearly 27 000 ID spent for Rice, 39 000 ID for chicken meat, 29 000 ID for red meat, 20 000
ID for prepared meals, and the lowest amount of money spent for providing wheat grain
which is about 1000 ID.
Table (2) Descriptive Statistics for the Expenditures of Peshmerga Households on Food Items
(Second Set of Variables)
Variables
N
Minimum
Maximum
Mean
Std. Deviation
Variance
Bread exp.
385
0
90.0
23.273
13.6521
186.381
Edible oil exp.
385
0
45
12.49
7.204
51.902
Rice exp.
385
0
85
26.51
13.798
190.391
Burghul exp.
385
0
28
5.34
4.115
16.937
4
233
Vol.23, No.4, 2019
Wheat exp.
385
0
15
1.11
2.417
5.840
Wheat grots exp.
385
0
20
4.39
3.492
12.196
Chicken meat exp.
385
0
150
38.77
16.651
277.255
Red meat exp.
385
0
100
29.47
15.704
246.620
Fish meat exp.
385
0
84
15.04
15.051
226.530
Sugar exp.
385
0
40
9.04
6.706
44.968
Flour exp.
385
0
58
7.09
8.364
69.955
Vegetables exp.
385
0
50
17.05
9.384
88.067
Fruits exp.
385
0
50
22.63
9.727
94.609
Beans exp.
385
0
70
14.69
8.431
71.090
Prepared meals exp.
385
0
200
19.99
27.558
759.419
prepared sweets exp.
385
0
75
12.80
14.344
205.740
Tea exp.
385
0
80
11.65
9.210
84.826
Nuts exp.
385
0
60
12.49
11.120
123.646
All expenses except
food
385
115.00
1858.00
563.0156
238.11143
56697.052
Source: calculated by the researcher using a random sample data of 385 survey forms/ Ministry of Peshmerga- KRG /2015.
3-2 Factor Analysis Results throughout the following steps:
There are four basic steps that have to be considered for a Factor Analysis experiment:
1- Collecting and creating data of the correlation matrix.
2- Extracting the initial factor solution.
3- Rotation and interpretation.
4- Measuring factor scores to use in further analyses.
3-2-1 Kaiser - Meyer – Olkin test (KMO- test):
In order to measure the sampling adequacy Kaiser - Meyer Olkin test (KMO- test) is
adopted. If KMO > 0.5 then the sample size is adequate. Using SPSS the anti-image matrix of
covariance and correlations could be calculated, where for the adequate sample size the entire
elements of the diagonal matrix must be greater than 0.5 (Field, 2000, p. 446).
Applying for the SPSS 21 program the Kaiser-Meyer- Olkin (KMO-test) Measure of
Sampling Adequacy for the first set variables of this study is 0.613 > 0.5. It means that the
sample meets adequacy measurement. This measure varies between 0 and 1, and values closer
to 1 are better.
The result of this factor analysis is rather satisfactory: the correlation matrix was
test of Sphericity is significant 0.000, as it is shown in table 3. The validity of the previous
measurements directs us to conclude that our data is appropriate to start applying factor
analysis.
Table 3 Suitability test of the first set of data for factor analysis
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
0.613
Bartlett's Test of Sphericity
Approx. Chi-Square
2034.513
df
231
Sig.
0.000
Correlation matrix Det.
0.004
Concerning the second set variables in this study Kaiser-Meyer-Olkin (KMO-test)
Measure of Sampling Adequacy is displayed in table 4 which is 0.736 > 0.5 meaning the
sample meets adequacy measurement.
determinant was
appendix 5. The researcher concludes that the data is appropriate for a factor analysis for the
second set variables as well.
4
234
Vol.23, No.4, 2019
3-2-2 Extraction of Significant Principal Components:
The shaded parts with blue of table 7 shows the eight significant principal components
that affect expenditure patterns of Peshmerga households in the Governorate of Erbil (first set
of variables) extracted because each component is having an eigenvalue greater than one. The first component explains15.657% of
total variance, the second explains 9.580%... and the eighth component explains 4.724% of total variances. These eight factors interpret
64.085 % of total variances of the variables. These factors are important to distinguish the most significant variables in Peshmerga
Table (7) Factor Analysis of the first set of variables by using PCM for Expenditure Patterns of
Peshmerga Households
Total Variance Explained
Compo
nent
Initial Eigenvalues
Extraction Sums of Squared
Loadings
Rotation Sums of Squared
Loadings
Total
% of
Variance
Cumulat
ive %
Total
% of
Variance
Cumulat
ive %
Total
% of
Variance
Cumulat
ive %
1
3.445
15.657
15.657
3.445
15.657
15.657
2.364
10.746
10.746
2
2.108
9.580
25.237
2.108
9.580
25.237
1.969
8.951
19.697
3
1.925
8.752
33.989
1.925
8.752
33.989
1.750
7.956
27.654
4
1.727
7.850
41.839
1.727
7.850
41.839
1.723
7.832
35.486
5
1.528
6.946
48.785
1.528
6.946
48.785
1.677
7.622
43.107
6
1.231
5.594
54.378
1.231
5.594
54.378
1.649
7.497
50.604
7
1.096
4.983
59.361
1.096
4.983
59.361
1.518
6.898
57.502
8
1.039
4.724
64.085
1.039
4.724
64.085
1.448
6.583
64.085
9
0.956
4.344
68.430
10
0.890
4.044
72.474
11
0.853
3.879
76.353
12
0.751
3.413
79.766
13
0.701
3.188
82.953
14
0.603
2.742
85.695
15
0.564
2.562
88.257
16
0.516
2.347
90.604
17
0.481
2.185
92.789
18
0.399
1.815
94.604
19
0.395
1.796
96.400
20
0.328
1.491
97.891
21
0.279
1.266
99.157
22
0.185
.843
100.000
As for the second set variables, the extraction of principal components is shown in
table 8 the significant components extracted which are demonstrated in the yellow shaded
area. Since each component is having eigenvalue greater than one, 6 significant principal
components have been extracted. These six components interpret 52.856% of total variances
of the variables.
Table (8) Factor Analysis of the second set of variables by using PCM of Food Items consumed by
Peshmerga Households.
Table 4 Suitability test of the second set of data for factor analysis.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
0.736
Bartlett's Test of Sphericity
Approx. Chi-Square
1592.825
df
231
Sig.
0.000
Correlation matrix Det.
0.014
4
235
Vol.23, No.4, 2019
3-2-3 Scree Plot:
Figure 1 the Scree plot for the first set variables
Figure 1 shows the 8 significant components their eigenvalues greater than one. The
elbow of the Scree plot breaks at component 8, which means the extraction of 8 components
their eigenvalues greater than one. So 8 components show 64.085% of variances in the first
set variables.
While for the second set variables the breaking point of the scree plot is at component
Total Variance Explained
Component
Initial Eigenvalues
Extraction Sums of Squared
Loadings
Rotation Sums of Squared
Loadings
Total
% of
Variance
Cumulati
ve %
Total
% of
Variance
Cumulati
ve %
Total
% of
Variance
Cumulati
ve %
1
3.574
16.248
16.248
3.574
16.248
16.248
2.576
11.710
11.710
2
2.489
11.312
27.560
2.489
11.312
27.560
2.367
10.758
22.468
3
1.779
8.085
35.645
1.779
8.085
35.645
2.185
9.931
32.399
4
1.508
6.854
42.499
1.508
6.854
42.499
1.791
8.142
40.541
5
1.230
5.592
48.091
1.230
5.592
48.091
1.465
6.661
47.201
6
1.048
4.765
52.856
1.048
4.765
52.856
1.244
5.655
52.856
7
0.998
4.536
57.392
8
0.952
4.328
61.720
9
0.872
3.964
65.683
10
0.812
3.691
69.374
11
0.776
3.527
72.902
12
0.738
3.355
76.257
13
0.697
3.166
79.423
14
0.659
2.995
82.418
15
0.596
2.711
85.129
16
0.563
2.561
87.690
17
0.557
2.534
90.224
18
0.487
2.215
92.439
19
0.478
2.174
94.614
20
0.439
1.996
96.610
21
0.401
1.822
98.432
22
0.345
1.568
100.000
4
236
Vol.23, No.4, 2019
Figure 2 the scree plot for the second set variables
3-2-4 Rotated Component Matrix
We select the orthogonal rotation in favor of oblique rotation because the factor scores
in oblique rotation correlate rather low less than the absolute value of 0.3. After applying
each factor on the base of factor loadings comparisons. Afterward, we need to search of factor
loadings greater than 0.4 and select the highest loading for each component. Then, compared
with other components loadings for the same variable. If this value is greater than the
remaining components loadings then the value of this variable in this particular component is
significant. And so on for the remaining components loadings, but if its value was less than
the other loadings then the value of this variable is not significant. In such a case we test
another component comes after this one, and so on concerning the remaining component
loadings.
The Interpretation of the First Set Variables:
If Communalities are high indicate that the extracted components representing the
ilitary
with Communalities of 0.784, 0.773
ommunality of 0.389.
respectively, i.e. these variables are having a most significant impac
remaining variables are not significant in component 1. As for the component 2 variables;
econd highest loadings (in
absolute values) of; 0.714, -0.686, 0.664, and 0.518 respectively. This means that
significant impact on
Meanwhile, the loadings of 18 remaining
variables are not significant in component 2. And so on for the loadings of remaining
variables shaded in blue and their corresponding extracted components.
4
237
Vol.23, No.4, 2019
Table 9 Rotated Component Matrix for the first set variables
Rotated Component Matrix
-
Component
hi 2
1
2
3
4
5
6
7
8
Years of service
0.822
-0.118
-0.047
0.109
0.066
0.036
-0.228
0.106
0.773
Military ranks
0.718
-0.058
0.040
0.094
0.118
-0.071
0.347
-0.274
0.743
Age
0.697
-0.016
0.135
-0.126
-0.031
0.157
-0.161
0.338
0.686
Salary
0.584
-0.097
0.003
-0.020
0.168
0.552
0.177
-0.263
0.784
Installment
-0.045
0.714
-0.131
0.007
-0.024
-0.028
-0.047
0.106
0.543
War effect
0.027
-0.686
0.015
-0.149
0.062
0.022
-0.136
0.095
0.525
Debts
-0.139
0.664
-0.136
0.043
-0.017
-0.131
-0.260
-0.274
0.
House rent
-0.038
0.518
0.201
-0.308
-0.192
0.118
0.276
-0.145
0.
Car Expenses
0.085
-0.147
0.783
0.191
0.190
-0.061
0.108
-0.031
0.731
Transportation
exp.
-0.001
0.000
-0.733
0.296
0.015
-0.024
0.065
0.148
0.5
Personal exp.
-0.061
-0.057
0.148
0.777
-0.093
-0.063
0.108
-0.240
0.
Health care type
-0.029
-0.082
0.192
-0.631
-0.189
-0.113
0.251
-0.045
0.
Total of food exp.
0.266
0.252
0.021
0.544
0.300
0.021
0.131
0.192
0.
House Equipment
exp.
0.078
-0.097
0.125
0.110
0.820
0.029
0.088
0.064
0.728
Cloths exp.
0.194
-0.111
0.122
0.123
0.620
0.346
-0.074
0.229
0.
Children education
exp.
0.302
0.012
0.421
0.170
-0.455
0.198
-0.007
0.418
0.719
Electricity exp.
0.236
-0.119
-0.060
0.089
-0.031
0.839
-0.059
-0.120
0.804
Water supply exp.
-0.157
0.063
0.044
-0.067
0.160
0.607
0.256
0.145
0.516
Certification
-0.004
0.127
0.044
-0.103
-0.047
0.146
0.730
-0.139
0.604
Wife income
-0.100
-0.170
-0.102
0.162
0.174
0.018
0.621
0.358
0.620
Delaying salary
-0.050
0.096
0.142
0.105
-0.108
0.067
-0.029
-0.574
0.389
Extra job income
-0.115
-0.253
0.444
0.168
0.289
0.023
-0.122
0.447
0.602
The Interpretation of the Second Set Variables:
Concerning the second set variables, i.e. food items expenditures, table 10 shows that
ommunality among the other 21
variables which is 0.669 then the second highest Communality Vegetable
ains the lowest Communality of
0.375.
The shaded cells with yellow in the rotated component matrix show that component 1 consists
of most significant variables affecting food expenditures of Peshmerga households such as;
absolute value of their loadings; 0.780, 0.734, 0.711, and 0.671 respectively. Next comes
Fouler expenses, sugar
of; 0.681, 0.661, 0.636, 0.631, and 0.401 respectively. The same interpretation concerning the
remaining components.
4
238
Vol.23, No.4, 2019
Table 10 Rotated Component Matrix for the second set variables
3-2-5 Component Score Coefficient:
For each item and each component, the component score is computed by multiplying
the item's standardized variable values (computed using listwise deletion) by the component's
score coefficients (IBM Knowledge Center). The bring about 8 component score variables are
can be used in place of the 22 original variables with only a 36 % losses of information for the
first set variables and % 47 losses of information for the second set variables. Depending on
component score coefficient matrix in appendix 2, component score variable representing the
first component for the first set of the original variables is computed as:
Component 1 = 0.187salary- 0.051wife income- 0.121 extra job income+ 0.382 years
of service+0.372 military ranks+0.311 age+ 0.010 certification+0.011 house rent – 0.041
delaying salary -0.014 debts +0.024 installment +0.047 healthcare type- 0.004 car expenses+
0.022 transportation+ 0.091 children education -0.007 clothes- 0.002 house equipment -
0.050 electricity -0.183 water supply -0.070 personal expenses -0.025 war effect + 0.103 total
of food expenses.
And so forth concerning the other seven remaining components.
Relying on component score coefficient matrix in appendix 3 the component score
variable that represents the first component for the second set of original variables is:
Component 1= 0.008 Salary -0.042 Wife income + 0.004 Extra job income + 0.006 Bread +
0.059 Edible oil – 0.111 Rice – 0.062 Burghul +0.055 Wheat – 0.049 Wheat groats +0.009
Chicken meat +0.097 Red meat +0.348 Fish meat -0.025 sugar + 0.025 Flour – 0.069
Vegetables – 0.031 Fruits + 0.100 Beans + 0.317 Prepared meals + 0.294 Prepared sweets -
0.128 Tea + 0.277 Nuts - 0.029 All expenses except food
And the same way concerning the other five remaining components.
The resulting variables which represent each component are then
commonly normalized. Using the saved components is also preferable to using the original
variables such as; salary in 1000 IDs, years of service, military ranks, healthcare type, food
expenses, certification,... etc., because the components are representative of all 22 original
Rotated Component Matrix
Component
1
2
3
4
5
6
hi 2
Prepared meals exp.
0.780
0.005
0.087
-0.057
0.098
0.199
0.6
Fish meat exp.
0.734
0.111
-0.023
0.217
0.049
-0.171
0.6
prepared sweets exp.
0.711
-0.043
0.006
-0.144
0.044
0.294
0.616
Nuts exp.
0.671
0.020
0.189
-0.027
0.066
-0.067
Burghul exp.
-0.003
0.681
0.086
0.101
0.227
-0.014
Flouer exp.
0.085
0.661
0.008
0.015
-0.193
0.018
Sugar exp.
-0.071
0.636
-0.006
0.186
-0.357
0.088
Wheat grots exp.
-0.025
0.631
-0.101
0.166
0.304
-0.023
Beans exp.
0.349
0.401
0.361
-0.145
0.046
-0.251
Rice exp.
-0.028
0.053
0.701
-0.020
0.073
-0.077
Tea exp.
-0.057
0.036
0.569
-0.017
0.119
0.179
Chicken meat exp.
0.172
0.035
0.569
0.244
-0.038
0.081
Red meat exp.
0.342
-0.184
0.530
0.169
0.078
-0.046
Edible oil exp.
0.327
0.413
0.455
-0.083
0.030
0.056
Salary
-0.041
0.199
0.002
0.689
-0.042
-0.052
0.521
All expenses except food
-0.040
0.209
0.052
0.687
-0.017
0.337
Bread exp.
0.014
-0.101
0.137
0.663
0.194
-0.078
Vegetables exp.
0.083
0.023
0.259
0.061
0.765
-0.023
Fruits exp.
0.163
-0.162
0.444
-0.049
0.489
0.050
Wheat exp.
0.178
0.234
-0.203
0.332
0.479
0.125
Wife income
0.056
0.177
-0.010
-0.045
0.085
0.709
Extra job income
0.095
-0.267
0.193
0.157
-0.079
0.565
4
239
Vol.23, No.4, 2019
variables despite their different measurement units, and the components are not linearly
correlated with each other.
4- Conclusions:
The most important conclusions achieved in this study are:
1- The descriptive statistics showed that the average salary of 385 Peshmerga was
788000 IDs per month, the minimum salary was 500000 IDs, and the maximum
salary was 3950000 IDs. The average of their service period was 9 years, the
military ranking average year 2015, the food expenditures
average was 283000 IDs per month, while the average of various types of
h
and 2.
2- The amount of the variables variances that is accounted for by the components
ommunalities are having higher values for some crucial variables of the first
set, such as; salary, years of service, and age by; 0.784, 0.773, and 0.743
respectively. While the highest Communalities for the second set variables
Prepared meal expenses, Vegetable
0.669, 0.665, and 0.634 respectively.
3- Throughout factor analysis compliance using the principal component method, 8
significant components extracted from the first set of variables. Observing the
breaking point of the Scree- plot, was in the eighth factor. And 6 significant
components extracted from the second set variables. Detecting the breaking point
of the Scree- plot, was in the sixth factor. These extracted components (factors) are
responsible for determining most of the variations may occur in expenditures of
4- After applying Varimax rotation and suppressing the results of loadings by size,
the most significant variables that are having impact on various patterns of
of service, Milit rank of importance
rank of importance (component 8), see table 9.
5- The most significant variables that are having an impact on food items
expenses, Flour expenses,
Sugar expenses, Wheat grots expenses, and Beans expenses
importance (component 2). And s
6- The results revealed that the ratio of the total variance explained by the eight
extracted factors of all expenditure patterns of Peshmerga households (first set
variables) is equal to 64.085 %. While the total variance explained by the six
extracted factors of expenditures on food items (second set variables) is equal to
52.856 %.
7- The resulting 8 component score variables are representative of and can be used in
place of, the 22 original variables with only a 36 % loss of information for the first
set variables and % 47 loss of information for the second set variables. Depending
on the component score coefficient matrix.
4
240
Vol.23, No.4, 2019
5- Limitations and Further Studies:
This study is the first of the kind in assessing the most essential expenditure patterns
of Peshmerga households, though it is limited to the extraction of the most significant factors
(components) affecting the expenditure patterns. The researcher tried to apply parallel
analysis in order to reduce the 8 extracted components to 5 extracted components. But the
total ratio of variation explained by the 5 factors declined to 48.785% for the first set
variables. Concerning the second set variables the 6 extracted components reduced to 4
components. The total ratio of variation explained by the factors declined to 42.499%.
Therefore, the researcher sees not to reduce the number of extracted components throughout
applying parallel analysis.
It is possible to build a linear regression model for the extracted factors in order to
diagnose and fix the problem of multicollinearities between variables. In this study, the
problem for both sets of variables. Also, the researcher
suggests applying confirmatory factor analysis to confirm the factor structure of observed
variables in further studies.
References
- Afifi, Abdelmonem, Clark, Verginia A., and May, Susanne. (2004). Computer-Aided Multivariate
Analysis, 4th ed. Chapman & Hall/ CRC, USA.
- Angus Deaton, and John Muellbauer. (1980). An Almost Ideal Demand System, The American
Economic Review, Vol. 70, No. 3, pp. 312-326.
- Bartholomew, D.J., Steele, F., Galbraith, J., and Moustaki, I. (2008). Analysis of Multivariate Social
Science Data: Statistics in the Social and Behavioral Sciences Series, (2nd ed.). [e book]. Taylor &
Francis. available at:
- Byrne, Patrick J., and Capps Oral; Saha Atanu. (1996). Analysis of Food-Away-from-Home
Expenditure Patterns for U.S. Households, 198289. American Journal of Agricultural Economics,
Volume 78, Issue 3, p.p. 614627.
- Child, Dennis. (2006). The Essentials of Factor Analysis, 3rd edition, Continuum, London, and New
York.
- Fabrigar, Leandre R., and Wegner, Duane T. ( 2012). Exploratory Factor Analysis: Understanding
Statistics, Oxford University Press.
- Field, Andy. (2000). Discovering Statistics Using SPSS for Windows: Advanced Techniques for
Beginners1st edition, Sage Publications, Inc. Thousand Oaks, CA, USA.
- Giri, N.C. (2004) Multivariate statistical analysis. Marcel Dekker Inc., New York.
- Gorsuch, Richard L. (2015). Factor Analysis: Classical edition, Tylor & Francis Group, New York, and
London.
- Harman, H. Harry. (1976). Modern Factor Analysis, 3rd edition, The University of Chicago Press, USA.
- https://doi.org/10.1086/662611 [accessed: 5 June 2018]
https://en.wikipedia.org/wiki/Peshmerga [accessed 31 July, 2017].
- https://www.crcpress.com/textbooks/evaluation/9781439889794 [accessed: July 31, 2018]
- IBM Knowledge Center. Component score coefficient matrix. [online], available at:
https://www.ibm.com/support/knowledgecenter/en/SSLVMB_23.0.0/spss/tutorials/fac_cars_compsc
coeffmtrx_01.html [accessed 30 July 31, 2018].
- Kamakura, Wagner A., and Du Rex Yuxing. (2012). How Economic Contractions and Expansions
Affect Expenditure Patterns, Journal of Consumer Research, Volume 39, Issue 2, pp. 229-247. [online]
available at:
- Miller Kent D., and Bromiley Philip. (2017). Strategic Risk and Corporate Performance: an Analysis of
Alternative Risk Measures. Academy of Management Journal,[online] Vol. 33, No. 4. available at
https://doi.org/10.5465/256289 [accessed 4 May 2018].
- Nunnally, J. C. (1978). Psychometric theory, 2nd ed, New York: McGraw-Hill.
- Ruscio, John, and Roche, Brendan. (2012). Determining the number of factors to retain in an
exploratory factor analysis using comparison data of known factorial structure, Psychological
Assessment, Vol 24, issue 2, p.p. 282-292.
- Suhr Diana D. (2006). Exploratory or confirmatory factor analysis?. [pdf]. Cary, NC: SAS Institute.
available at:
http://tx.liberal.ntu.edu.tw/~PurpleWoo/Literature/!DataAnalysis/FactorAnalysis_SAS.com_200-31.pdf
[accessed 2 Oct.2017]
4
241
Vol.23, No.4, 2019
- Tiebout, Charles M. A Pure Theory of Local Expenditures. [pdf], University of Chicago Press.
Available at: https://www.journals.uchicago.edu/doi/pdfplus/10.1086/257839
- Wikipedia the free encyclopedia. [online] , available at :
https://en.wikipedia.org/wiki/History_of_Peshmerga , [accessed 23 Oct.2016].
Appendices
Appendix (1): Survey Form
- The following information is for married Peshmerga only and is used for university research purposes
With the help of your family, please complete this form for a period of not more than one week
Mark in the circle you choose. Thanks for your cooperation.
1- Salary in Iraqi Dinar:
2- Years of service as Peshmerga:
6- Military rank:
7- Age:
8- Do you carry an academic certification? No Primary Preparatory College
9- If your wife earns income, how much she gets per month in Iraqi Dinars?
10- If you live in a house you pay rent, how much you pay?
11- Does the government delay paying your salary? Yes No
12- Do you fall under the influence of loans most months of the year? Yes No
13- Do you have to pay monthly installments? Yes No
14- If you do any extra job other than the work of Peshmerga, how much is your monthly income in Iraqi Dinars?
15- When you and your family get sick, what kind of hospital do you go to? Public. Private
16- If you own a car how much you pay for fuel and maintenance monthly?
17-
18- How much is your children education expenses per month in Iraqi Dinar?
19- How much is clothed expenses for you and your family, per month?
20- How much does it cost to buy home equipment per month?
21- How much are the public electricity expenditures and generators per month in Iraqi Dinars?
22- How much is water expenditures in Iraqi Dinars per month?
23- How much are your personal expenses like cigarettes and other stuff per month?
24-
Very bad Bad Non
25- Please write down the monthly expenses to purchase the following food patterns accurately and in Iraqi Dinars:
1- Bread:
2- Edible oil:
3- Rice:
4- Burghul:
5- Wheat:
6- Wheat groats:
7- Chicken meat:
8- Red meat:
9- Fish meat:
10- Sugar:
11- Flour:
12- Vegetables:
13- Fruits:
14- Beans:
15- Prepared meals outside the home:
16- Prepared sweets outside the home:
17- Tea:
18- Nuts:
Appendix 2
The Component Score Coefficient Matrix of all expenditures patterns (first set variables)
Component
1
2
3
4
5
6
7
8
Salary
.187
-.037
-.041
-.049
.071
.250
.057
-.229
Wife income
-.051
-.072
-.116
.092
.005
-.054
.450
.269
Extra job income
-.121
-.035
.239
.067
.100
.021
-.093
.250
Years of service
.382
-.006
-.069
.004
-.006
-.107
-.121
.039
Military ranks
.372
-.030
-.023
.008
.062
-.233
.254
-.216
Age
.311
.087
.039
-.122
-.064
-.005
-.092
.242
Certification
.010
.039
.002
-.037
-.041
.025
.475
-.042
House rent
.011
.279
.152
-.163
-.042
.075
.146
-.010
Delaying salary
-.041
-.020
.120
.100
-.026
.078
-.065
-.422
Debts
-.014
.333
.001
.009
.112
-.022
-.196
-.134
Installment
.024
.410
-.023
-.019
.052
.007
-.037
.174
Health care type
.047
-.046
.096
-.359
-.033
-.110
.172
-.002
Car Expenses
-.004
-.016
.450
.098
.075
-.092
.040
-.090
Transportation exp.
.022
-.049
-.446
.175
-.048
-.012
.101
.137
Children education exp.
.091
.056
.232
.174
-.436
.126
.015
.350
Cloths exp.
-.007
.050
.039
-.034
.352
.165
-.106
.075
4
242
Vol.23, No.4, 2019
House Equipment exp.
-.002
.044
.044
-.075
.541
-.077
.017
-.060
Electricity exp.
-.050
-.068
-.054
.078
-.113
.569
-.126
-.112
Water supply exp.
-.183
.070
.011
-.041
.051
.419
.095
.111
Personal exp.
-.070
-.096
.103
.526
-.185
-.014
.083
-.193
War effect
-.025
-.364
-.044
-.087
-.006
.005
-.078
-.025
Total of food exp.
.103
.192
.006
.268
.106
-.058
.096
.137
Appendix 3
The Component Score Coefficient Matrix of Food items Expenditures (second set variables)
Component
1
2
3
4
5
6
Salary
.008
.007
-.007
.409
-.095
-.107
Wife income
-.042
.101
-.037
-.126
.085
.598
Extra job income
.004
-.133
.090
.078
-.104
.437
Bread exp.
.006
-.127
.028
.407
.062
-.133
Edible oil exp.
.059
.180
.196
-.109
-.058
.034
Rice exp.
-.111
.021
.374
-.039
-.038
-.070
Burghul exp.
-.062
.297
.008
-.043
.163
-.002
Wheat exp.
.055
.068
-.214
.130
.356
.078
Wheat grots exp.
-.049
.270
-.103
.001
.245
-.011
Chicken meat exp.
.009
-.025
.281
.133
-.148
.020
Red meat exp.
.097
-.123
.227
.123
-.067
-.092
Fish meat exp.
.348
-.014
-.121
.163
-.055
-.212
Sugar exp.
-.025
.265
.044
.061
-.277
.067
Flour exp.
.025
.291
.014
-.051
-.151
.019
Vegetables exp.
-.069
.007
.021
-.042
.544
-.018
Fruits exp.
-.031
-.070
.152
-.070
.306
.035
Beans exp.
.100
.174
.148
-.112
-.036
-.216
Prepared meals exp.
.317
-.022
-.072
-.036
-.003
.117
prepared sweets exp.
.294
-.027
-.098
-.089
-.016
.209
Tea exp.
-.128
.023
.295
-.063
.025
.149
Nuts exp.
.277
-.021
.006
.003
-.040
-.102
All expenses except food
-.029
.021
.009
.363
-.074
.218
Appendix 4
The Correlation Matrix of the First Set Variables
Sal
ary
Yea
rs
of
serv
ice
Milit
ary
rank
s
Ag
e
Certific
ation
Wif
e
inco
me
Ho
use
ren
t
Dela
ying
salar
y
De
bts
Install
ment
Extr
a
job
inco
me
He
alth
car
e
typ
e
Car
Expe
nses
Transport
ation
exp.
Child
ren
educa
tion
exp.
Clo
ths
exp
.
House
Equip
ment
exp.
Electr
icity
exp.
Wa
ter
sup
ply
exp
.
Pers
onal
exp.
W
ar
eff
ect
To
tal
of
fo
od
ex
p.
Correlation
Salary
1.0
00
.40
0
.505
.29
2
.135
.018
.10
5
.002
-
.23
2
-.086
-
.076
-
.01
4
.125
-.002
.114
.30
1
.229
.617
.16
3
.020
.01
9
.09
1
Years of
service
.40
0
1.0
00
.405
.59
9
-.193
-
.057
-
.18
9
-.069
-
.11
4
-.084
.016
-
.13
6
.087
.049
.205
.25
5
.105
.234
-
.06
6
.026
.17
2
.23
3
Military
ranks
.50
5
.40
5
1.00
0
.23
4
.252
-
.018
-
.03
4
.061
-
.12
8
-.068
-
.037
-
.03
8
.148
-.005
.122
.12
0
.134
.121
-
.00
5
.073
.00
9
.21
5
Age
.29
2
.59
9
.234
1.0
00
-.130
.001
-
.04
1
-.094
-
.19
3
-.096
.054
.01
9
.113
-.108
.308
.24
1
.062
.225
.07
4
-.133
.06
7
.13
0
Certificat
ion
.13
5
-
.19
3
.252
-
.13
0
1.000
.166
.21
8
.022
.01
0
.038
-
.117
.14
9
.023
-.075
.010
.00
7
-.055
.084
.20
2
-.015
-
.07
0
.04
6
Wife
income
.01
8
-
.05
7
-
.018
.00
1
.166
1.00
0
.00
5
-.103
-
.28
6
-.089
.067
-
.03
0
.091
.123
-.026
.16
5
.236
.000
.15
0
.089
.03
4
.13
0
House
rent
.10
5
-
.18
9
-
.034
-
.04
1
.218
.005
1.0
00
.142
.21
8
.240
-
.190
.20
7
-.006
-.133
.046
-
.18
9
-.153
-.006
.02
3
-.090
-
.25
9
-
.13
6
Delaying
salary
.00
2
-
.06
9
.061
-
.09
4
.022
-
.103
.14
2
1.000
.13
7
.004
-
.167
.05
5
-.037
-.074
-.005
-
.10
6
-.064
-.003
.01
1
.096
-
.09
5
.02
8
4
243
Vol.23, No.4, 2019
Debts
-
.23
2
-
.11
4
-
.128
-
.19
3
.010
-
.286
.21
8
.137
1.0
00
.383
-
.291
-
.08
7
-.178
.100
-.151
-
.17
0
-.159
-.141
-
.09
6
.064
-
.31
6
-
.00
5
Installme
nt
-
.08
6
-
.08
4
-
.068
-
.09
6
.038
-
.089
.24
0
.004
.38
3
1.000
-
.129
-
.04
6
-.165
.125
-.027
-
.12
3
-.098
-.075
-
.01
8
.011
-
.28
1
.04
0
Extra job
income
-
.07
6
.01
6
-
.037
.05
4
-.117
.067
-
.19
0
-.167
-
.29
1
-.129
1.00
0
-
.03
7
.324
-.086
.215
.34
2
.300
.002
.03
0
.059
.23
2
.13
7
Health
care type
-
.01
4
-
.13
6
-
.038
.01
9
.149
-
.030
.20
7
.055
-
.08
7
-.046
-
.037
1.0
00
.027
-.158
.019
-
.21
0
-.130
-.113
.04
0
-.234
.05
6
-
.32
5
Car
Expenses
.12
5
.08
7
.148
.11
3
.023
.091
-
.00
6
-.037
-
.17
8
-.165
.324
.02
7
1.000
-.412
.196
.20
2
.252
-.022
.05
0
.281
.08
2
.07
2
Transport
ation
exp.
-
.00
2
.04
9
-
.005
-
.10
8
-.075
.123
-
.13
3
-.074
.10
0
.125
-
.086
-
.15
8
-.412
1.000
-.124
.01
4
.034
-.007
-
.02
3
.132
-
.01
7
.08
7
Children
education
exp.
.11
4
.20
5
.122
.30
8
.010
-
.026
.04
6
-.005
-
.15
1
-.027
.215
.01
9
.196
-.124
1.000
.07
3
-.101
.132
.03
0
.015
.05
8
.10
3
Cloths
exp.
.30
1
.25
5
.120
.24
1
.007
.165
-
.18
9
-.106
-
.17
0
-.123
.342
-
.21
0
.202
.014
.073
1.0
00
.445
.264
.16
0
-.009
.15
9
.25
3
House
Equipme
nt exp.
.22
9
.10
5
.134
.06
2
-.055
.236
-
.15
3
-.064
-
.15
9
-.098
.300
-
.13
0
.252
.034
-.101
.44
5
1.000
.035
.13
6
.065
.10
8
.26
8
Electricit
y exp.
.61
7
.23
4
.121
.22
5
.084
.000
-
.00
6
-.003
-
.14
1
-.075
.002
-
.11
3
-.022
-.007
.132
.26
4
.035
1.000
.23
3
.031
.05
5
.06
4
Water
supply
exp.
.16
3
-
.06
6
-
.005
.07
4
.202
.150
.02
3
.011
-
.09
6
-.018
.030
.04
0
.050
-.023
.030
.16
0
.136
.233
1.0
00
-.074
-
.00
6
.11
2
Personal
exp.
.02
0
.02
6
.073
-
.13
3
-.015
.089
-
.09
0
.096
.06
4
.011
.059
-
.23
4
.281
.132
.015
-
.00
9
.065
.031
-
.07
4
1.00
0
-
.08
1
.19
8
War
effect
.01
9
.17
2
.009
.06
7
-.070
.034
-
.25
9
-.095
-
.31
6
-.281
.232
.05
6
.082
-.017
.058
.15
9
.108
.055
-
.00
6
-.081
1.0
00
-
.19
6
Total of
food exp.
.09
1
.23
3
.215
.13
0
.046
.130
-
.13
6
.028
-
.00
5
.040
.137
-
.32
5
.072
.087
.103
.25
3
.268
.064
.11
2
.198
-
.19
6
1.0
00
a.
Determi
nant =
.004
Appendix 5
The Correlation Matrix of the Second Set Variables
Bre
ad
exp
.
Edi
ble
oil
exp.
Ric
e
exp
.
Burg
hul
exp.
Wh
eat
exp.
Wh
eat
grot
s
exp.
Chick
en
meat
exp.
Re
d
me
at
exp
.
Fis
h
me
at
exp
.
Sug
ar
exp
.
Flou
er
exp.
Vegetab
les exp.
Fru
its
exp
.
Bea
ns
exp.
Prepar
ed
meals
exp.
prepar
ed
sweet
s exp.
Tea
exp
.
Nut
s
exp
.
All
expen
ses
except
food
Sala
ry
Wife
inco
me
Extr
a job
inco
me
Correlation
Bread
exp.
1.0
00
.053
.08
3
.057
.155
.151
.143
.16
9
.13
3
.01
0
-
.077
.146
.10
6
-
.01
9
.016
-.083
.07
6
.01
7
.255
.233
.034
.092
Edible
oil exp.
.05
3
1.00
0
.23
9
.254
.083
.200
.168
.26
1
.23
4
.17
9
.145
.119
.18
8
.37
3
.298
.190
.23
2
.25
8
.088
-
.004
.102
.022
Rice
exp.
.08
3
.239
1.0
00
.118
-
.035
.014
.233
.30
5
.03
1
-
.04
6
-
.001
.180
.21
7
.18
8
.061
.036
.28
3
.15
3
.007
.037
.026
.031
Burghul
exp.
.05
7
.254
.11
8
1.000
.239
.383
.084
-
.00
7
.04
7
.26
0
.318
.134
-
.02
2
.24
0
.088
.002
.07
1
.05
7
.190
.208
.037
-.086
Wheat
exp.
.15
5
.083
-
.03
5
.239
1.00
0
.212
.081
.01
4
.22
2
.07
9
.091
.231
-
.00
8
.02
5
.130
.066
.07
6
.09
9
.179
.174
.055
.070
Wheat
grots
exp.
.15
1
.200
.01
4
.383
.212
1.00
0
-.010
-
.06
7
.10
9
.23
1
.295
.109
-
.03
3
.15
3
-.002
-.036
.00
9
.01
5
.196
.182
.094
-.124
Chicken
meat
exp.
.14
3
.168
.23
3
.084
.081
-
.010
1.000
.25
4
.13
6
.06
5
.053
.196
.23
5
.28
7
.161
.105
.20
0
.17
2
.149
.102
.060
.179
Red
meat
exp.
.16
9
.261
.30
5
-.007
.014
-
.067
.254
1.0
00
.22
6
-
.14
1
-
.018
.221
.29
6
.12
5
.307
.181
.17
4
.21
9
.082
-
.006
.015
.109
Fish
meat
exp.
.13
3
.234
.03
1
.047
.222
.109
.136
.22
6
1.0
00
.05
3
.133
.134
.10
7
.27
1
.418
.297
-
.00
6
.41
3
.079
.036
.038
.007
Sugar
exp.
.01
0
.179
-
.04
6
.260
.079
.231
.065
-
.14
1
.05
3
1.0
00
.398
-.128
-
.16
1
.11
4
-.100
-.079
-
.03
2
-
.06
4
.265
.162
.095
-.101
Flouer
exp.
-
.07
7
.145
-
.00
1
.318
.091
.295
.053
-
.01
8
.13
3
.39
8
1.00
0
-.032
-
.09
1
.16
6
.065
-.004
.01
4
.04
1
.114
.114
.048
-.066
Vegetab
les exp.
.14
6
.119
.18
0
.134
.231
.109
.196
.22
1
.13
4
-
.12
8
-
.032
1.000
.41
6
.20
3
.145
.069
.14
1
.14
6
.094
.022
.032
.034
Fruits
exp.
.10
6
.188
.21
7
-.022
-
.008
-
.033
.235
.29
6
.10
7
-
.16
1
-
.091
.416
1.0
00
.12
9
.192
.208
.18
8
.19
8
.041
-
.048
.034
.055
Beans
exp.
-
.01
9
.373
.18
8
.240
.025
.153
.287
.12
5
.27
1
.11
4
.166
.203
.12
9
1.0
00
.206
.172
.06
0
.24
7
-.032
-
.014
.003
-.060
Prepare
d meals
exp.
.01
6
.298
.06
1
.088
.130
-
.002
.161
.30
7
.41
8
-
.10
0
.065
.145
.19
2
.20
6
1.000
.581
.09
4
.40
1
-.011
-
.040
.122
.135
prepare
d
sweets
exp.
-
.08
3
.190
.03
6
.002
.066
-
.036
.105
.18
1
.29
7
-
.07
9
-
.004
.069
.20
8
.17
2
.581
1.000
.03
7
.36
9
-.015
-
.034
.152
.119
4
244
Vol.23, No.4, 2019
Tea
exp.
.07
6
.232
.28
3
.071
.076
.009
.200
.17
4
-
.00
6
-
.03
2
.014
.141
.18
8
.06
0
.094
.037
1.0
00
.14
4
-.004
.064
.023
.134
Nuts
exp.
.01
7
.258
.15
3
.057
.099
.015
.172
.21
9
.41
3
-
.06
4
.041
.146
.19
8
.24
7
.401
.369
.14
4
1.0
00
-.059
.000
.043
.070
All
expense
s except
food
.25
5
.088
.00
7
.190
.179
.196
.149
.08
2
.07
9
.26
5
.114
.094
.04
1
-
.03
2
-.011
-.015
-
.00
4
-
.05
9
1.000
.423
.180
.161
Salary
.23
3
-
.004
.03
7
.208
.174
.182
.102
-
.00
6
.03
6
.16
2
.114
.022
-
.04
8
-
.01
4
-.040
-.034
.06
4
.00
0
.423
1.00
0
.018
-.076
Wife
income
.03
4
.102
.02
6
.037
.055
.094
.060
.01
5
.03
8
.09
5
.048
.032
.03
4
.00
3
.122
.152
.02
3
.04
3
.180
.018
1.00
0
.067
Extra
job
income
.09
2
.022
.03
1
-.086
.070
-
.124
.179
.10
9
.00
7
-
.10
1
-
.066
.034
.05
5
-
.06
0
.135
.119
.13
4
.07
0
.161
-
.076
.067
1.00
0
a.
Determin
ant =
.014
