EFFECT OF SELECTED INSECTICIDE ON WHITEFLY (Bemisia tabaci) INFESTING BRINJAL PLANTS 324 Examination of Urban Consumers’ Propensity to Consume Margarine by Applying Correspondence Analysis Slobodan Nicin Assistant Professor; Faculty of European Legal and Political Studies, Novi Sad, Serbia Abstract It is well known that margarine as food is of great nutritious significance to the nourishment of the population. The aim of this paper is to carry out research, which is based on the data obtained from the survey, regarding the propensity of urban population market consumers to consume margarine. The survey has been carried out on a representative sample of the citizens of urban area. Corre- spondence analysis will be used in this paper, with the aim of determining the impact of selected factors on consumers’ preferences for margarine. Keywords: Margarine market, margarine consumption, consumers’ preferences, correspondence analysis Introduction 1 Маrgarine as food plays an important role in human nutrition. This paper aims at examining propensity of urban market consumers’ to consume margarine based on a sample of citizens of Novi Sad through the application of contemporary statistical methods. Novi Sad is the second largest city as well as second biggest industrial centre in Serbia. The impact of a larger number of representative factors which have an effect on supply and demand of margarine, will also be examined, as well as interdependency of the above factors. Correspondent analysis methods will be used in the paper which enables the examination and determination of matching – correspondencies of modalities of several factors by decomposition of a singular value, as well as χ 2 – independency test. Correspondence analysis is used in situations where it is necessary to quantify Corresponding author’s Email address: bobanicin@yahoo.com the quality of information contained in a nominal variable . As the main objective of Correspondence Analysis can also be specified testing connectivity between categories of rows or columns. correspondent analysis method can be considered as a technique used in the description of the data ,as the manner of their presentation in an aappropriate graphic form. This presentation accelerates and simplifies the testing and interpretation of results . Case studies are , on the one hand , factors that may have an impact on consumption and consumer preferences for margarine, and on the other hand consumers prefer- ences in using and purchasing margarine. Working methods and data sources The aim of this paper is to apply correspon- dence analysis method to the data obtained on the basis of consumers’ preferences for margarine. The data to which correspon- dence analysis is applied have been obtained Asian Journal of Agriculture and Rural Development journal homepage: http://aessweb.com/journal-detail.php?id=5005 mailto:bobanicin@yahoo.com Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 325 on the basis of the survey carried out on the sample of citizens of Novi Sad. The research includes 500 people surveyed, according to the plan of a simple random sample.The task of this paper would be to determine, on the basis of representative variables taken as factors in correspondence analysis, the cor- relation of margarine consumption (consum- ers’ propensity to consume margarine) and the examined variables (factors). Population age structures (5 categories) have been taken as variables in this paper, house- hold size (5 categories), and degree of pro- fessional qualifications (4 categories), fre- quency of buying margarine (7 categories), main reason for using this foodstuff (7 cate- gories) and type of margarine (4 categories). 4 categories of margarine were indicated in the original survey questionnaire, in relation to the last category as follows: dairy marga- rine, vegetable margarine , diet margarine and soya margarine. In respect to the fact that there was not a single response regard- ing the consumption of soya margarine, the research results refer to the remaining 3 categories. Statistical software Statistica10.0 has been used for the purpose of statistical data proc- essing in this paper. Correspondence analysis is widely used in research related to the marketing and market (particularly in research of consumers’ pref- erences). Greenacre (2007) indicates in his book ”Correspondence Analysis in Prac- tice“that correspondence analysis is a statis- tical technique which serves a useful pur- pose to a great number of researchers, scien- tists and people working in the practical field who deal with the collection and proc- essing of categorical data, especially in terms of sociological research. This tech- nique is a great help to the analysis of the data from cross tables, provided in the form of numerical frequencies, and as a result, they provide a simple graphic representation which enables faster interpretation and un- derstanding of the data. According to the group of authors (Hair et al., 2006), corre- spondence analysis represents an explana- tory technique used for data analysis which is adjusted to the aim of the analysis of sim- ple two-directional and multidirectional ta- bles which contain some of the measures re- garding correspondence between rows and columns. Unlike the usual, traditional testing of hypotheses which serve to prove (dis- prove) a priori assumptions regarding the re- lation that exists among variables, a more detailed, i.e. in-depth data analysis is used for the identification of systematic relations among variables in the events when there is no information or a priori expectations are incomplete in terms of the character of such relations. Detroja et al. (2006) emphasize that corre- spondence analysis is a powerful multi- variation technique based on a general de- composition of a singular value. Considering a great importance of graphic representation of tables of contingency in a low- dimensional space for a simple and efficient interpretation of some dependency among rows and columns, the need for correspon- dence analysis arises which enables optimal graphic representation of contingency tables in a low-dimensional space (Jobson,1992). Correspondence analysis can be treated as a decomposition technique of total - χ 2 – value by determining a small number of dimen- sions which can serve to represent devia- tions of real values from the expected val- ues. Thereby χ 2 – value is based upon the sum of square deviations of real frequencies from the expected ones in the contingency table (Bendixen, 2003). Correspondence analysis also includes de- termination of inertia value. Overall or total inertia value is determined as a quotient of total χ 2 – value and total number of units of observation in the contingency table. Inertia is a measure of variation in the table and it does not depend on the size of a sample. Provided that correspondence analysis is carried out in greater detail, decomposition of total inertia into the components of rows and columns is possible, similar to the method of variance analysis, at which inertia Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 326 of rows and columns is determined. That way more precise explanations are obtained regarding homogeneity, considering that zero hypotheses in the analysis of the con- tingency table are a hypothesis of homoge- neity of rows, that is, columns. Contribution to the inertia of rows and columns is usually expressed relatively in relation to the total inertia. Maximum number of characteristic values which arises from the two- dimensional table equals the produce of number of columns reduced by 1 and the number of rows reduced by 1(STATISTICA 10.0, 2011). There are numerous examples of the appli- cation of correspondence analysis in the re- search of consumers’ preferences for differ- ent sorts of agricultural farm products and foodstuffs. Green et al. (1987) use correspondence analysis as a technique which enables the analysis of relations regarding the prediction of profiles of consumers’ choices and their demographic features. Hoffman and DeLeeuw (1992) use corre- spondence analysis in the context of con- necting objects – different brands of goods, on the basis of centroid principles. Аuthors state that according to this principle, trade- marks which are geometrically close to each other are actually similar trade marks. Pantelić and Savić (2000) state that tests based on - distribution are frequently used in the course of analysis of the data obtained on the basis of marketing research. Lacomba (2001) refers to correspondence analysis as a visual technique which is faster and simpler than other similar techniques and which provides an insight into the structure of consumers’ preferences with reference to different products. Kleij and Musters (2003) study consumer preferences according to different varieties of mayonnaise by using correspondence analysis which provides visualization comparable to preference maps. Liggett et al. (2008) apply the method of multi-dimensional scaling as a appropriate method that in some way enables estimation of perfect product on the basis of consumers’ opinion about similar products. Panea et al. (2009) refers to a correspondence analysis as an usefull technique for simultaneous graphical presentation of continous and categorial variables and use this method in the analysis of beef-quality based on consumers’profile. Avilez et al. (2010) apply correspondence analysis as a method for investigating of re- lationship between milk production, farm size and producers’ educational level. Guerrero et al. (2010) apply correspondence analysis in investigation of relationship of European consumers’ preferences towards the certain foods and expression “tradi- tional”. Nićin (2010) uses a correspondence analysis method in the research of preferences of consumers in Novi Sad according to the consumption of cheese. Beh et al. (2011) use correspondence analy- sis in studying perceptions of food of con- sumers in Europe. Chollet et al. (2011) use correspondence analysis in examination of eight sensor char- acteristics of beer. De Souza et al. (2011) apply correspondence analysis with purpose to determine the best sweetener that producers use in production of Swiss cheese. Larsson and Orsini (2011) use correspon- dence analysis in researching the consump- tion of fish products. Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 327 Examination results According to the survey results approxi- mately equal number of the examinees out of the total number of the surveyed house- holds (250) state that they buy margarine once a month (24.7%), once a week (23.4 %) and once every two weeks (22.6%). The percentage of the examinees who buy mar- garine once a month is in decline propor- tionately to the household size. The fourth position is taken by the examinees who have responded that they buy margarine 2-3 times a week (13%).Less frequently than once a month, margarine is bought by 8.4% of the examinees. Margarine is bought on a daily basis by 7.1% of the examinees. Response “something else” (which is not defined by the questionnaire) has been provided by 0.84% of the examinees. Household size has displayed a statistically significant correspondence with the fre- quency of buying margarine χ 2 = 38.99 (р=0,028, 05.0 ), whereas total inertia is in the amount of 0.163 (Table 1.). 50.5% of total inertia has been explained by means of the first dimension, 27.9% of total inertia has been explained by means of the second one,18.5% of total inertia has been ex- plained by means of the third dimension, аnd 3.1% of total inertia has been explained by means of the fourth dimension- variabil- ity, which means that in total about 78% of total inertia – variability of the observed phenomenon has been explained by means of the first two dimensions. The first two dimensions explain in a statistically significant way the variability of the observed phenomenon. Examinees from households with one member who buy margarine once a month or less frequently have contributed to a great extent to the explanations of inertia, as well as those whose response to this question has been “ something else”, examinees from households with two members who buy margarine once a week or once a month, examinees from households with 3 and 4 members who buy margarine 2-3 times a week, as well as the examinees from the largest households who buy margarine every day or less frequently than once a month. Definitely the greatest contribution to the explanation of inertia is made by the exami- nees from households with one member whose response is “something else”– re- sponse which is not provided within the questions by the questionnaire, as well as the examinees from households with three members who buy margarine 2-3 times a week. It can be concluded on the basis of two- dimensional graphic representation of coor- dinates of rows and columns that there is a correspondence of examinees from house- holds with two members with baying marga- rine everyday and once a month, as well as of the examinees from the largest house- holds and households with four members with buying margarine once a week or once every two weeks (Figure 1). Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 328 Figure 1: Two-dimensional plot of the coordinates of rows (household size) and columns (frequency of purchasing the margarine) Source: Authors calculation based on questionnaire results In addition, it can be noticed that with reference to the vertical axes of two- dimensional graphic representation, in terms of the frequency of buying margarine, groups from any household size except for the households with one member are quite homogenous. In other words, with reference to the vertical axes, on the one hand, house- holds with two members and largest house- holds can be noticed as sub-groups, and on the other hand, households with three and four members. Thereby it can be noticed that households with one member are signifi- cantly different from the above sub-groups. With reference to the horizontal axes of the Figure 1 homogenous groups of the exami- nees from both the smallest and largest households can be noticed. The largest number of the examinees from all age categories indicates habit as a main reason for using margarine (42.5%). Table 1: Eigen values and total inertia for all dimensions (household size of examinees  fre- quency of margarine purchasing) Eigenvalues and total inertia for all dimensions: Tоtal inertia = 0.163, χ 2 = 38.99, Degrees of freedom= 24;p = 0.027, 05.0 Dimension Number Singular Val- ues Eigenvalues % of Total Inertia Cumulative In- ertia χ 2 value 1 0.29 0.08 50.48 50.48 19.68 2 0.21 0.05 27.93 78.41 10.89 3 0.17 0.03 18.48 96.89 7.20 4 0.07 0.01 3.11 100.00 1.21 Source: Authors calculation based on questionnaire results Nice flavour of margarine as a main reason has been provided by 30.5% of the exami- nees. Nutritional value of margarine is in the third place as a reason for using it – 14.6%. 2D Plot of Row and Column Coordinates; Dimension: 1 x 2 Row .Coords 1- 1 member 2- 2 members 3- 3 members 4- 4 members 5- 5 or more members Col.Coords 1- Every day 2- 2-3 times a w eek 3- Once in a w eek 4- Once in tw o w eeks 5- Once in a month 6- Less frequently than once in a month 7 - Something else 1 2 3 4 5 1 2 3 4 5 6 7 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Dimension 1; Eigenvalue: .08235 (50.48% of Inertia) -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 D im en si on 2 ; E ig en va lu e: .0 45 56 ( 27 .9 3% o f I ne rt ia ) Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 329 Reasonable price of the product has been in- dicated as a main reason by 3.4% of the ex- aminees. High quality of the product has been indicated as a main reason by 3% of the examinees. Age of the examinees has showed a statistically significant correspon- dence with the responses related to the main reason for margarine consumption, χ2 = 39.84 (р=0.022), whereas total inertia is 0.171 (Table 2.) Тable 2: Eigen values and inertia for all dimensions (age of examinees  main purpose of us- ing margarine) Eigenvaluesandtotalinertiaforalldimensions:Total inertia = 0.171, χ 2 =39.84, Degrees of freedom= 24;p = 0.022 05.0 Dimension Number Singular Values Eigenvalues % of Total Inertia Cumulative Iner- tia χ2 value 1 0.31 0.09 55.07 55.07 21.94 2 0.22 0.05 28.92 83.99 11.52 3 0.14 0.02 12.25 96.24 4.88 4 0.08 0.01 3.76 100.00 1.50 Source: Authors calculation based on questionnaire results About 55% of total inertia – variability has been explained by means of the first dimen- sion, 28.9% of total inertia – variability has been explained by means of the second di- mension, 12.2% of total inertia – variability has been explained by means of the third dimension, аnd 3.8% of total inertia – vari- ability has been explained by means of the fourth dimension ,which means that in total about 84% of total inertia variability of the observed phenomenon has been explained by means of the first two dimensions. The first two dimensions explain variability of the observed phenomenon in a statistically significant way. Great contribution to the explanations of the inertia has been made by the youngest examinees, who indicate high quality of the product, good market supply and “some other reasons” as main reasons for margarine consumptions“,then by the examinees aged 30-39 who indicate habit, high quality of the product and nice flavour, next by the examinees aged 40-49 who indi- cate high quality of the product and “some other reasons”, followed by the examinees aged 50-59 and the oldest examinees who refer to the high quality of the product, good market supply and nice flavour. Definitely the greatest contribution to the explanation of inertia has been provided by the examinees aged 30-39, who indicate nice, flavour as a main reason for margarine consumption, as well as the oldest exami- nees who indicate high quality of the prod- uct. It can be concluded on the basis of two- dimensional graphic representation of coordinates of rows and columns that there is a correspondence of the examinees aged 50-59 with the reasonable price of margarine as a main reason for its use, correspondence of the examinees aged 30-39 with the nice flavour as a main reason, as well as the ex- aminees aged 40-49 who indicate nutritional value of margarine as a main reason. Fur- thermore, in relation to the vertical axes of two-dimensional graphic representation, in terms of the frequency of margarine con- sumption, pretty homogenous youngest group of the examinees and group of the ex- aminees aged 50-59 can be noticed, whereas in relation to the horizontal axes of the graphic, pretty homogenous group of the ex- aminees aged 30-39 and the oldest group of the examinees on the one hand, and a group of the examinees aged 40-49 and 50-59 on the other hand can be noticed (Figure 2). Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 330 Figure 2: Two-dimensional plot of the coordinates of rows (age of examinees) and columns (main purpose of using margarine) Source: Authors calculation based on questionnaire results The greatest number of the examinees of all the categories buys dairy margarine (50.4%). There is an equal percentage of those who buy both vegetable and diet margarine (25.6%, i.e. 23.9% of the examinees). Degree of professional qualifications has showed no statistically significant correspondence with the type of margarine that consumers most frequently buy, χ 2 =7.22 (р=0.301, 05.0 ), whereas total inertial amounts to 0.0303 (Table 3). 68.86% of to- tal inertia – variability has been explained by means of the first dimension, and the rest of the total inertia – variability has been ex- plained by means of the second dimension, which means that in total a complete total inertia-variability of the observed phenome- non has been explained by means of the first two dimensions. The first two dimensions explain in a statistically significant way variability of the observed phenomenon. Significant contribution to the explanations of the inertia has been made by the exami- nees with primary education, who buy dairy margarine and diet margarine, then by the examinees with secondary education who buy vegetable margarine, as well as by the examinees with high school education who buy both dairy and vegetable margarine. Тable 3: Eigen values and total inertia for all dimensions (level of education of examinees х margarineе type) Eigenvaluesandtotalinertiaforalldimensions: Total inertia = 0.030, χ 2 = 7.22, Degrees of- freedom= 6;p = 0.301, 05.0 Dimension Number Singular Values Eigenvalues % of Total Inertia Cumulative Iner- tia χ2 value 1 0.14 0.02 68.9 68.9 4.97 2 0.10 0.01 31.1 100.00 2.25 Source: Authors calculation based on questionnaire results 2D Plot of Row and Column Coordinates; Dimension: 1 x 2 Row .Coords 1-19-29 years 2- 30-39 years 3- 40-49 years 4- 50-59 years 5- 60 and more years Col.Coords 1- Habit 2- High Product Quality 3- Reasonable Price 4- Good Market Supply 5- Nice Flavour 6- Nutritional Value 7-Some Other Reasons 1 2 3 4 5 1 2 3 4 5 6 7 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 Dimension 1; Eigenvalue: .09416 (55.07% of Inertia) -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 D im en si on 2 ; E ig en va lu e: .0 49 45 ( 28 .9 2% o f I ne rt ia ) Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 331 It can be concluded on the basis of two- dimensional graphic representation of coor- dinates of rows and columns that there is a correspondence of the examinees with high school education with vegetable margarine purchase. Furthermore, pretty homogenous groups of the examinees with secondary education and high school education can be noticed in relation to the vertical axes of two-dimensional graphic representation, in terms of the type of margarine that is bought, whereas pretty homogenous groups of the examinees with primary education and high school education can be noticed in relation to the horizontal axes of the graphic (Figure 3). Figure 3:Two-dimensional plot of the coordinates of rows (level of education of examinees) and columns (margarine type) Source: Authors calculation based on questionnaire results Conclusions Based on the survey that has been carried out and statistical processing through the application of up-to-date statistical software STATISTICA, it has been established that over 70% of the examinees buy margarine with the frequency varying from once a week to once a month.The age of the exami- nees has displayed a statistically significant correspondence with the responses related to the main reason of using margarine (χ2 = 39.84 ;р = 0.022, 05.0 ), meaning that the main reason of using margarine is not independent of the age of the exami- nees.Household size has also displayed a statistically significant correspondence with the frequency of buying margarine (χ2 = 38.99 ; р = 0,028, 05.0 ), which indi- cates that the frequency of buying margarine is not independent of the number of house- hold members.Since a significant correspon- dence of modalities of previously crossed questions has been established on the basis of the derived method of correspondence analysis, the resulting conclusions indicate that this technique can be applied to the ex- amination of some performances of marga- rine as foodstuff. References Avilez, J. P., Escobar P., Von Fabeck G., Villagran K., García F., Matamoros R., & A. García Martínez (2010). Dairy farms productive characteriza- 2D Plot of Row and Column Coordinates; Dimension: 1 x 2 Row .Coords 1- Primary education 2- Secondary education 3- High School 4- Faculty Col.Coords 1- Dairy Margarine 2- Vegetable Margarine 4- Diet Margarine 1 2 3 4 1 2 4 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Dimension 1; Eigenvalue: .02089 (68.86% of Inertia) -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 D im en si on 2 ; E ig en va lu e: .0 09 45 (3 1. 14 % o f I ne rti a) Asian Journal of Agriculture and Rural Development, 4(5)2014: 324-332 332 tion using multivariate analysis meth- odology. Revista Científica, FCV- LUZ, 20(1), 74-80. Beh, E. J., Lombardo R., & B. Simonetti (2011). A European perception of food using two methods of corre- spondence analysis. Journal of Food and Quality Preference, 22, 226-231. Bendixen, M. (2003). A Practical guide to the use of correspondence analysis in marketing research. Marketing Bulle- tin, 14(2), 16-38. Chollet, S., Lelievre M., Abdi H., & D. Valentin (2011). Sort and beer every- thing you wanted to know about the sorting task but did not dare to ask. Food Quality and Preference, 22(6), 507-520. Detroja, K. P., Gudi R. D., Patwardhan S. C., & Roy K. (2006). Fault detection and isolation using correspondence analysis. Ind. & Eng. Che. Re, 45(1), 223-235. Green, P E., Carroll D., & A. M. Krieger (1987). Conjoint analysis and multi- dimensional scaling a complementary approach. Journal of Advertising Re- search, 27(5), 21-27. Greenacre, М. (2007). Correspondence analysis in practice (second edition), Chapman & Hall: Boca Raton, FL. Guerrero, L., Claret A., Verbeke W., En- derli G., Zakowska-Biemans S., Vanhonacker F., Issanchou S., & Marta (2010). Perception of tradi- tional food products in six European regions using free word association. Food Quality and Preference, 21, 225-233. Hair, J. F., Black W. C., Babin B. J., Ander- son R. E., & R. L. Tatham (2006). Multivariate data analysis, Pearson- prentice hall: New Jersey. Hoffman, D. L., & Leeuw J. D. (1992) In- terpreting multiple correspondence analysis as an MDS method. Market- ing Letters, 3(3), 259-272. Jobson, D. J. (1992). Applied multivariate data analysis –volume ii categorical and multivariate methods, Springer- Verlag: New York. Larsson, S. C., & Orsini N. (2011). Fish consumption and the risk of stroke a dose–response meta-analysis. Stroke, 42(12), 3621-3623. Liggett, R. E., Drake M. A., & Delwiche J. F. (2008). Impact of flavour attributes on consumer liking of Swiss cheese. J. Dairy Sci, 91, 466–476. Nicin, S. (2010). Possibilities of investiga- tion of cheese market applying corre- spondence analysis, Book of Ab- stracts, International Symposium Livestock Production, Veterinary Medicine and Economics in Rural Development and Production of Safety Food, Divcibare, 20-27. June 2010. Panea, B., Casasús I., Blanco M., & Joy M. (2009). The use of correspondence analysis in the study of beef quality a case study on Parda de Montaña breed. Spanish Journal of Agricul- tural Research, 7(4), 876-885. Pantelic, D., & М. Savic (2000). Application of statistical methods in marketing re- search. Аnnals of Faculty of Econom- ics, 5, 651-656. De Souza, R. V., Marques Pinheiro A. C., De Deus Souza Carneiro J., Pinto S. M., Abreu L. R., & Carvalho Mene- zes C. (2011). Analysis of various sweeteners in petit suisse cheese de- terminationof the ideal and equivalent sweetness. Journal of Sensory Stud- ies, 26(5), 339–345. Kleij, T. F., & Musters P. A. (2003). Text analysis of open-ended survey re- sponses: a complementary method to preference mapping. Food Quality and Preference, 14(1), 43-52. http://www.researchgate.net/researcher/71836658_Anna_Claret http://www.researchgate.net/researcher/40094192_Wim_Verbeke http://www.researchgate.net/researcher/2030588302_Geraldine_Enderli http://www.researchgate.net/researcher/2030588302_Geraldine_Enderli http://www.researchgate.net/researcher/2005937724_Sylwia_Zakowska-Biemans http://www.researchgate.net/researcher/44456066_Filiep_Vanhonacker http://www.researchgate.net/researcher/44456066_Filiep_Vanhonacker http://www.researchgate.net/researcher/9207764_Sylvie_Issanchou http://www.researchgate.net/researcher/31456681_Marta_Sajdakowska