- Email: [email protected]
Identification of Consumer Psychological Profile
Available online at www.sciencedirect.com
ScienceDirect Procedia Computer Science 102 (2016) 351 – 358
12th International Conference on Application of Fuzzy Systems and Soft Computing, ICAFS 2016, 29-30 August 2016, Vienna, Austria
Identification of consumer psychological profile R.N. Mikayilova* 6 Istiglaliyat, Baku 370001, Azerbaijan
Abstract One of the pressing topics of the marketing analysis, the problem of studying and modeling of consumer character, is examined. An approach to the problems is proposed, based on the modern concepts of cognitive psychology and new methods of personality psycho-diagnostics. The basic premises of the approach are formulated. The logical-linguistic formalism for modeling and identification of consumer psychological profile is developed. A description of an applied method of identification is given. The results of reliability and validity check of the method are discussed. 2016Published The Authors. Published Elsevier B.V.access article under the CC BY-NC-ND license © 2016 by Elsevier B.V.by This is an open (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of ICAFS 2016. Peer-review under responsibility of the Organizing Committee of ICAFS 2016 Keywords: consumer; psychological profile; identification; cognitive approach; logical-linguistic paradigm; personality theory; theory of psychological types; fuzzy sets; questionnaire test; repertory grid
1. Introduction Consumer behavior (CB) research is one of the most important and pressing problems in the free market economy. The marketing theory considers a qualitative model of CB that gives a general idea of the factors affecting the “black box” of consumer consciousness1,2. Practice, however, has long required more accurate and informative models that allow for “measuring”, studying and predicting CB. Many attempts at constructing mathematical models of CB yielded no practically significant results3. It became clear in the late 1980s that adequate modeling of CB within the framework of axiomatic theories and the classical calculus is impossible in principle, and the solution should be sought in the so-called cognitive approach4,5. This paper proposes a technique of identification of consumer psychological profile (CPP) as the essential and the most hard-to-measure component of CB. The technique was developed as part of mentioned studies.
*E-mail address: [email protected]
1877-0509 © 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of ICAFS 2016 doi:10.1016/j.procs.2016.09.411
352
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
2. Overall characteristic of the cognitive approach 2.1. Ideology and structure The foundation of the cognitive approach comprises: x cognitive metaphor—likening human cognitive activity to operation of computer systems; x applied personality theories determining the methodology of identification (knowledge acquisition) and conceptualization of a person’s psychological space—theory of personal constructs6,7, theory of psychological types8, implicit personality theory9, theory of psychological patterns10,11,12; x knowledge engineering, which provides tools for formalization and implementation of psychological models13. Cognitive methodology, like other methodologies of experimental sciences, is not axiomatic but hypotheticaldeductive14 and includes: x basic premises– postulates, hypotheses and principles of character modeling; x idealized object –a conceptual framework of the model constructed in compliance with the basic tenets; x core–concepts and definitions used in the conceptual and applied (working) models; x applied calculus – a deductive system that determines the logic of character identification based on various knowledge forms (expert, mathematical, experimental knowledge, etc.) and allows for easy correction of the model following the results of testing and operation. 2.2. Basic premises The axiomatics of the cognitive approach was formulated on the basis of an analysis of4,11 and included the following basic premises (P): P1. Consumer behavior is determined by consumer’s certain character traits and can therefore be predicted. P2. Consumer characteristics modeled as a complex of relatively stable traits (unchangeable or staying unchangeable for a long time) that are identified through observation, introspection or psycho-diagnostic testing. P3. Consumer character is modeled within the context of a problem at hand; instead of a universal model, a problem-oriented model is constructed, i.e. a model oriented “personalized marketing”. P4. The complex of character traits forms the implicit structure functioning on the principle of “psychological balance”—the condition of trait normalization. P5. In trade practices, a specific consumer’s character is identified within the domain of reference consumer types, whose stereotypical behavioral patterns are usually well known to experienced market workers. P6. Reference consumer types formed by an observer (a salesperson, an expert—an experienced marker worker) in practice result from personal reflection processes and are, to a varying degree, of virtual nature. P7. Reference consumer types are differentiated by clustering and weighing certain traits, with the weight factors reflecting the psychological role of a particular trait in the formation of a particular reference type. P8. Traits of a particular reference type are clustered based on the implicit personality theory8,12. This term mirrors a very specific experimental fact: upon learning that a person possesses some character trait named in the everyday language, practically any individual who is a native speaker of that language instinctively concludes that the person in question also possesses other traits associated with the known trait. P9. As a rule, consumer characters are not so pronounced that they comprise traits of only one reference type. They usually combine properties of different reference types, different characters sometimes possessing certain identical traits. P10. Categories of reference consumer types can be different in different segments of the consumer market. In terms of practical identification, the “hypothesis of reference principle of psychic reflection” (P5) and the “hypothesis of blurred outline of consumer character” (P9) are the most important ones among the basic premises. These hypotheses currently acknowledged by the world’s leading psychologists and sociologists drastically change the formulation of CPP identification problem and its solution methods. Instead of the probabilistic-statistical paradigm of CPP identification used in similar studies, the logical-linguistic paradigm based on the methods of knowledge engineering and fuzzy sets theory15 becomes predominant. This is reflected in the conceptual model of the subject area of CPP identification problem.
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
353
2.3. Conceptual model of the subject area of CPP identification problem We use the following basic concepts in the conceptual model. Consumer psychological type is the combination of consumer’s distinctive psychological characteristics that maintain relative coherence and consistency of consumer’s responses to the environment2,6. Attributes are consumer’s inherent character traits that can be used as a basis for describing his/her psychological type. Reference consumer psychological type is the most pronounced, “contrasting” (reference) consumer types of used to identify the psychological type of a specific consumer. The concept of reference types is widely used in psychology, sociology, ecology, decision-making theory, when one needs some benchmarks (stereotypes) to identify and predict processes that are difficult-to-formalize and phenomena of the real world. Applied model of consumer psychological type is a model oriented on the application in a specific segment of the consumer market in the capacity of a working model. The conceptual model of the subject area of CPP identification problem is represented by the four components: K = < T, G , A , R >,
{ } (ȡ 䌜 P = {1,2,...m}) is the term set of reference consumer psychological types (named images of the
where T = Tȡ0
most pronounced, contrasting consumer types occurring at the market, e.g. T = {T10 : “conceited”; T20 :“cautious”;
{}
T30 :“simple-minded”; T40 : “sensible”}31); G = IJ ȡ0
is the set of applied models of reference psychological
•
types; A is the set of attributes of a specific consumer; R is the procedure of fuzzy identification (classification) of consumer psychological type on the set G .In compliance with the hypothesis of blurred outline of character (P9), the procedure determines the degree of consumer’s membership in each of the reference types Tȡ0 䌜 T . This is an essential property of the procedure that distinguishes it from classic classification models, which cannot contain double classified objects, that is, objects that fell into two classes at the same time (principle of noncontradiction), and unclassified objects (principle of excluded middle). Thus, the compactness hypothesis in the procedure is irrelevant. 3. Applied model of consumer psychological type
An applied model is represented as a system of definitions that formalize the semantics of the concepts of the subject area and the logic of CPP identification. Definition1 .A frame-based structure13 of the following form will be referred to as an applied model (further on referred to simply as model) of the reference consumer psychological type IJ 0 :
(
n
) 䌥Ȝ i = ȁ ,
T 0 = {< (i ) ; a i ; Ȝ i > }; i = 1, n ; a i 䌜 A 0 ;
i =1
where T 0 is the name of a reference psychological type (RPT); (i) is the number of the i-that tribute of RPT, which is unique for a specific RPT; A 0 is the name of the i-th attribute of RPT; Ȝ i is the basic set of attributes of RPT; is the weight factor of the i-th attribute determining the psychological role of an attribute in the formation of RPT; Ȝ i = 1, L;
䌥Ȝ
i
= ȁ ; ȁ is the overall weight factor (OWF) RPT; ȁ > L .
The values and are established by expert judgement during constructing an applied model of RPT.
354
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358 n
Definition2. A model of RPT is referred to as normal if the condition
䌥Ȝ
* i
= 1 holds.
i =1
If a model of RPT is not in the normal form, then it can always be brought to the norm by dividing all values of weight factors by ȁ . m
Definition 3.The union A = U A ȡ0 will be referred to as the universal set of attributes of consumer model, with ȡ =1
A ȡ0 A 0Ș 䍴 ; ȡ 䍴Ș, i.e. different reference types can contain identical attributes.
Definition 4.The set of pairs of the form IJ = {< ȝ IJ ( IJ 0 ) / IJ 0 >}, where IJ 0 䌜 T, ȝ IJ ( IJ 0 ) 䌜 [0,1], will be referred to as the fuzzy consumer psychological type on the set . The function : T 䊻 [0,1]—the mapping of the set into the unit segment [0;1]—is called the membership function of the fuzzy type IJ; T is the term set of consumer RPT. For every specific RPT IJ 0 䌜 T , ȝ IJ (IJ 0 ) accepts a certain value from the closed interval [0;1]. That value is called the degree of membership of the type in the reference type IJ 0 . Let us assume that the set does not include pairs < ȝ IJ (IJ 0 ) / IJ0 > , for which ȝ IJ (IJ 0 ) = 0 .
{
}
Definition 5.The set PIJ = IJ IJ 䌜 T & ȝ IJ ( IJ 0 ) > 0 will be referred to as the carrier of the fuzzy psychological type
IJ. In other words, the carrier of the type is the subset PIJ of the term set T , for the elements of which the membership function ȝ IJ is strictly greater than zero. Definition 6.The fuzzy psychological type IJ is referred to as normal, if the condition sup ȝ IJ (IJ 0 ) = 1 holds. IJ䌜T If the fuzzy type is not normal, then it can always be made normal by dividing all values of the membership functions by its maximum value. According to Definition4, the recognition (identification) of the psychological type of a specific consumer
( 0)
requires determining the function ȝ c IJ of membership of the identifiable type in the term set T . The problem is solved by establishing the degree of membership of the type in each of RPTs. Let us introduce the definitions establishing the semantic identification rule that reflects the specifics of the problem. Definition 7.The overall weight coefficient ȁ IJ ( IJ 0 ) of the identifiable consumer type in the structure of the reference type W is the value n
( ) 䌥ıi
ȁIJ IJ0 =
Ȝi ,
i =1
where ıi is the Kronecker delta; ıi = 1 , if the i-th attribute of reference type IJ 0 is present among the consumer's identified attributes A*; ıi = 0 , otherwise.
Oi is the weight factor of the i-that tribute of the reference psychological type
IJ0 .
Definition 8.The aggregate overall weight factor of the identifiable fuzzy type IJ on the set of RPT (the term set m
T ) is the value ȁ IJ (T ) =
np
䌥䌥ı
pi
Ȝ pi ; ȡ 䌜 P; i 䌜 I ȡ
ȡ =1 i =1
where P = {1,2,...m} is the set of consumer RPT; I ȡ is the set of indices of the attributes of the U-th RPT.
( )
( )
( )
Definition 9.The value ȝ IJ T ȡ0 = ȁ IJ IJ ȡ0 / ȁ IJ (T ) will be referred to as the degree of membership ȝ IJ T ȡ0 of the identifiable type in the reference type .
355
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
4. Co onsumer psycho ological profilee
{ }
(ȡ 䌜 P = {1,2,...m}) is the basic term t set of RPT T; is the fuuzzy
Lett us a above thaat T = Tȡ0
psychoological type of a specific consumer; WU
^ P W /W !`; P is 0
0
WU
W
the carrier of thee type IJ 0
G
(Definnition5). We shaall also assume that ȡ R 1 , where w R 1 is the real axis. Lett us then arrangee the set of term ms of the carrier P IJ in line with h the expressionn
(䌔T
ȡ
)(
)
(
)(
)(
)
䌜 T 䌔TȘ 䌜 T (ȡ < Ș ) ˩ 䌗ȝ IJȡ 䌜 P IJ 䌗ȝ IJȘ 䌜 P IJ ȝ IJȡ > ȝ IJȘ ,
whhich means that the t term corresp ponding to the above a degree off membership acccepts a lesser number. n Deffinition 10. Let us refer to the model m of the fuzzzy psychologiccal type IJ ȡ arrannged by the estaablished rule as a connsumer psycholo ogical profile. Tabble 1 and figurre 1 below show examples off tabular and grraphic represenntation of consu umer psychologgical profilees, obtained thro ough a survey of o the “sale of ap partments” segm ment. Table 1. Taabular representation n of CPP Consumer
Consumer pssychological profilee,
T10
T20
1. M.A.I.
0,0
2. A.F.K. 3. I.U.G.
0,2 0,1
PW (W 0 )
T30
T40
0,35
0,55
0,1
0,6 0,2
0,1 0,0
0,1 0,7
Fig. 1. Graaphic representation n of CPP.
5. CP PP identificatio on technique
Thee basic premisees of the cognittive approach and a the logical--linguistic form malism given eaarlier were usedd for develooping the applied technique for fo identifying CPP. C In the prrocess of develoopment, some quite complex and extenssive work was done d at the unformalized heurisstic stages relateed to: identifying the term m set of consum mer RPT in partiicular segmentss of the consumeer market; connstructing the frrame-prototype of RPT; connstructing the frrame-example of o RPT; testting the reliabiliity and validity of the techniqu ue. Exttensive literaturre on trade psy ychology (one of o the first pub blications in thee field appeared d as far back aas in 1909— —Parsous F. Choosing C a vo ocation. London n. 1909), personality theoriees, psycho-diag gnostics, theoryy of psychoological types, psychosomatics p s and positive psychology, cognitive psycholoogy was used in this work. We also emplooyed the express diagnostics data that we had d obtained by means m of questionnaire tests ad dapted to particcular markeet segments: salee of apartments,, cars, computerr equipment, ou uterwear, buildinng supplies.
356
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
In accordance with the common practice of psychometrics, an attempt at testing the reliability and validity of the technique was made30. Some of the essential points of the performed research are outlined in the following paragraphs. The identification of term sets of RPT was carried out by means of typological questionnaire tests drawn up in collaboration with market workers with great experience in “personal sale”. Data from similar studies was also used19,22,23,27. During developing the frame-prototype of RPT, the structuring of consumer personality was the key issue. The model of “actual conflict”17 adapted to the trading conditions was taken as the structuring model. The model compactly but comprehensively mirrors the peculiarities of the trading process, in which the antagonistic interaction between a salesperson and a consumer—one of the form of “actual conflict”—is the psychological dominant. The model establishes four groups of attributes (superordinate constructs6) that reflect the basic forms of conflict processing by a consumer during trading: through body (“Activeness”), actions (“Activity”), contacts (“Communication”) and intuition (“Decision-making”). Processing “through body” means that conflict is experienced as a combination of physical sensations and reflects a person’s general activeness. Processing “through actions” means that actual conflict is resolved rationally, i.e. through the structures of consumer’s practical activity, by actualizing his/her cognitive abilities and consumer skills regarding specific types of goods and services. Processing “through contacts” reflects consumer’s psychological ability to establish and maintain interpersonal relations with a salesperson during trading. Processing “through intuition” reflects the peculiarities of consumer’s mental functions that determine his/her behavior at different stages of decision-making—perception of situation, product evaluation, decision-making, dependence on salesperson, response to failure. The principle of “psychological balance” in the model of actual conflict reflects the experimentally confirmable fact that preferential development of some forms of conflict processing leads to the reduction of other forms, i.e. hypertrophy of one response form pushes other forms into the background8,17. Each of the described response forms is determined by four bipolar attributes (subordinate constructs6). The universal set of subordinate attributes of (normal, not psychotic) consumers was identified in accordance with the trait theory1, which is currently fundamental in the study of consumer behavior. We also used the information from studies in trade psychology and the data we had obtained by means of personality questionnaire tests directly from the points of sale (specialty stores, supermarkets, exhibitions, real estate offices). Verbal labels of the attributes were chosen as is customary in the psychometric practice8,30. Namely, words and phrases were used that are present in the everyday language, close to market vocabulary and generally comprehensible, e.g. “familiar with the product and studies it carefully”—“unfamiliar with the product and does not study it or pretends to be studying it”, “makes rational decisions”—“makes emotional decisions”, “docile and accommodating”—“stubborn and unaccommodating”. Frame-examples of RPT (composition of the basic attributes, their weight factors, normalization conditions)were constructed by means of Hinkle’s implication grids7, which do a good job identifying generally accepted views of RPT structure in different market segments. We also used data from similar studies19,22,23,27. Following the requirements of the psychometric practice30, we attempted to evaluate the reliability (reproducibility of the results of CPP identification in time) and validity of the technique(reproducibility of the results of multi-level analysis: consumer’s personality from the viewpoints of specific groups (experts), a specific “other” and itself). There are currently no universal techniques of assessing reliability and validity of psychometric tools31. Assessment techniques are usually developed to fit objectives and conditions of a specific psychometric study. For this reason, we have developed an evaluation scheme that takes into account the specifics of the problem under investigation32. More than 70 subjects have been studied. 12 experts and 26 observers—persons who knew the subjects well enough (close relatives, coworkers, fellow students)—were involved in the study. To evaluate the reliability of the technique, the identification of CPP was performed twice, 6 months apart. The data obtained by means of personality questionnaires and repertory grids were used in the assessment of validity. Comparing the identification results, we confined ourselves to a simple arithmetic method7, because the use of the correlation analysis methods (Spearman’s rank method, Pearson’s method of mixed moments, etc.),which are
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
357
traditional in psychometrics, led to serious complications caused by the fuzzy nature of the results and their relationship with the situational background. The methods, forms of questionnaire and grids, which are currently in development, will, in our opinion, contribute to overcoming these challenges. Meanwhile, even the preliminary qualitative analysis has identified the presence of a correlation between the obtained results, which indicates good potential of this technique and of the approach in general. In the process of testing, we have obtained new useful information,which is to be used to improve the CPP identification technique for the first and foremost purpose of increasing its adequacy, usability and establishing the “range of convenience”1,6. 6. Conclusion
The proposed approach and developed conceptual framework do a fairly good job structuring the subject area of CPP identification problem. The introduced formalisms based on knowledge and fuzzy logic determine the scientific technology of studying and identifying CPP for different segments of the consumer market. Marketing specialists are usually well familiar with the stereotypical behavior of consumers with reference psychological types. For this reason, the CPP identification technique presented in the paper opens up new possibilities for solving a wide range of practical problems: x operational prediction of a specific consumer’s behavior; x choosing the line of behavior during the preparation and conduct of trade negotiations; x prevention of conflict situations during trading; x studying the statistical structure of consumer behavior in different market segments; x career guidance, professional advice and training trade personnel and students of higher educational institutions of trade and economic profile. This range of problems today, due to the mass nature of the trade sector and the socio-economic tension of the transition period, goes beyond particular psychometric research, also having a social context33. References 1. 2. 3. 4. 5. 6. 7. 8. 9.
Engel J, Blackwell R, Miniard P. Consumer behavior. Saint Petersburg, Moscow: PiterKom; 2000. Kotler Ph. Marketing essentials. New York: McGraw-Hill; 1991. Kinnear T, Taylor J. Marketing research: an applied approach. New York: McGraw-Hill; 1989. Caines B, Shaw M. Cognitive and logical foundation of knowledge acquisition. New York: North- Holland Publ. Co.; 1988. Hansen F. Consumer choice behavior. A cognitive theory. New York, London: Wiley; 1989. Kelly G. The psychology and personal constructs. New York: Norton Co.; 1955. Fransella F, Bannister D. A manual for repertory grid technique. Moscow: Progress; 1987. Jung C. Psychological types. Moscow: University book; 1998. Bruner JS, Tagiuri R. The perception of people. In: Lindzey G, editor. Handbook of social psychology. Reading, MA: Addison-Wesley; 1974, p. 634–654. 10. Zaporozhets AV, Zinchenko VP, Venger LA, Rusekon AG. Perception and action. Moscow: Science; 1977. 11. Zabrodin YM. On some trends of development of home psychophysics. Psychological journal 1982; 2: 55-69. 12. Kozlovsky SM. Role of references in psychophysical measurements. Issues of psychology 1985; 4: 102-9. 13. Ueno H, Ishizuka M (Eds.). Knowledge Representation and Usage. Moscow: Peace; 1989. 14. Ruzavin TI. Hypothetic-deductive method. Logic and empirical knowledge. Moscow: Science; 1972. 15. Zadeh L, Fu K, Tanaka K, Shimura M. Fuzzy sets and their applications to cognitive and decision process, in Proc. of the US̢Japan Seminar on Fuzzy Sets and their Applications, Berkeley, CA 1975, pp. 27-44. 16. Asmolov AG. Psychology of personality. Principles of general psychological analysis. Moscow: Science; 1990. 17. Peseschkian N. Psychosomatics and positive psychotherapy. Moscow: Science; 1996. 18. Hjelle L, Ziegler D. Personality theories. Saint Petersburg: ATP; 1997. 19. Fabiunke G, Grunewald O, Lehm J. Culture of commodity sales. Moscow: Ekonomika; 1973. 20. Yanovsky MI. Role of self-observation (introspection) in psychology. Issues of psychology 2001; 1: 91-6. 21. Blass J. Social psychology and personality: toward a convergence. Journ. of Pers. and Social Psychology 1984; 47; 27-41. 22. Buzzotte V, Lefton R, Sherberg M. Effective Selling through Psychology. New York: Psychological Associates; 1981 23. Character Discipline. 17th ed. El Paso: National Character Laboratory 24. Cattell R. The scientific analysis of personality. Baltimore: GPH; 1965. 25. Eysenk M, Keane M. Cognitive psychology: A student’s Handbook. Hillsdale: Erelbaum; 1993.
358
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358 26. Yogan R, Desoto C, Solano C. Traits, tests and personality research. Am Psychol 1977; 32(4): 255-264. 27. Levine R, Drang D, Edelson B. A comprehensive guide to ai and expert systems. New York: McGraw-Hill; 1991. 28. Melvin S, Hattwick Ph. The new psychology of selling. New York, Toronto, London: Wiley; 1980 29. Rosenzweig S. Investigating and appraising personality. In: Andews TG, editor. Methods of Psychology. New York: Wiley; 1984. p. 530568. 30. Anastasi A. Psychological testing. New York: Macmilan; 1982 31. Carlson JF, Geisinger K.F, Jonson JL. The ninth mental measurements yearbook. Lincoln, NE: Buros Institute of Mental Measurements; 2014. 32. Karayev RA, Mikayilova RN. Cognitive modeling of consumer behavior. Trans. ofANAS. Physics, Engineering and Mathematics 2001;2:47-52. 33. Mikayilova RN. Using methods of social psychology to predict and prevent conflict situations in trade. In Proc. of the International scientific practical conference “Influence of globalization processes on the formation of national labor markets and employment problem”, Baku, 2001, p. 111-6.
ScienceDirect Procedia Computer Science 102 (2016) 351 – 358
12th International Conference on Application of Fuzzy Systems and Soft Computing, ICAFS 2016, 29-30 August 2016, Vienna, Austria
Identification of consumer psychological profile R.N. Mikayilova* 6 Istiglaliyat, Baku 370001, Azerbaijan
Abstract One of the pressing topics of the marketing analysis, the problem of studying and modeling of consumer character, is examined. An approach to the problems is proposed, based on the modern concepts of cognitive psychology and new methods of personality psycho-diagnostics. The basic premises of the approach are formulated. The logical-linguistic formalism for modeling and identification of consumer psychological profile is developed. A description of an applied method of identification is given. The results of reliability and validity check of the method are discussed. 2016Published The Authors. Published Elsevier B.V.access article under the CC BY-NC-ND license © 2016 by Elsevier B.V.by This is an open (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of ICAFS 2016. Peer-review under responsibility of the Organizing Committee of ICAFS 2016 Keywords: consumer; psychological profile; identification; cognitive approach; logical-linguistic paradigm; personality theory; theory of psychological types; fuzzy sets; questionnaire test; repertory grid
1. Introduction Consumer behavior (CB) research is one of the most important and pressing problems in the free market economy. The marketing theory considers a qualitative model of CB that gives a general idea of the factors affecting the “black box” of consumer consciousness1,2. Practice, however, has long required more accurate and informative models that allow for “measuring”, studying and predicting CB. Many attempts at constructing mathematical models of CB yielded no practically significant results3. It became clear in the late 1980s that adequate modeling of CB within the framework of axiomatic theories and the classical calculus is impossible in principle, and the solution should be sought in the so-called cognitive approach4,5. This paper proposes a technique of identification of consumer psychological profile (CPP) as the essential and the most hard-to-measure component of CB. The technique was developed as part of mentioned studies.
*E-mail address: [email protected]
1877-0509 © 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of ICAFS 2016 doi:10.1016/j.procs.2016.09.411
352
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
2. Overall characteristic of the cognitive approach 2.1. Ideology and structure The foundation of the cognitive approach comprises: x cognitive metaphor—likening human cognitive activity to operation of computer systems; x applied personality theories determining the methodology of identification (knowledge acquisition) and conceptualization of a person’s psychological space—theory of personal constructs6,7, theory of psychological types8, implicit personality theory9, theory of psychological patterns10,11,12; x knowledge engineering, which provides tools for formalization and implementation of psychological models13. Cognitive methodology, like other methodologies of experimental sciences, is not axiomatic but hypotheticaldeductive14 and includes: x basic premises– postulates, hypotheses and principles of character modeling; x idealized object –a conceptual framework of the model constructed in compliance with the basic tenets; x core–concepts and definitions used in the conceptual and applied (working) models; x applied calculus – a deductive system that determines the logic of character identification based on various knowledge forms (expert, mathematical, experimental knowledge, etc.) and allows for easy correction of the model following the results of testing and operation. 2.2. Basic premises The axiomatics of the cognitive approach was formulated on the basis of an analysis of4,11 and included the following basic premises (P): P1. Consumer behavior is determined by consumer’s certain character traits and can therefore be predicted. P2. Consumer characteristics modeled as a complex of relatively stable traits (unchangeable or staying unchangeable for a long time) that are identified through observation, introspection or psycho-diagnostic testing. P3. Consumer character is modeled within the context of a problem at hand; instead of a universal model, a problem-oriented model is constructed, i.e. a model oriented “personalized marketing”. P4. The complex of character traits forms the implicit structure functioning on the principle of “psychological balance”—the condition of trait normalization. P5. In trade practices, a specific consumer’s character is identified within the domain of reference consumer types, whose stereotypical behavioral patterns are usually well known to experienced market workers. P6. Reference consumer types formed by an observer (a salesperson, an expert—an experienced marker worker) in practice result from personal reflection processes and are, to a varying degree, of virtual nature. P7. Reference consumer types are differentiated by clustering and weighing certain traits, with the weight factors reflecting the psychological role of a particular trait in the formation of a particular reference type. P8. Traits of a particular reference type are clustered based on the implicit personality theory8,12. This term mirrors a very specific experimental fact: upon learning that a person possesses some character trait named in the everyday language, practically any individual who is a native speaker of that language instinctively concludes that the person in question also possesses other traits associated with the known trait. P9. As a rule, consumer characters are not so pronounced that they comprise traits of only one reference type. They usually combine properties of different reference types, different characters sometimes possessing certain identical traits. P10. Categories of reference consumer types can be different in different segments of the consumer market. In terms of practical identification, the “hypothesis of reference principle of psychic reflection” (P5) and the “hypothesis of blurred outline of consumer character” (P9) are the most important ones among the basic premises. These hypotheses currently acknowledged by the world’s leading psychologists and sociologists drastically change the formulation of CPP identification problem and its solution methods. Instead of the probabilistic-statistical paradigm of CPP identification used in similar studies, the logical-linguistic paradigm based on the methods of knowledge engineering and fuzzy sets theory15 becomes predominant. This is reflected in the conceptual model of the subject area of CPP identification problem.
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
353
2.3. Conceptual model of the subject area of CPP identification problem We use the following basic concepts in the conceptual model. Consumer psychological type is the combination of consumer’s distinctive psychological characteristics that maintain relative coherence and consistency of consumer’s responses to the environment2,6. Attributes are consumer’s inherent character traits that can be used as a basis for describing his/her psychological type. Reference consumer psychological type is the most pronounced, “contrasting” (reference) consumer types of used to identify the psychological type of a specific consumer. The concept of reference types is widely used in psychology, sociology, ecology, decision-making theory, when one needs some benchmarks (stereotypes) to identify and predict processes that are difficult-to-formalize and phenomena of the real world. Applied model of consumer psychological type is a model oriented on the application in a specific segment of the consumer market in the capacity of a working model. The conceptual model of the subject area of CPP identification problem is represented by the four components: K = < T, G , A , R >,
{ } (ȡ 䌜 P = {1,2,...m}) is the term set of reference consumer psychological types (named images of the
where T = Tȡ0
most pronounced, contrasting consumer types occurring at the market, e.g. T = {T10 : “conceited”; T20 :“cautious”;
{}
T30 :“simple-minded”; T40 : “sensible”}31); G = IJ ȡ0
is the set of applied models of reference psychological
•
types; A is the set of attributes of a specific consumer; R is the procedure of fuzzy identification (classification) of consumer psychological type on the set G .In compliance with the hypothesis of blurred outline of character (P9), the procedure determines the degree of consumer’s membership in each of the reference types Tȡ0 䌜 T . This is an essential property of the procedure that distinguishes it from classic classification models, which cannot contain double classified objects, that is, objects that fell into two classes at the same time (principle of noncontradiction), and unclassified objects (principle of excluded middle). Thus, the compactness hypothesis in the procedure is irrelevant. 3. Applied model of consumer psychological type
An applied model is represented as a system of definitions that formalize the semantics of the concepts of the subject area and the logic of CPP identification. Definition1 .A frame-based structure13 of the following form will be referred to as an applied model (further on referred to simply as model) of the reference consumer psychological type IJ 0 :
(
n
) 䌥Ȝ i = ȁ ,
T 0 = {< (i ) ; a i ; Ȝ i > }; i = 1, n ; a i 䌜 A 0 ;
i =1
where T 0 is the name of a reference psychological type (RPT); (i) is the number of the i-that tribute of RPT, which is unique for a specific RPT; A 0 is the name of the i-th attribute of RPT; Ȝ i is the basic set of attributes of RPT; is the weight factor of the i-th attribute determining the psychological role of an attribute in the formation of RPT; Ȝ i = 1, L;
䌥Ȝ
i
= ȁ ; ȁ is the overall weight factor (OWF) RPT; ȁ > L .
The values and are established by expert judgement during constructing an applied model of RPT.
354
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358 n
Definition2. A model of RPT is referred to as normal if the condition
䌥Ȝ
* i
= 1 holds.
i =1
If a model of RPT is not in the normal form, then it can always be brought to the norm by dividing all values of weight factors by ȁ . m
Definition 3.The union A = U A ȡ0 will be referred to as the universal set of attributes of consumer model, with ȡ =1
A ȡ0 A 0Ș 䍴 ; ȡ 䍴Ș, i.e. different reference types can contain identical attributes.
Definition 4.The set of pairs of the form IJ = {< ȝ IJ ( IJ 0 ) / IJ 0 >}, where IJ 0 䌜 T, ȝ IJ ( IJ 0 ) 䌜 [0,1], will be referred to as the fuzzy consumer psychological type on the set . The function : T 䊻 [0,1]—the mapping of the set into the unit segment [0;1]—is called the membership function of the fuzzy type IJ; T is the term set of consumer RPT. For every specific RPT IJ 0 䌜 T , ȝ IJ (IJ 0 ) accepts a certain value from the closed interval [0;1]. That value is called the degree of membership of the type in the reference type IJ 0 . Let us assume that the set does not include pairs < ȝ IJ (IJ 0 ) / IJ0 > , for which ȝ IJ (IJ 0 ) = 0 .
{
}
Definition 5.The set PIJ = IJ IJ 䌜 T & ȝ IJ ( IJ 0 ) > 0 will be referred to as the carrier of the fuzzy psychological type
IJ. In other words, the carrier of the type is the subset PIJ of the term set T , for the elements of which the membership function ȝ IJ is strictly greater than zero. Definition 6.The fuzzy psychological type IJ is referred to as normal, if the condition sup ȝ IJ (IJ 0 ) = 1 holds. IJ䌜T If the fuzzy type is not normal, then it can always be made normal by dividing all values of the membership functions by its maximum value. According to Definition4, the recognition (identification) of the psychological type of a specific consumer
( 0)
requires determining the function ȝ c IJ of membership of the identifiable type in the term set T . The problem is solved by establishing the degree of membership of the type in each of RPTs. Let us introduce the definitions establishing the semantic identification rule that reflects the specifics of the problem. Definition 7.The overall weight coefficient ȁ IJ ( IJ 0 ) of the identifiable consumer type in the structure of the reference type W is the value n
( ) 䌥ıi
ȁIJ IJ0 =
Ȝi ,
i =1
where ıi is the Kronecker delta; ıi = 1 , if the i-th attribute of reference type IJ 0 is present among the consumer's identified attributes A*; ıi = 0 , otherwise.
Oi is the weight factor of the i-that tribute of the reference psychological type
IJ0 .
Definition 8.The aggregate overall weight factor of the identifiable fuzzy type IJ on the set of RPT (the term set m
T ) is the value ȁ IJ (T ) =
np
䌥䌥ı
pi
Ȝ pi ; ȡ 䌜 P; i 䌜 I ȡ
ȡ =1 i =1
where P = {1,2,...m} is the set of consumer RPT; I ȡ is the set of indices of the attributes of the U-th RPT.
( )
( )
( )
Definition 9.The value ȝ IJ T ȡ0 = ȁ IJ IJ ȡ0 / ȁ IJ (T ) will be referred to as the degree of membership ȝ IJ T ȡ0 of the identifiable type in the reference type .
355
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
4. Co onsumer psycho ological profilee
{ }
(ȡ 䌜 P = {1,2,...m}) is the basic term t set of RPT T;
Lett us a above thaat T = Tȡ0
psychoological type of a specific consumer; WU
^ P W /W !`; P is 0
0
WU
W
the carrier of thee type IJ 0
G
(Definnition5). We shaall also assume that ȡ R 1 , where w R 1 is the real axis. Lett us then arrangee the set of term ms of the carrier P IJ in line with h the expressionn
(䌔T
ȡ
)(
)
(
)(
)(
)
䌜 T 䌔TȘ 䌜 T (ȡ < Ș ) ˩ 䌗ȝ IJȡ 䌜 P IJ 䌗ȝ IJȘ 䌜 P IJ ȝ IJȡ > ȝ IJȘ ,
whhich means that the t term corresp ponding to the above a degree off membership acccepts a lesser number. n Deffinition 10. Let us refer to the model m of the fuzzzy psychologiccal type IJ ȡ arrannged by the estaablished rule as a connsumer psycholo ogical profile. Tabble 1 and figurre 1 below show examples off tabular and grraphic represenntation of consu umer psychologgical profilees, obtained thro ough a survey of o the “sale of ap partments” segm ment. Table 1. Taabular representation n of CPP Consumer
Consumer pssychological profilee,
T10
T20
1. M.A.I.
0,0
2. A.F.K. 3. I.U.G.
0,2 0,1
PW (W 0 )
T30
T40
0,35
0,55
0,1
0,6 0,2
0,1 0,0
0,1 0,7
Fig. 1. Graaphic representation n of CPP.
5. CP PP identificatio on technique
Thee basic premisees of the cognittive approach and a the logical--linguistic form malism given eaarlier were usedd for develooping the applied technique for fo identifying CPP. C In the prrocess of develoopment, some quite complex and extenssive work was done d at the unformalized heurisstic stages relateed to: identifying the term m set of consum mer RPT in partiicular segmentss of the consumeer market; connstructing the frrame-prototype of RPT; connstructing the frrame-example of o RPT; testting the reliabiliity and validity of the techniqu ue. Exttensive literaturre on trade psy ychology (one of o the first pub blications in thee field appeared d as far back aas in 1909— —Parsous F. Choosing C a vo ocation. London n. 1909), personality theoriees, psycho-diag gnostics, theoryy of psychoological types, psychosomatics p s and positive psychology, cognitive psycholoogy was used in this work. We also emplooyed the express diagnostics data that we had d obtained by means m of questionnaire tests ad dapted to particcular markeet segments: salee of apartments,, cars, computerr equipment, ou uterwear, buildinng supplies.
356
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
In accordance with the common practice of psychometrics, an attempt at testing the reliability and validity of the technique was made30. Some of the essential points of the performed research are outlined in the following paragraphs. The identification of term sets of RPT was carried out by means of typological questionnaire tests drawn up in collaboration with market workers with great experience in “personal sale”. Data from similar studies was also used19,22,23,27. During developing the frame-prototype of RPT, the structuring of consumer personality was the key issue. The model of “actual conflict”17 adapted to the trading conditions was taken as the structuring model. The model compactly but comprehensively mirrors the peculiarities of the trading process, in which the antagonistic interaction between a salesperson and a consumer—one of the form of “actual conflict”—is the psychological dominant. The model establishes four groups of attributes (superordinate constructs6) that reflect the basic forms of conflict processing by a consumer during trading: through body (“Activeness”), actions (“Activity”), contacts (“Communication”) and intuition (“Decision-making”). Processing “through body” means that conflict is experienced as a combination of physical sensations and reflects a person’s general activeness. Processing “through actions” means that actual conflict is resolved rationally, i.e. through the structures of consumer’s practical activity, by actualizing his/her cognitive abilities and consumer skills regarding specific types of goods and services. Processing “through contacts” reflects consumer’s psychological ability to establish and maintain interpersonal relations with a salesperson during trading. Processing “through intuition” reflects the peculiarities of consumer’s mental functions that determine his/her behavior at different stages of decision-making—perception of situation, product evaluation, decision-making, dependence on salesperson, response to failure. The principle of “psychological balance” in the model of actual conflict reflects the experimentally confirmable fact that preferential development of some forms of conflict processing leads to the reduction of other forms, i.e. hypertrophy of one response form pushes other forms into the background8,17. Each of the described response forms is determined by four bipolar attributes (subordinate constructs6). The universal set of subordinate attributes of (normal, not psychotic) consumers was identified in accordance with the trait theory1, which is currently fundamental in the study of consumer behavior. We also used the information from studies in trade psychology and the data we had obtained by means of personality questionnaire tests directly from the points of sale (specialty stores, supermarkets, exhibitions, real estate offices). Verbal labels of the attributes were chosen as is customary in the psychometric practice8,30. Namely, words and phrases were used that are present in the everyday language, close to market vocabulary and generally comprehensible, e.g. “familiar with the product and studies it carefully”—“unfamiliar with the product and does not study it or pretends to be studying it”, “makes rational decisions”—“makes emotional decisions”, “docile and accommodating”—“stubborn and unaccommodating”. Frame-examples of RPT (composition of the basic attributes, their weight factors, normalization conditions)were constructed by means of Hinkle’s implication grids7, which do a good job identifying generally accepted views of RPT structure in different market segments. We also used data from similar studies19,22,23,27. Following the requirements of the psychometric practice30, we attempted to evaluate the reliability (reproducibility of the results of CPP identification in time) and validity of the technique(reproducibility of the results of multi-level analysis: consumer’s personality from the viewpoints of specific groups (experts), a specific “other” and itself). There are currently no universal techniques of assessing reliability and validity of psychometric tools31. Assessment techniques are usually developed to fit objectives and conditions of a specific psychometric study. For this reason, we have developed an evaluation scheme that takes into account the specifics of the problem under investigation32. More than 70 subjects have been studied. 12 experts and 26 observers—persons who knew the subjects well enough (close relatives, coworkers, fellow students)—were involved in the study. To evaluate the reliability of the technique, the identification of CPP was performed twice, 6 months apart. The data obtained by means of personality questionnaires and repertory grids were used in the assessment of validity. Comparing the identification results, we confined ourselves to a simple arithmetic method7, because the use of the correlation analysis methods (Spearman’s rank method, Pearson’s method of mixed moments, etc.),which are
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358
357
traditional in psychometrics, led to serious complications caused by the fuzzy nature of the results and their relationship with the situational background. The methods, forms of questionnaire and grids, which are currently in development, will, in our opinion, contribute to overcoming these challenges. Meanwhile, even the preliminary qualitative analysis has identified the presence of a correlation between the obtained results, which indicates good potential of this technique and of the approach in general. In the process of testing, we have obtained new useful information,which is to be used to improve the CPP identification technique for the first and foremost purpose of increasing its adequacy, usability and establishing the “range of convenience”1,6. 6. Conclusion
The proposed approach and developed conceptual framework do a fairly good job structuring the subject area of CPP identification problem. The introduced formalisms based on knowledge and fuzzy logic determine the scientific technology of studying and identifying CPP for different segments of the consumer market. Marketing specialists are usually well familiar with the stereotypical behavior of consumers with reference psychological types. For this reason, the CPP identification technique presented in the paper opens up new possibilities for solving a wide range of practical problems: x operational prediction of a specific consumer’s behavior; x choosing the line of behavior during the preparation and conduct of trade negotiations; x prevention of conflict situations during trading; x studying the statistical structure of consumer behavior in different market segments; x career guidance, professional advice and training trade personnel and students of higher educational institutions of trade and economic profile. This range of problems today, due to the mass nature of the trade sector and the socio-economic tension of the transition period, goes beyond particular psychometric research, also having a social context33. References 1. 2. 3. 4. 5. 6. 7. 8. 9.
Engel J, Blackwell R, Miniard P. Consumer behavior. Saint Petersburg, Moscow: PiterKom; 2000. Kotler Ph. Marketing essentials. New York: McGraw-Hill; 1991. Kinnear T, Taylor J. Marketing research: an applied approach. New York: McGraw-Hill; 1989. Caines B, Shaw M. Cognitive and logical foundation of knowledge acquisition. New York: North- Holland Publ. Co.; 1988. Hansen F. Consumer choice behavior. A cognitive theory. New York, London: Wiley; 1989. Kelly G. The psychology and personal constructs. New York: Norton Co.; 1955. Fransella F, Bannister D. A manual for repertory grid technique. Moscow: Progress; 1987. Jung C. Psychological types. Moscow: University book; 1998. Bruner JS, Tagiuri R. The perception of people. In: Lindzey G, editor. Handbook of social psychology. Reading, MA: Addison-Wesley; 1974, p. 634–654. 10. Zaporozhets AV, Zinchenko VP, Venger LA, Rusekon AG. Perception and action. Moscow: Science; 1977. 11. Zabrodin YM. On some trends of development of home psychophysics. Psychological journal 1982; 2: 55-69. 12. Kozlovsky SM. Role of references in psychophysical measurements. Issues of psychology 1985; 4: 102-9. 13. Ueno H, Ishizuka M (Eds.). Knowledge Representation and Usage. Moscow: Peace; 1989. 14. Ruzavin TI. Hypothetic-deductive method. Logic and empirical knowledge. Moscow: Science; 1972. 15. Zadeh L, Fu K, Tanaka K, Shimura M. Fuzzy sets and their applications to cognitive and decision process, in Proc. of the US̢Japan Seminar on Fuzzy Sets and their Applications, Berkeley, CA 1975, pp. 27-44. 16. Asmolov AG. Psychology of personality. Principles of general psychological analysis. Moscow: Science; 1990. 17. Peseschkian N. Psychosomatics and positive psychotherapy. Moscow: Science; 1996. 18. Hjelle L, Ziegler D. Personality theories. Saint Petersburg: ATP; 1997. 19. Fabiunke G, Grunewald O, Lehm J. Culture of commodity sales. Moscow: Ekonomika; 1973. 20. Yanovsky MI. Role of self-observation (introspection) in psychology. Issues of psychology 2001; 1: 91-6. 21. Blass J. Social psychology and personality: toward a convergence. Journ. of Pers. and Social Psychology 1984; 47; 27-41. 22. Buzzotte V, Lefton R, Sherberg M. Effective Selling through Psychology. New York: Psychological Associates; 1981 23. Character Discipline. 17th ed. El Paso: National Character Laboratory 24. Cattell R. The scientific analysis of personality. Baltimore: GPH; 1965. 25. Eysenk M, Keane M. Cognitive psychology: A student’s Handbook. Hillsdale: Erelbaum; 1993.
358
R.N. Mikayilova / Procedia Computer Science 102 (2016) 351 – 358 26. Yogan R, Desoto C, Solano C. Traits, tests and personality research. Am Psychol 1977; 32(4): 255-264. 27. Levine R, Drang D, Edelson B. A comprehensive guide to ai and expert systems. New York: McGraw-Hill; 1991. 28. Melvin S, Hattwick Ph. The new psychology of selling. New York, Toronto, London: Wiley; 1980 29. Rosenzweig S. Investigating and appraising personality. In: Andews TG, editor. Methods of Psychology. New York: Wiley; 1984. p. 530568. 30. Anastasi A. Psychological testing. New York: Macmilan; 1982 31. Carlson JF, Geisinger K.F, Jonson JL. The ninth mental measurements yearbook. Lincoln, NE: Buros Institute of Mental Measurements; 2014. 32. Karayev RA, Mikayilova RN. Cognitive modeling of consumer behavior. Trans. ofANAS. Physics, Engineering and Mathematics 2001;2:47-52. 33. Mikayilova RN. Using methods of social psychology to predict and prevent conflict situations in trade. In Proc. of the International scientific practical conference “Influence of globalization processes on the formation of national labor markets and employment problem”, Baku, 2001, p. 111-6.