Research on Online Education Consumer Choice Behavior Path Based on Informatization

Yongrui Su,Ling Zhao

1 School of Economics and Management,Beijing University of Posts and Telecommunications,Beijing 100876,China

2 School of Marxism,Beijing University of Posts and Telecommunications,Beijing 100876,China

Abstract:The application of information and communications technology(ICT)in the education industry is becoming more and more extensive,and online education realized through ICT is developing in full swing.The influence of ICT on online education consumer’s choice behavior is the core issue of online education industry development research.The research on the interactive path and methods of information and online education consumer choice behavior is worth exploring and revealing.This study introduces the word-of-mouth factor as a new research variable under the framework of the Rational Choice Theory model(RCT)and the structural equation method to conduct empirical research and theoretical analysis to verify the validity of the hypothesis and model.The fifth-Generation mobile communication system(5G)analyses the factors affecting online education consumer behavior choices based on the premise of ICT.Research on the path between ICT and choice behavior provides new ideas for online education consumer choice behavior research and ICT and content and provides a new scenario.This article is a cross-disciplinary research content in theory,and its innovation opens up a new path for the management of ICT research.The research results have innovative significance and value at both the theoretical and practical levels.

Keywords:information and communications technology(ICT);consumer choice behavior;rational choice theory;the fifth-Generation mobile communication system(5G);online education

I.INTRODUCTION

The 45th“Statistical Report on the Development of China’s Internet”of the China Internet Network Information Center(CNNIC)clearly stated that online education shows explosive growth.As of March 2020,online education consumers in China have reached 423 million,an increase of 110.2%from the end of 2018,accounting for 46.8% of the total netizens.Despite the impact of the“COVID-19”,universities,middle schools and primary schools have implemented“stop classes without suspension”measures,and online education has become the primary teaching medium.[1]Data reflects that the development of online education in China is about to enter a period of rapid development.After the epidemic dissipates,online education,significantly higher education,will still be available to people as an essential means of learning.How to manage and apply this emerging service industry from the perspective of consumers is the key to the development of online education[2—4].The integration and development of information technology and online education have a history of more than 30 years.The earliest form of online education is distance education,which is spread through the wired network and educational practice in television and radio.This kind of distance education has been popular worldwide since the 1980s.The beginning of the online education model based on the internet is the Khan Academy establishment.The formal establishment of Khan college in 2009 marks that the online education mode has entered a new development model.We can find from the context of online education that it is constantly moving forward with the improvement of information technology and citizenship.

The current research online education can be roughly divided into five areas:platform construction,course research,market promotion,consumer behavior[5,6]and future development.Therefore,online education consumer’s choice behavior is the core issue of online education consumer’s behavior.However,the key to this core issue is the influence of ICT on choice behavior and information content on choice behavior.The research on consumer choice behavior still needs to be improved,so it is based on the rational choice theory(“Rational Choice Theory,referred to aspect”)[7]applied in this behavior.The leading factor research is a crucial way to explore the internal mechanism and influence path of the relationship between online education consumer’s choice behavior and ICT.It provides a more solid theoretical foundation for further empirical research.So the research on online education consumer choice behavior solves the theoretical and practical problems and provides an indispensable strategic planning resource for the development of the online education industry in the development of 5G.

The key contributions of this paper can be summarized as follows:The first is to combine rational choice theory with structural equation methods to analyze the factors that affect online education consumer’s choice behavior.This method brings a new perspective and experience to the study of online education user selection behavior.It can more comprehensively and realistically study the universality and particularity of consumer behavior.Secondly,the other is to prove that online education consumers’choice behaviors are studied from 5G business development.Thirdly the innovative introduction of word-of-mouth factors and verification of its establishment.With the gradual promotion and realization of 5G application scenarios,online education will have a new application in this era.Therefore,clarifying the generation mechanism and communication law of online education consumer choice behavior will provide a solid theoretical basis and scientific basis for the new application of online education.However,the current research on the choice behavior of online education consumers is still not perfect,and there is much room to make up for it.As an essential content of user behavior research in recent years,word-of-mouth factors have been widely discussed and studied.Online education relies on the internet for knowledge dissemination,which is bound to involve the influence and role of word-of-mouth factors in selecting knowledge content.However,there is still much space for the research of word-of-mouth factors in online education user selection behavior.Based on the above reasons,this paper wants to take this as an innovative perspective to study online education user selection behavior more deeply,comprehensively,and scientifically.It provides new ideas and methods for studying the interaction of information and choice behavior.Furthermore,the next one is the research of online education consumer choice behavior.Determining the influencing factors of the choice behavior is the most considerable difficulty to promote the research.Therefore,the rational choice theory and structural equation method can be used to research the online education consumer choice behavior,which can well limit the research boundary of the consumer choice behavior[8].

The research in this perspective accords with both the actual situation and the requirements of scientific demonstration[9].Rational choice theory is combined with the structural equation method applied to online education consumer choice behavior research,the theoretical analysis of online education consumer choice behavior factors.Propagation path exploration has paramount theory significance and the practical level of online education consumer choice guide and has a guiding role in the enterprise strategy making[10—13].The rest of this paper is organized as follows.Section II presents the Literature review.In section III,a hypothesis and experimental design are proposed.It was following that the research results and analysis are presented in section IV.Next is the discussion in section V.Section VI is the conclusion.Finally,insufficient are drawn in section VII.

II.LITERATURE REVIEW

2.1 Rational Choice Theory

The development of rational choice theory has roughly gone through three periods.Hosman[14]introduced rational choice theory into sociological research,introducing the rational principle of economic man to study exchange behavior in small groups,and this kind of research is the application of rational choice theory.Granovy conducted social network research in the 1970s,extending the rational choice theory to broader social networks.Until the 1990s,Coleman further improved and developed rationality the application of choice theory in sociology.In his book the foundation of“Social Theory”put forward the“individual-level theory of action”[7],Coleman put the three foundations of the concept of personal behavior:purpose,rationality,the pursuit of maximization of efficiency internalizes individual behavior into the internal analysis of the social system,trying to realize the system architecture of rational choice theory“from micro to macro”[15].

In the book“The Foundation of Social Theory”published in 1990,Coleman proposed that the rational choice theory is that individual actors will consider their value preferences and structural constraints when facing choices,and the ultimate choice is the most effective oriented action plan.

According to the literature,RCT research was applied to the exploration of politics,collectives,law,economy,resource allocation,and related fairness issues in the early stage[16—19].Around 2010,RCT research[20,21]gradually tended to research on daily life issues at the micro-level.RCT is individual research on the choice behavior of organizations and organizations that provides a theoretical basis,and many scholars have conducted in-depth research on the adoption and use behavior of information technology and information systems based on RCT[22].Therefore,using the RCT model to study online education consumer behavior can reflect the results of consumer behavior from an objective perspective and analyze the influencing factors that affect consumer’s choice behavior from an emotional level.Online education consumer behavior conforms to the characteristics of rational choice theory,which is mainly reflected in the goal that online education consumers want to achieve is to maximize their learning effect.Choosing an online education resource or product that suits them is the prerequisite for optimization.The irrational choice is difficult to meet this prerequisite,so it is rational to analyze online education consumers from the perspective of universality.Through quantitative and qualitative analysis can be better combined to enrich theoretical connotation in order to harvest more accurate and forward-looking research conclusions.The analysis of factors affecting the rational choice behavior of online education consumers needs to be scientifically confirmed from the perspective of mathematical modeling combined with the structural equation method(SEM).

2.2 Structural Equation Model Based on Rational Choice Theory(RCT-SEM)

Online education consumer choice behavior research has much room for supplementation stems from the three dilemmas of online education consumer choice behavior research.First,the factors that affect consumer choice are extensive and difficult to generalize,and the second is that the online education industry constantly changed.Consumer choice will continue to change with the development of the industry.Third,the subject of choice behavior research is“people”because people’s subjective initiative determines that people have unstable characteristics when choosing.However,choice behavior may have a potential regular and connection.The degree of connection between this connection and the regular is low,which is not enough to reflect the degree of internal connection.However,RCT provides a good breakthrough for this research.Based on the empirical analysis and deductive reasoning of online education consumer behavior,online education consumers abide by the principles of rational behavior when selecting online education resources,which conforms to the characteristics of rational choice behavior.Therefore,the use of RCT to analyze online education consumer choice behavior is in line with actual research needs.At the same time,RCT is not a closed-loop system of quantitative analysis.According to the needs of literature compilation and experimental research,we design this research model as follows:

This model introduces word-of-mouth factors based on the traditional RCT model.On the one hand,it is derived from many mature literature studies showing that“Consumers often refer to word-of-mouth information when purchasing products or services.Wordof-mouth has an important impact on consumers purchasing decisions.It has a better effect than other forms of advertising;word-of-mouth has a more significant impact on consumers than text,radio advertising”.[23—25]On the other hand,it stems from the particularity of online education resources.The core purpose of consumers choosing online education resources is to learn knowledge or skills.It is a spiritual aspect and not a material need.Therefore,the spiritual level test is by no means a measurement of a single price element.However,the information interaction between people in the community and then through individual value judgments can draw rational conclusions that meet their own value needs.In order to better verify that the RCT has theoretical and innovative significance for the research of online education consumer’s choice behaviors,the SEM is introduced to demonstrate the degree of matching and relevance of RCT in this scenario from the perspective of mathematics and sociology.At the same time to explore online education consumer’s choices,the main influencing factors of behavior provide analysis channels.Due to the diverse and complex influencing factors of online education consumers choice behavior,the research on online education consumers choice behavior based on RCT-SEM provides four entry levels for the selection of influencing factors:personal interests and value preferences,structural constraints,word-ofmouth factors,and rationality starting from the four dimensions of choice behavior,collecting and understanding the subjective wishes of the subjects through questionnaire surveys,using SEM to analyze the relationship between various factors and variables deriving the main influencing factors,combined with the RCT theoretical framework,researching and summarizing online education consumer choice.The main characteristics and conclusions of the behavior.

Based on our analysis and research,the rational choice behavior of online education consumers belongs to the category of rational choice theory.However,the more rational choice behavior will also be affected and induced by relevant factors.Exploring the main factors affecting the choice behavior of online education consumers is a vital prerequisite to clarify the development law of online education consumer’s choice behavior.The importance of influence factors on online education consumer choice behavior reflects in three aspects.First of all,influencing factors are the direct or indirect motives that influence online education consumer choices[26,27].To clarify the composition of influencing factors,better analyze the value orientation and behavior orientation of online education consumer choices.Secondly,the research conclusions can provide empirical better conclusions for online education consumer behavior spreading research[28—30].Analysis of influencing factors is the basis for studying online education consumer spreading behavior’s internal mechanism and infection path.Finally,the conclusions of the research on the influence factors of online education consumer choice behaviors can pave the way for academic research.

III.HYPOTHESIS AND EXPERIMENTAL DESIGN

Because the data in our research comes from questionnaires,the factors cognitions and reflections of online education consumer’s choice behaviors investigated are subjective perceptions of the subjects,and biases are inevitable.The SEM can incorporate multiple dependent variables and allow latent variables and measurement variables.It can also estimate the factor structure and factor relationship simultaneously,which is suitable for processing and analyzing the questionnaire data.Concerning the theoretical research of Qiu Haozheng and Lin Bifang,this paper establishes an SEM to study the selection needs of online education consumers[31].

According to the theoretical framework of rational choice and the actual research situation,our research summarizes and designs ten influencing factors,as shown in Figure 2,and scientifically demonstrates these ten influencing factors.For the relationship between them,reasonable assumptions are made in combination with relevant research,and the specific contents are described in the following chapters.

Figure 1.Online education consumer choice behavior research model.

Figure 2.Online education consumers choose behavioral research routes.

3.1 Personal Value Preference Hierarchy Design and Assumptions

3.1.1 The Impact of Individual Aspiration on Personal Value Preference

Our research hypothesis is constructed according to the rational choice theoretical framework,as shown in Figure 2.The development of the hypothesis follows a systematic scientific approach to exploring the relationship between measured variables and the relationship between measured variables and latent variables.Therefore,personal interests and value preference divide into individual aspiration,personal needs,the direct influence of advertising,and the indirect influence of advertising[7].Structural constraints are divided into purchasing power and social policy influence[32],and interpersonal word-of-mouth and authoritative word-of-mouth[25]innovation are included in the rational choice theoretical framework for measurement.

Figure 3.Research path of factors affecting online education consumers choice behavior.

Figure 4.Research path of factors affecting online education consumers choice behavior after adjustment.

When a person has a strong will for a certain behavior,he is more likely to implement the behavior[33].Before online education consumers choose online education products and resources,personal value preference is an important factor that affects their choice.The influence of personal will on personal value preference reflects in“Personal behavioral willingness reflects the implementation of certain specified behavior by humans”.[34]Personal will affect the choice of personal value preference.According to Maslow’s hierarchy of needs theory,Herzberg’s two-factor theory,it is explained that personal will is an internal driving force that produces a direct impact.The following hypotheses are proposed:

H1:Individual aspiration affects personal interests and value preferences.

3.1.2 The Impact of Personal Needs on Personal Value Preferences

For studying the level of personal needs,Maslow’s theory of personal needs is the best interpretation[35].Compared with personal wishes,personal needs occur before personal wishes are generated.Personal needs,as a human instinct,are generated by wishes.The direct basis is the fundamental source of behavior.The personal needs of online education consumers already belong to the spiritual value category sequence.This demand state is at the top level in the Maslow demand theory framework.Therefore,personal needs are more like the“source power”of personal value preferences.This hypothesis is proposed:

H2:Personal needs positively affect personal interests and value preferences.

3.1.3 The Impact of Advertising on Personal Value Preferences

As the name suggests,advertisements are widely advertised and are everywhere in daily life.In modern society,advertisements are divided into non-economic advertisements and economic advertisements.For online education consumers,they can learn about the characteristics of online education products and resources through advertisements.Educational advertising is roughly classified as economical advertising for discussion.Advertising itself is the transmission of information,but the effect of advertising can have a substantial impact on personal value preferences and behavior choices[39].Advertising effect refers to the effect obtained by consuming and occupying social labor in advertising activities[40].There are mainly the following three categories:communication effect;promotion of sales;changing consumer attitudes[41].These three main goals affect the personal value preferences of online education consumers and directly or indirectly improve themselves through external forms.For the cognition of selected online educational resources and products,this hypothesis is proposed:

H3:Advertising positively affects personal interests and value preferences.

(Advertising=direct influence of advertisement+indirect influence of advertisement)

H4:The direct impact of advertising is more significant than the indirect impact.

3.2 Word-of-mouth Hierarchical Design and Assumptions

Consumers often refer to word-of-mouth information when purchasing products or services,and word-ofmouth has an essential impact on consumers purchasing decisions[23];research results show that wordof-mouth has a more significant impact on consumers than text,radio advertising[42,25].One of the key points and innovations of this paper is introducing word-of-mouth factors into the RCT model and verifying the influence of word-of-mouth factors on the rational choice behavior of online education consumers through SEM.Combined with empirical research,online education consumers who choose online educational resources or products are affected by word-ofmouth factors as much as the experience after use.So word-of-mouth factors are an essential measurement indicator for consumers when choosing online education products.According to the actual needs of the research content in our research,the word-of-mouth level is divided into two indicators consider:Interpersonal word-of-mouth and Authority word-of-mouth;therefore,the following assumptions are drawn:

H5:Interpersonal word-of-mouth positively affects rational choice behavior.

H6:Authority word-of-mouth positively affects rational choice behavior.

3.3 Structural Constraints Hierarchical Design and Assumptions

In Coleman’s“Social Theoretical Foundation”,structural constraints are summarized as market structure,authority structure and trust structure.Economic mechanisms play a leading role in the market structure.In the authority structure and trust structure,power or strength,trust,and social norms play an essential role.Of course,some scholars believe that structural constraints determine and restrict the position and function of the units inhabited by the“potential logic and deep order of the system”.In general,the rules of specific social relations and the potential logic of policies are particularly emphasized—implementation,development,and change of social relations and the impact of specific actors[43].The structural constraints in online education consumers’choice behaviors are prominent.They are restricted by the economic purchasing power structure and social factors when choosing online education resources and products.

3.3.1 The Influence of Purchasing Power on Structural Constraints

Purchasing power is also called“consumer purchasing power”.Refers to the sum of the abilities necessary for consumers to complete the purchase of commodities[36].Purchasing ability is the root of consumers purchasing behavior.Online education products have different types and prices.Consumers will have doubts when choosing online education products.After all,this“product”is a virtual and spiritual item.It needs to be transformed through practice and media;purchasing power is an economic constraint in structural constraints,and purchasing power is at the core of economic constraints.Therefore,the following hypotheses are proposed:

Structural constraints Q19[38,7]Q19.1 How subject to policy structure when choosing an online education product?Q19.2 How socially constrained are you when choosing an online education product?Q19.3 How economicly constrained are you when choosing an online education product?Personal interests and value preferences[7]Q20 Q20.1 What is your overall assessment of the online education products you use?Q20.2 Do you think the cost of online education products is worth it?Q20.3 Do you think using this online education product has met your expectations?Social and policy implications[7]Q21 Q21.1 How does your job needs affect your choice of online education products?Q21.2 How does the national policy affect your choice of online education products?Q21.3 How does the unit policy affect your choice of online education products?Q21.4 How important is this dimension to you when you choose an online education product that teaches knowledge and skills that meet the needs of social development?Q21.5 If an online education product is widely used by the public,how important is it to get you interested?

H7:Purchasing power positively affects structural constraints.

3.3.2 The Impact of Social and Policy Factors on Structural Constrints

Social and policy factors are important influencing factors in structural constraints and the most direct and clear objective influencing factors.As individuals in society,consumers have to adapt to the needs of social development and follow the trend of social development.Online education consumers’ choice of online education resources and products results from consideration of social and policy factors.Through online education,consumers can continuously learn and continuously improve their comprehensive strength,but the content and direction of learning will be affected.The promotion and restriction of social and policy factors,so hypotheses are put forward:

H8:Social and policy influences positively affect structural constraints.

3.4 Hierarchical Design and Assumptions of Rational Choice Willingness

According to the above discussion,the original RCT model only includes personal interests,value preferences,and structural constraints.Our research innovatively embeds these factors into the RCT model to form word-of-mouth with a rational choice model.So two hypotheses H5 and H6,are proposed;in order to verify whether the RCT-SEM is valid and whether the effect is significant,the following hypotheses are proposed:

H9:Personal interests and value preferences positively affect rational choice behavior.

H10:Structural constraints positively affect rational choice behavior.

When sorting out the above measurement indicators,we can roughly summarize them into four categories:personal factors,advertising factors,social factors,and word-of-mouth factors,among which personal factors include:personal wishes,personal needs,and purchasing power;advertising factors reflect in advertising direct influence and indirect influence of advertising;word-of-mouth factors include:interpersonal word-of-mouth and traditional word-of-mouth;social factors are social and policy influences.In the final analysis,our research is studying the influence of the above four factors on online education consumers choices,and therefore assumption:

H11:The whole model confirms the influence of personal factors,social factors,advertising factors,and word-of-mouth factors on rational choice behavior.

IV.RESEARCH RESULTS AND ANALYSIS

4.1 Descriptive Statistics

Our research will combine rational choice theory and the structural equation method(SEM)to expole how research and analysis affect online education consumer choice behavior.RCT occupies a significant position in economics and sociology,but the analysis of quantitative empirical problems is minimal.Therefore,the combination of RCT and SEM can analyze the research content and problems from both qualitative and quantitative fusion.[44,16]In addition to the Likert 5-point scale for the measurement questions in Table 1,includes descriptive statistical questions,screening questions,quantitative statistical questions,and subjective questions in order to more comprehensively reflect the fundamental choices of online education consumers from multiple angles and methods.Despite the diverse characteristics and complex types of online education consumers,it is still difficult to cover everything.Therefore,our research limits the age of the participants to 18 years old and below 70 years old.The types of online education for young people,such as K12,are outside the scope of our research.The educational level ranges from college to doctoral students,looking for similarities in online education consumers of different ages,genders,and educational backgrounds.Draw the research roadmap of our research based on experimental hypotheses.

Table 1.Online education consumers to select behavior measurement scales.

The questionnaire distribution platform for our research is Credamo.It was released on September 22,2020.The collection was completed on October 10.A total of 1804 questionnaires were collected.Moreover,three rounds of pre-experiments were conducted before the questionnaires were issued.Next,the number of pre-experimental questionnaires was 300.This 300pre-experimental questionnaires are to test whether the validity and reliability of the questionnaire meet the standard,whether the model structure design is reasonable and in line with the use method of the structural equation.Seven hundred thirty-four valid questionnaires,770 invalid questionnaires,valid questionnaires accounted for 48%(excluding pre-experimental questionnaires).Using IBM SPSS 24.0 to analyze the sample and provide descriptive results,the age composition of the subjects is shown in Table 2.Among them,seven missing persons did not fill in the specific age or fill in the text characters during the filling process,so the statistics are missing.The completeness and authenticity of the questionnaire are guaranteed.According to the statistical results,19 to 30-yearolds are the main groups to fill in this questionnaire.Because the questionnaire sets the screening option“Have you ever used online educational resources and products?”,selecting non-consumers is not the target of this survey.So 19 years old people in their 30s are also concentrated groups using online education products or resources.

Table 2.Descriptive statistics of respondents’characteristics.

According to Table 2,it can be more clearly reflected in this survey that the age distribution of consumers who use online education products or resources.It shows a reverse distribution since the age of 19,the fewer online education products and resources are used as future.However,by no means are still many middle-aged and older people who are also consumers of online education products and resources.The minimum age group for this survey is 18 years old and above.Online education consumers under 18 years old are minors.Most of their behaviors in choosing online education platforms and resources should follow the opinions of parents.It is not the research of this paper to meet the needs of primary education.The focus is therefore not within the scope of research.

Table 3.Scale properties of the measurement model.

Among the 734 valid questionnaires in the questionnaire survey,there were 311 male subjects,accounting for 42.4%;423 female subjects 57.6%.Educational background distribution:the maximum number of candidates with a bachelor’s degree is 360,accounting for 49%;the second with a graduate degree is 289,accounting for 39.4%.

The questionnaire was distributed to the whole society through the Credamo platform,and 1,804 questionnaires were collected from 35 regions of the world,proving that the survey’s scope is relatively broad and sufficient.The number of items in this questionnaire is 69,and the number of valid questionnaires is 734.Although there is no precise regulation on the necessary sample size using the SEM[38],some scholars have stated that the ratio of the sample size to the estimated parameters should be 5:1 or 10:1[32].The research in this paper focuses on online education consumer behavior.The research coverage is relatively comprehensive,and consumer behavior is diverse and complex.Therefore,the sample size sets to about ten times the item.The survey results are expected to be more accurate and reliable;the recovery of 734 valid questionnaires meets the calculation requirements of the SEM,which paves the way for the correctness of the conclusions of the paper.

4.2 Measurement Model Evaluation Reliability and Validity

4.2.1 Model Checking

In terms of model verification,partial least squares structural equation modeling(PLS-SEM)is used to verify the research model[45,46].Compared with covariance-based structural equation modeling(CBSEM),covariance-based structural equation modeling(CB-SEM)requires the smallest measurement scale,sample size and residual distribution[47],and PLSSEM relaxes the assumption of normal distribution[31].The model established in this paper is complex and includes many structures,indicators,and relationships,which means that the PLS-SEM is more suitable for use[48].In addition,this study adopts a predictive perspective for analysis.PLS-SEM maximizes the variance of endogenous variables explained by exogenous variables[49],and the focus of prediction is suitable for achieving the research goals.Based on the above reasons,this article uses the PLS method for research.According to the recommendations of Anderson and Gerbing[50],data analysis adopts a two-stage process:measurement model and structural model evaluation.Measurement model evaluation checks reliability and validity,while structural model evaluation allows hypothesis testing by examining the relationships between constructions.

In order to verify the measurement model,the measurement model evaluated the reliability,average extraction amount of variation,and discriminative validity.The study used Joreskog’s[51]combined reliability and Cronbach’s alpha to evaluate the internal consistency reliability.Bagozzi,according to Yi(1988)’s suggestion,the combined reliability value should reach 0.60.However,the classical test theory holds that the scale’s reliability should reach 0.70,and Nunnally(1978)believes that Cronbach’s alpha coefficient value should be at least greater than 0.5 in practical applications.It is more significant than 0.7,so to study more rigorously,this paper adopts the combined reliability and Cronbach’s alpha coefficient value of 0.70.For the average variation extractionamount(AVE),this paper adopts the judgment of Anderson &Gerbing.The AVE criterion is 0.50[50].When the AVE value is greater than 0.50,it indicates that the convergence ability of the latent variable is ideal;at the same time,the questionnaire and validity standards are carried out in the research object.The degree of correlation between the measured results is usually low,mostly between 0.20 and 0.60,rarely exceeding 0.70,generally between 0.4 and 0.8,which is ideal[52].With the development and advancement of research,it is more appropriate to define the upper limit of the ideal threshold of validity coefficient as 0.9 or less.The judgment of discriminative validity is based on the fact that each variable should be greater than the fair value of the correlation coefficient between each pair of variables[52].There is a pair of variables with a correlation coefficient greater than any of the variables.The amount of variation extracted from a term means that in the pair of variables,the measurement item of one variable may also be the measurement item of another variable[50].That is,the test fails,and the model is not valid.

Based on the above theory,three rounds of preexperiments were conducted on the model established in this paper,and it found that the Q9 series of questions,Q12.1,2,a Q14.4,5,Q15.4,and Q21.5 were unqualified.Achieve the desired goal;since the Q9 series of questions failed the test,the H1 hypothesis is not valid,and the adjusted research path is obtained.The revised drawing is shown in Figure 5:

4.2.2 Reliability and validity test

According to model testing and pre-experiment adjustments,SmartPls3 software is used to test the reliability and validity of the data.The test results are shown in the following figure:

According to data analysis in Table 3,the combined reliability in this study is between 0.876 and 0.929,and all Cronbach’s alpha values are above 0.724;Table 6 shows that all factor loading coefficients are above 0.7 and Wixom and Watson proposed that factor loading exceeds 0.5[53].The average coefficient of variation of each structure exceeds 0.5,which indicates that the structure explains at least 50%of the variance of its items as shown in Table 4.In order to evaluate the discriminative validity[47],this article uses more than 0.7 to indicate that the item set well,reflecting the individual construct is remarkable.At the same time,the study tested the correlation ratio between the Fernell-Lake coefficient and the single heterogeneous trait(HTMT).As shown in Table 4,the correlation coefficients between one construct and other constructs are all less than the square root of the AVE of the construct[49],Which verifies the validity of the model[54].The verification method of discriminative validity has been innovated in recent years.Voorhees et al.proposed a related heterogeneity-single trait ration[55].The smaller the HTMT test value,the higher the discriminative degree and the difference between the two constructs,as shown in Table 5.In principle,the HTMT cannot be greater than 0.85.The maximum HTMT value in this paper is 0.72,which meets the verification standard and has blatant discrimination,indicating that all the model constructions have established discrimination validity.

4.3 Model Evaluation

Before evaluating the structure,the structural model should check its collinearity to ensure that the regression results are not biased[31].The collinearity statis-tic should be lower than 3.The external VIF and internal VIF of this study are both lower than 3.Therefore,collinearity is not the subject of this study,and we show Table 7 and Table 8 for the analysis results.

Table 4.Differential validity of the questionnaire(Fornell-Luck criterion).

In order to test the hypothesis,the structural model was checked using the bootstrapping technology of the SmartPls software specifying 5000 sub-samples.In the structural model analysis,the focus is to determine the significance and relevance of each hypothesis path and the explained variance.The model explains that the variance of personal interest and value preference is 14%,the variance of structural constraints is 27.8%,and the variance of rational choice behavior is 11.9%.Personal needs and personal interest value and value preferences have a significant positive effect(B= 0.292,P＜0.001);the direct impact of advertising and personal interests and value preferences have a significant positive impact(B=0.132,P＜0.01).The indirect impact of advertising and personal interests and value preferences have a significant positive impact(B = 0.091,P＜0.05);personal interests and value preferences and rational choice behavior have a significant positive impact(B= 0.091,P＜0.05).Social and policy influences and structural constraints have a significant positive impact(B= 0.456,P＜0.001),purchasing power and structural constraints have a significant positive impact(B= 0.221,P＜0.001),and structural constraints and rational choice behaviors are significant Positive influences(B=0.238,P＜0.001).Authoritative word-ofmouth transmission and rational choice behavior have a significant positive effect(B=0.106,P＜0.05).Interpersonal word-of-mouth transmission and rational choice behavior have no significant effect because P=0.238 is greater than the significant upper limit of Pvalue 0.05.So interpersonal word-of-mouth transmission has no obvious effect on rational choice behavior and further proves that H5 is not valid.According to the comprehensive analysis,the previous assumptions are valid except for H1 and H5.Table 9 and Figure 5 show the results of the structural model:

Figure 5.Path diagram of structural model results.

Table 5.Discriminant validity:Heterotrsait-monotrait(HTMT).

Table 6.External factor loading factor.

Table 7.VIF of external model.

Table 8.VIF of internal model.

Table 9.Structural model assessment for direct and indirect effects.

In the previous article,by summarizing and mea-suring the characteristics of the indicators,the indicators are divided into four categories:personal factors,advertising factors,social factors,and word-ofmouth factors.Therefore,H11 is proposed in the entire model to confirm that personal factors,social factors,advertising factors,and word-of-mouth factors are all for rational choice the influence of behavior.Combined with the model analysis results,the H1 and H5 hypotheses are not valid.However,personal factors include personal wishes and personal needs,the direct impact of advertising,and the indirect impact of advertising.Word-of-mouth factors include not only interpersonal word-of-mouth transmission but also authority word-of-mouth transmission.Other hypothetical factors that are immediately measured by Juncheng impact the rational choice behavior of online education consumers.Therefore,the argument H11 is established.Based on the model analysis results,we can conclude that personal factors,social factors,advertising factors,and word-of-mouth factors are the choices of online education consumers—the main factors affecting online education products and resources.

V.DISCUSSTION

Online education consumer choice behavior research based on the rational choice theory(RCT)treats online education consumers as“rational people”from a theoretical level[56].The conclusions of this research prove from an empirical level that online education consumers follow the rational choice theory when they choose online education products or resources.The internal mechanism and conforms to its external development law.Studies have proved that in the field of personal factors,personal needs have a significant positive impact on personal interests and value preferences[57].From this,it can be inferred that online education consumers should only consider from a personal perspective when choosing online education products and resources.What we consider is our own needs.At the same time,we are also pleasantly surprised to find that advertising factors can indeed positively affect the choices of online education con-sumers.Therefore,the reasonable effective and extensive publicity of online education products and resources is bound to arouse more online education consumers attention.Interest,but from another perspective,when online education consumers choose online education products or resources,the ways to understand product information and content are still not rich enough.The emerging industry of online education has emerged,but its essential products and core content have demand the popularity of consumers and the general public still needs to be further improved.

Although purchasing power belongs to the category of personal factors,research shows that it has a very significant positive impact on structural constraints[58].This conclusion reflects that knowledge is priceless,and learning knowledge is costly.So as online in the development of the education industry,whether it is the main body of the enterprise or the consumer object,the objective factors of purchasing power should be considered rationally and sincerely when choosing online education products.As an essential measurement standard,social factors have been demonstrated in the research to impact online education consumers’rational choice behavior.On the one hand,it reflects that continuous learning in the current society is gradually becoming prominent.

On the other hand,it reflects the independent learning of the public.Motivation still needs to be strengthened.If online education is regarded as the main channel of continuous learning and innovative learning,the entire society’s cultural atmosphere,and knowledge base will be improved by leaps and bounds,thereby promoting the overall development of various fields.The addition of word-of-mouth factors has innovatively enriched the connotation of the theoretical selection model and provided important factor indicators for studying online education consumer’s choice behavior[59].However,the interpersonal word-ofmouth transmission factor failed the structural test.The authoritative word-of-mouth transmission has an evident influence on the rational choice behavior of online education consumers,which proves that the word-of-mouth factor has a significant influence on the rational choice behavior[60,61].There has been an endless stream of research on internet word-ofmouth in recent years,and genuine and effective internet word-of-mouth should be attributed to traditional word-of-mouth.Whether it is comments,scoring,orother forms of online word-of-mouth when it is publicly presented,it not only represents personal subjective opinions.It also represents the objective feelings of consumers after use research has shown that compared with interpersonal.

Regarding the question of the failure of the H1 and H5 hypotheses,we believe that the main reason for the failure of H1 is due to the similarity in the dimensions of understanding between individual aspiration[62]and personal needs[63],there is a significant difference in the academic argument,but the subjects are regarded as age and education level.Different groups may not clearly distinguish the difference between the two when filling in the questionnaire,and they are more focused on personal needs in terms of cognition.Of course,they may also be affected by the characteristics of the test group and various subjective factors,resulting in personal willingness.Did not pass the inspection H5 failed the test.We believe that there is a big gap between the influence of word-of-mouth transmission among people and the influence caused by the transmission of authoritative word-of-mouth,which is embodied in three aspects:First,in addition to acquaintance factors,most online education users.It is impossible to communicate and trust directly,so the strength of interpersonal word-of-mouth transmission will be weakened[64];secondly,compared with authoritative word-of-mouth,the emotional tendency brought by interpersonal word-of-mouth will be affected by the environment and special time,so it is relatively objective[65,66].In the end,online education users,as“rational people”,are analyzed from the cognitive level.Compared with interpersonal word-ofmouth,they will believe in the influence of authoritative word-of-mouth.Therefore,the transmission of interpersonal word-of-mouth is not pass the inspection.

The addition of word-of-mouth factors has innovatively enriched the connotation of the theoretical selection model and provided important factor indicators for studying online education consumer’s choice behavior.However,the interpersonal word-of-mouth transmission factor failed the structural test.The authoritative word-of-mouth transmission has an evident influence on the rational choice behavior of online education consumers,which proves that the word-ofmouth factor has a significant influence on the rational choice behavior.It also represents the objective feelings of consumers after use;research has shown that compared with interpersonal word-of-mouth transmission,online education consumers pay more attention to authoritative word-of-mouth factors when choosing products or resources.However,the establishment of authoritative word-of-mouth depends entirely on the product or resource itself.The actual situation to effectively improve product quality and establish an excellent after-sales service system is the basis for enhancing its comprehensive competitiveness.

VI.CONCLUSION

The conclusions are as follows:

First,breaking through the traditional research model,combine Rational Choice Theory(RCT)with Structural Equation Modeling(SEM)to analyze and derive the main influencing factors affecting online education consumers’ choice behavior.Through empirical tests,raise and prove the establishment of the model and draw reliable conclusions.The influencing factors are classified and summarized as:personal factors,social factors,advertising factors,and wordof-mouth factors.

Second,based on the RCT to research online education consumers’ choice behavior,online education consumers choose behavior to conclude.Personal interests,value preferences,structural constraints,and word-of-mouth are essential aspects that affect online education consumers’choices.

Third,innovatively build and introduce the wordof-mouth factor into the RCT-SEM,and verify their establishment.The analysis shows that the word-ofmouth factor is an important indicator that affects online education consumer’s choice behavior.

Fourth.The conclusions of our research on the influencing factors of online education user selection behavior can better lay a theoretical foundation for online education consumer behavior research.The influencing factors of online education consumer choice behavior directly affect the mechanism and path of online education consumer spreading behavior.Research conclusions can better reveal the application of communication dynamics in online education consumer spreading behavior.

The theoretical significance of this research has three aspects.Firstly,unlike traditional research using SEM,this paper combines RCT with SEM to study online education consumer choice behavior.It is a new theoretical attempt.Studies have confirmed that the newly integrated RCT-SEM meets the verification standards,laying a theoretical and empirical basis for the future use of RCT-SEM to analyze similar problems.

The second is to research and analyze online education consumers’choice behaviors from rational choice theory and combine the structural equation method.It finds the main factors affecting online education consumers,revealing the overall path of online education consumer’s choice behavior development.The use of RCT provides a clear conceptual boundary for online education consumer choice behavior research.Under the RCT perspective,through the micro-level empirical exploration to make up for the lack of this area and carry out rigorous and solid scientific theories,Systematic research provides innovative supplements at the theoretical level and guides at the practical level.

Thirdly,enrich and perfect the theoretical content of consumer behavior and management through the combination of RCT and SEM.This research is distinguished from other consumer behavior studies,specifically as follows:1.Online education consumers are“rational people”;2.Online education consumer’s benefits are potentially unable to be“realized”immediately;3.Online education consumer’s behavior goals are clear,and the random ones are rarely generated.Therefore,analyzing the main influencing factors of online education consumer choice behavior has universal and guiding value for the study of online education consumer behavior and supplements consumer behavior research.For management,the main influencing factors can enrich organizational management and behavior management theory and can lead to innovative management paths for management optimization and strategic management.

What needs to be further added here is that the signs mentioned in the conclusion are a specific description of the contribution of this paper in the introduction,which is a specific elaboration from the perspectives of theory and practice.The explanation of the significance of this part of the research is also better to show the foresight and innovation.Our research can better guide online education enterprises to formulate more scientific and reasonable strategies and policies,provide a public reference standard and basis for online education consumers,and guide online education users to make reasonable and correct choices.The practical significance of this study is as important as the theoretical level.It is the actual content that urgently needed to be supplemented in developing the current online education industry.

VII.INSUFFICIENT

Although this paper has made many innovative attempts,some shortcomings cannot be avoided.The RCT is mainly reflected in the research boundary for our research,but online education consumer behavior research is a more extensive and complex problem.Although highly representative,the research cannot represent all possibilities.Secondly,this paper adopts a questionnaire.Although it can genuinely reflect the subjective behavior of consumers,there may still be deviations due to various objective conditions.The effect of objective data support will be more obvious.

ACKNOWLEDGEMENT

This work was supported by National Social Science Fund Youth Project“Research on the Group Behavior of ‘Post-95’ College Students Based on Complex Networks”of China(Project Number:17CKS047)