CORPORATE SOCIAL RESPONSIBILITY AND MARKETING: A LITERATURE REVIEW OF EBSCO DATABASE 2008-2010

Penerapan prinsip GCG menghendaki terakomodasinya kepentingan stakeholder dalam pengelolaan bisnis. Di sisi lain, penerapan CSR secara benar berarti juga memenuhi prinsip responsibilitas yang merupakan prinsip dasar dari GCG. Tujuan paper ini adalah melanjutkan ulas balik (literature review) artikel yang telah dilakukan sebelumnya oleh Egri and Raltson (2008). Perbedaan yang ada, dibandingkan literature review sebelumnya, adalah bahwa paper ini secara lebih detail menggambarkan metode apa yang digunakan dalam artikel dan trend dimana penelitian dilakukan serta teori yang digunakan. Data yang diperoleh dari database EBSCO dianalisis menggunakan analisis isi. Hasil analisis menunjukkan bahwa dari 83 artikel publikasi, didominasi oleh penelitian empiris. Artikel non penelitian atau teoritikal berisi tentang kerangka kerja penelitian CSR dan marketing. Sedangkan untuk artikel penelitian empiris, unit analisisnya didominasi oleh masyarakat umum, pengumpulan data didominasi dengan survey, teknik sample yang digunakan didominasi oleh sample non probability. Analisis data dilakukan kebanyakan dengan menggunakan SEM, lokasi penelitian kebanyakan dilakukan di USA, dan terdapat perubahan focus isu dari teori agency ke teori stakeholder dan teori etika.


INTRODUCTION
On 12 October 2005 the Government proposed the Draft of Limited Liabilities to the House of Representatives.The draft was intended to replace the Law of 1/1995 on the Limited Liabilities.Wahyudi and Azheri (2008) wrote that when the Government proposed the draft, the Government wants to support the implementation of Good Corporate Governance (GCG).Although it took difficult process, eventually on 20 July 2007 the ten fractions totally agreed with the enactment of the draft to become the Law of Limited Liabilities.The tough process was due to the existence of article 74 that rules about the obligation of companies running the business in natural resource exploitation (companies resource-based) to have social and environmental responsibility.
Good Corporate Governance, according to Organisation Economic Co-operation and Development (OECD) includes four main sectors (Fuadi in Wahyudi and Azheri, 2008).Those are: (1) shareholders' right and the protection; (2) role of employees and others stakeholder; (3) accurate, transparent, and on-time disclosure of corporate structure and operation; and (4) responsibilities of commissioners and board of directors to stakeholders.The main four sectors are then put into 4 basic principles.Those are fairness, transparency, accountability, and responsibility.Responsibility means corporate as the member of society to conform the prevailing regulation and law and to meet social needs.Responsibility emphasizes on specified system to arrange the mechanism of corporate responsibility to shareholders and stakeholders.This is relevant with the objective to create GCG that accommodates stakeholders such as the society, the government, business associations and other parties (Arifin, 2005).Wibisono (2007) wrote that corporate responsibility takes the form of corporate conformity to the prevailing regulations (for example tax, industrial relation, work health and safety, environmental protection, and conducive to business environment).Wanadjaja (2008) wrote that the implementation of GCG principle requires the accommodation of stakeholders' interest in business management.Appropriate implementation of CSR also means to meet the criteria of responsibility principle, which is the basis of GCG.In other words, CSR is closely related with GCG and can be used as the vehicle in implementation of GCG (Gill, 2008).
The concept of CSR appears in management and marketing literature (Maignan et al., 2005).Sen and Bhattacharya (2001) stated that there are a number of CSR concepts, extending between neo-classical view about corporate responsibility to shareholders and corporate view of corporate obligation to carry out the responsibility to stakeholders.From marketing perspective, CSR is seen as the concept of social process to meet the expectation, norms, and values of stakeholders (Sweeney, 1972).Maignan and Ferrell (2004); Maignan et al., (2005) stated that expectation, norms, and values reflect the issue that may directly affect stakeholders such as consumer rights, environmental protection, product safety, information transparency; and indirectly affect the stakeholders such as the use of child labors in corporate operation, ignorance of labor rights, and labor discrimination.Dawkins and Lewis (2003) stated that at present time consumers give strong pressure to the implementation of CSR as an important factor in purchasing decision.
The objective of this study is to have literature review of how researchs about CSR and marketing are performed in scientific publication.Baumgartner and Pieter (2003) wrote that literature review can contribute in the access of the impacts of some journals on knowledge.Margolis and Walsh (2003) informed that literature review can give direction to future researchs.Meanwhile Egri and Ralson (2008) stated that literature review can serve as the basis of future researchs in similar fields.
Literature review about CSR can be seen in articles written by Bakker et al.,, (2005).They collected published articles of CSR from 1969 to 2002.The results of analysis showed that the evolution of CSR in the last 30 years is a progressive mode and takes in various forms.The weaknesses of this article are that it only used the titles and abstracts of sample articles for the analysis.In other research by Lockeet et al., (2006), they collected some published articles about CSR in 10 journals of management, from 1992 to 2002.The results of analysis showed that most of researchs about CSR in the sample were quantitative rather than qualitative.In addition, there are 4 dominant themes in the publication about CSR.They are social responsibility, environmental responsibility, business ethics, and stakeholder approach.Other review performed by Egri and Raltson (2008), sought and collected articles about CSR published articles in 13 international journals from 1998 to 2007.From the 321 articles reviewed, they found that 75% of publication were empirical researchs, more than half (51%) of the empirical articles used survey in collecting the data.This paper refers to the previous researchs.The difference from those researchs is that this paper does not select the journals to be reviewed.The articles collected are those from 2008 to the beginning of 2010 and are obtained electronically from only the database of EBSCO.Another difference is that, according to me it is the advantage of this paper, is that this paper attempts to be in more detail look at the method that the researchs used in the articles published.Therefore, the questions proposed for the articles are: (1) what is the method used in the published empirical articles?; and (2) what is the trend of data analysis, location of the research, and theory used in the empirical articles?

AGENDA SETTING THEORY AND HYPOTHESES
The theory of agenda setting was initially proposed by Walter Lippman.An empirical research of this theory was performed by McComb and Shaw (Rakhmat 2007).McComb and Shaw (1972) studied the determination of the agenda in the presidential campaign in 1968 and hypothesized that mass media determined the agenda for each political campaign and affected attitude projection to political attitude.Severin and Tankard Jr (2007) wrote that the theory of agenda setting refers to the capability of the mass media, with the repeated news release, to promote the importance of an issue into public mind.
The assumption of the theory of agenda setting is that mass media filter the news, articles or texts to be released.Then selectively, the" gatekeepers" such as the proofreaders, editorial staffs, and journalists determine which news will be released and which other is concealed.Because the readers, viewers, and listeners receive most information from mass media, the agenda of mass media is closely related with public agenda.For example, when the Constitution Court played the recorded talk between the man thought to be Anggodo and the men thought to be government officials, both electronic and printed mass media, collectively and repeatedly put it on news that Anggodo became a very popular person thought to be a "legal case intermediator".If mass media is able to shape image about the person thought to be a legal case intermediator by publishing the information, then it is understandable to assume that mass media plays an important role in informing about something.
In this study, agenda setting theory was applied in the publication by one or more researchers in an academic journal.Therefore it is assumed that an academic journal filter some articles or passages to be published.Then selectively, a peer review determines which ones are good for publication, and which others are not.Because researchers have a lot of information and materials of research from journals, then the agenda setting of the journals is certainly related with the agenda of the researcher.
From the aforementioned description, the hypotheses of this study are: : There is a different frequency between articles of empirical researchs and articles of theoretical in the publication H 2 : For articles of empirical researchs, there is a different frequency of the use of qualitative, quantitative, and mixed approach of researchs in the publication H 3 : For articles of empirical researchs, there is a different frequency of the types of unit of analysis used in the publication H 4 : For articles of empirical researchs, there is a different frequency of the methods of data collection used in the publication H 5 : For articles of empirical researchs, there is a different frequency between the researchs using probability sampling method and non probability sampling method H 6 : For articles of empirical researchs, there is a different frequency between the researchs which perform validity testing and those which do not.H 7 : For articles of empirical researchs, there is a different frequency between the researchs which perform reliability testing and those who do not H 8 : There is a changing trend of the method of data analysis, location of research, and the theory used in the publication Different from hypotheses one to seven, hypothesis eight is not binding hypothesis, it is only a guideline to perform the analysis.

METHOD
There are 2 types of content analysis: conceptual analysis and relational analysis.Siregar et al., (2009) wrote that while relational analysis deals with examining relationships among concepts in texts, conceptual analysis further by quantifying the presence frequency of concepts represented by words of phrase.Titscher et al., (2000) wrote that the main core and tool in content analysis is the categorization system: each unit of analysis has to be coded or, in other words, allocated in one of more categories.Although Weber (1990) wrote that there is not any single best method to perform content analysis, this study carefully attempted to use the aforementioned procedure suggested by many researchers.Graphically, the procedure performed is presented in the following figure (see Figure 1).
Determining sample method: Identifying the text of artucle to be examined based on unit of analysis Determining coding sheet: validity assessment Reliability assessment using the coding sheet and judges coded: coefficient of reliability from Holsti Analysing and interpreting the data Specifying unit of analysis

Population and Sampling
Athough it is not meant to generalize, the population of this study is all articles found in the database of EBSCO, while the scope of the study is all articles meeting the criteria: (1) they are obtained by using the key phrase of "CSR and marketing"; (2) the articles are downloadable or in other words they have the extension of pdf (print document file); and (3) the articles are published from 2008 until beginning 2010.To select the sample, saturation sampling method is used.Saturation sampling is defined as the acquisition of all element of samples in a population having the characteristics requried by the researchers (Black and Champion, 1976).

Operational Definition of Variable
Study variables are the variables stated in statements of study.The operational definition of this variable is as follow: a) EBSCO is the the database of paid on line scientific article, accessible on subscription.b) Corporate Social Responsibility (CSR) is a sustainable commitment of business sector to ethically behave and contribute for the economic development, while improving the life quality of the employees and the family members as well as the local community and the larger society c) Marketing is the activity of deciding how to advertise a product and firm, what price to charge for it etc.d) Articles are the manuscript written by an author or more on a particular subject.This article may result from empirical study , or review of previous articles.e) Approach of study is the paradigm used in the research including: (1) positivism, using quantitative and experimental methods to test the generalization of hypotetikodeductive; and (2) phenomenological, using qualitative and naturalistic approach, which holistically understands human experiences in special context (Idrus, 2007).f) Unit of analysis is the single member of the sample, which is the element of the population member having single or multiple treatments.g) Data collection is the method in which data of research is obtained.Adam et al., (2007) wrote that there are some methods to collect data, including observation, experiment, survey, interview, diary methods, case studies, data storage, and triangulation.h) Validity represents the extent of which a measurement measures what it wants to measure (Ancok, 1999).Meanwhile reliability is defined as the ability of a measure to measure consistently (Black and Champion, 1976).i) Trend of research is a general tendency in the way a research is changing.j) Location is a place where data research is collected.k) Theory is the general principles or ideas of a scientific subject.

Data Collection
Data collection is performed by using the computer at the Faculty of Economics and Business, Master and Doctoral Program of UGM.The data needed in the study is secondary data, that is the articles published in the database of EBSCO.The use of database of EBSCO is based on the results of research performed by Blessinger and Olle (2004).They compared 3 online database of EBSCO, ASAP and ProQuest collected in November 2002.The comparison, that includes scope of subject; title of article; accessibility; and article trend, showed that EBSCO is the "deepest" database and presents the largest number of articles, has the longest duration, and has the best peer review.In addition, EBSCO is also the most aggressive online database in adding new titles of article in the recent.Berelson (1952) wrote that in content analysis, validity is not a big deal.In careful operational definition and correct and good indicator selection, coding sheet is assumed to measure what it should measure.In other words, it is enough to measure the validity by using face validity.Meanwhile reliability testing in this study is performed by inter coder reliability.This testing is performed to ensure the objectivity of data that will be analyzed and also to ensure the reliability of the used analysis technique.In this study reliability is tested by Holsti's coeficient of realibility (1963).

Data Analysis
The number of catagory assignments on which all coder agree C 1 , C 2 = The sum of all catagory assignments by all coder After validity and reliability testing, the data is analyzed by statistical program of SPSS, the results of which are descriptively analysed.

RESULTS
The preparation of the coding sheet is consulted to researcher at Faculty of Mathematic and Natural Science, Gadjah Mada University.The objective of the consultation was to have face validity test in the coding sheet (see Appendix 1 and 2).After that, with the assistance from a reasercher from Institute of Social Study, reliability is tested (see Appendix 3).There is a difference in determining the threshold score of acceptance of reliability coefficient.Krippendorf (1980) reports that the lower limit of acceptance of reliability coefficient is 0.80; Scott in Hasrullah (2001) put it over 0.75; while Berelson (1952) put the coefficient score between 0.79 and 0.96. the calculation of reliability at this study is 0.94 or above the score proposed by Krippendorf and Scott, and between the range suggested by Berelson.

General Description of Article Sample
In the beginning of 2008, in World Economic Forum in Davos, Switzerland, attended by the Secretary General of UN Ban Ki-Moon, 30 state heads, approximately 100 cabinet ministers, and hundreds of chiefs of large corporate all over the world, Bill Gates, delivered his speech as well as introduced the term of creative capitalism.Djatmiko (2008) wrote that creative capitalism is a new form of capitalism that gives more emphasis on poor groups that have been marginalized by the tremendous wave of capitalism.Although what he said is actually an old paradigm, but because it was stated by a person who has been in position of the richest person in the world, it sounds very beautiful.Soon after that, the theme became hot issue around the world.The concept of creative capitalism or previously known as CSR is essentially a concept that maximizes the positive impact of corporate operation and minimizes the negative impact towards sustainable development.Vogel (2005) wrote that since 1990an if the term Corporate Social Responsibility was put into the search engine of Google, there will be more than 30000 sites displayed, more than 15 millions of pages in World Wide Web referring the term CSR, more than 1000 corporate having signed the code of conduct about their commitment to give more attention to social, environmental, and human right issues in their operation, and more than 2000 corporate issuing reports about the CSR that have been implemented.The results of the collected articles based on names of journals can be seen in Table 1 in the following.

Figure 2 Name of Journal and Total Sample
Table 1 shows that the journal that most frequently published articles about CSR and marketing between 2008 and the beginning of 2010 was Journal of Business Ethics (39.29%).When viewed from the year of publication, in 2009 there were more articles published than in 2008 (a detailed on Appendix 4).

Hypothesis testing
Hypothesis testing is performed by using chi-square analysis.Siegel and Castellan (1988); Cooper & Schindler (2006) wrote that chi-square analysis is used in the testing that uses nominal data.This technique tests the difference between the distribution of sample data and the expected distribution.At the confidence level of α= 5%, all alternative hypotheses were suppported by the collected data (see Appendix 5 for detailed).

Testing of hypothesis 1 H 1
: There is a different frequency between articles of empirical researchs and articles of theoretical in the publication (χ 2 = 6.857 ; d.f.= 1 ; n = 84; Asymp.Sig = 0.009) Of 84 articles, there were 54 empirical research articles and 31 non-research articles (theoretical).In theoretical articles, the writer described the framework or model of researchs about CSR and marketing.

Testing of hypothesis 3
H 3 : For articles of empirical researchs, there is a different frequency of the types of unit of analysis used in the publication (χ 2 = 20.415 ; d.f.= 2 ; n = 54; Asymp.Sig = 0.000) Unit of analysis is the single member of a sample, which is the element of population members.In general, there are 3 (three) unit of analysis used in the researchs.They are the public (commond people), students, and others.What it means by others is annual reports, magazine advertisements, and corporate websites.The largest number of unit of analysis mostly found in the research conducted by Crittenden (2009) was 6226 students in 36 countries and the least number was 3 companies as found in the research conducted by Maon et al., (2009).

Testing of hypothesis 4
H 4 : For articles of empirical researchs, there is a different frequency of the methods of data collection used in the publication (χ 2 = 20.259 ; d.f.= 3 ; n = 54; Asymp.Sig = 0.000)   The data used in the research articles was primary data.Adam et al., (2007) wrote that there are several methods of data collection.Those are observation, experiment, survey, interview, diary methods, case studies, data storage, and triangulation.On the other hand, Churchill (2001) wrote that primary data can be collected by observation and communication.Meanwhile Jogiyanto (2004) wrote that data can be collected by observation, interview, experiment, computer-based survey, and archived data.
Using the opinion of Jogiyanto ( 2004), there are 22 articles using survey, 14 articles using observation, 8 articles using experiment, 3 articles using interview, and 7 articles using mixed method in the data collection.Of the articles using survey for the data collection, 4 articles used mail survey, 5 articles used telephone survey, and 7 articles used questionnaire survey.

Testing of hypothesis 5
H 5 : For articles of empirical researchs, there is a different frequency between the researchs using probability sampling method and non probability sampling method (χ2 = 29.630; d.f.= 1 ; n = 54; Asymp.Sig = 0.000) Sampling technique can be classified into two broad catagories.They are probability and non probability.Of 53 articles of empirical researchs 47 articles employed non probability sampling technique and only 6 using probability sampling technique.It is interesting to see the use of theoretical sample in the articles written by Valor & Marino (2009) and Maon et al., (2009).Theoretical sample is the sample collected based on the concepts theoretically related with the constructed theory (Strauss and Corbin, 1998).
They also mentioned that the objective of this sampling technique is that it collected the sample of events, incidences, and others indicating the catagory, nature, and size, in order that we can arrange and connect them conceptually.Poerwandari in Salim (2006) wrote that theoretical sample is determined on the basis of articular criteria underlined by the previous researchs (constructed from initial assumption).With this techniqeu, sample was collected in the researchs with grounded theory.Glasser and Straus in Strauss and Corbin (2009) wrote that this type of theory will, " … bridge the disparity between theoretical researchs and empirical researchs."

Testing of hypothesis 6 and 7
H 6 : For articles of empirical researchs, there is a different frequency between the researchs which perform validity testing and those which do not (χ 2 = 7.407 ; d.f.= 1 ; n = 54; Asymp.Sig = 0.006) H 7 : For articles of empirical researchs, there is a different frequency between the researchs which perform reliability testing and those who do not (χ2 = 6.000 ; d.f.= 1 ; n = 54; Asymp.Sig = 0.014) In general, validity is related with two matters.They are validity of measurement and validity of the research.Validity of measurement is the extent to which a measurement measures what it wants to measure, while validity of research is the control to secondary variable.On the other hand, reliability gives focus on the 'consistency of measurement in measuring a phenomenon.Of 54 articles of empirical researchs, 36 articles had reliability testing and 37 articles had validity testing, the rest did not mention any testing of measurement.
For the quantitative empirical researchs, reliability testing was performed with 3 methods.They are inter-rater reliability; test-retest; and internal consistency reliability using Cronbach's alpha/ α.Meanwhile the validity testing was performed by Exploratory Factor Analysis, Confirmatory Factor Analysis, face validity, and construct validity (discriminant and convergen validity).Meanwhile for qualitative method, reliability was tested by credibility standard, transferrability standard, and dependability standard (i.e.Maon et al., (2009); Muzellec and Lambkin (2008); Ahlstrom and Egels-Zanden (2008).
Credibility standard and dependability standard were tested by involving peers (not involved in the research) to discuss, give input, and criticism.The involvement is made at the making of questionnaires, and reporting the results of research.

Trend of Research Location and the Theory Used
In seeing this trend, we count the mode of data analysis, location of the research, and the theory used.
Table 1 continued Table 1 shows that 10 articles employed Structural Equation Modeling (SEM), 9 articles employed regression and descriptive analysis, 6 articles employed case study, 8 articles employed content analysis, and 7 articles employed Analysis of Variant (ANOVA).
In the articles using SEM for the data analysis, 5 articles were analyzed by AMOS, 4 articles were analyzed by Partial Least Square (PLS), and 1 article was analyzed by LISRELL.Partial Least Square is a technique that generalizes and combines principal component analysis and multiple regressions.In PLS, optimal linear relation among latent variables was calculated and interpreted as the best predictive relation despite the limitations (Ghozali, 2008).Additional advantages of the use of PLS are: (1) that it does not require large samples.Nijssen and Douglas (2008) wrote that PLS is effective at formative and reflective latent construct, is freely distributed (not necessarily normal), is very strong to analyze small-size samples.Pinto et al., (2008) wrote that PLS requires small samples; Mahmood et al., (2004) wrote that technique of analysis in PLS required minimum scale, sample size, and residual distribution; Lee (2001) wrote that PLS can accommodate only small samples; Birkinshaw et al., (1995) suggested that PLS used when only small-size of samples could be collected, assumption of multivariate normality and interval scale cannot be proved, and when the focus of study is to predict dependent variables; Graham et al., (1994) wrote that PLS is more robust with small samples; (2) it is not affected by multi-colinearity.Inkpen and Birkenshaw (1994) wrote that in their study, all relations were modeled simultaneously to avoid multi-colinearity, so that they used PLS; (3) it does not require normal distribution of the data and normal distribution of the error term.Pirouz (2006) wrote that some advantages of PLS, are: (a) that it can be used in the model with many dependent and independent variables; (b) that it is a robust testing for the missing data and noise data; (c) that it can be used for reflective and formative latent data; (d) that it can handle nominal, ordinal, and continuous scales.In brief, the difference (vis-à-vis) between variant-based and covariance-based is presented in Table 2.

Number of Analysis Unit
Large.Small.

Complexity of model
Little to medium complexity (less than 100 indicators).

Relation between latent variable and the indicators
Only able to handle reflective indicator.
Able to handle both reflective and formative indicators.
Table 2 shows that the locations of the research were mostly performed in the United States of America (42.6%).It is not surprising since CSR as part of GCG has developed since the disclosure of Enron case in America.On the other hand, Indonesia as a location of research is found in only 1 article, written by Wanderley et al.,, (2008).They studied 127 websites of large corporates in developing countries.Therefore, although the name of Indonesia was mentioned it is probable that they accessed it from other countries.
Initially, the most frequently used theory in GCG research was predominantly by agency theory.The results of analysis did not show any article using this theory.Most of the theory used to explain empirical researchs were stakeholder theory (9.6%) and ethic theory (7.4%).Freeman (1984) explained stakeholder as any "individuals and groups having interest in the achievement of organizational objective that in turn can result in affecting the achievement of their objectives."Therefore, stockholders is a part of stakeholder other than other parties.Etes (1995) wrote that, the investors in a business is not only stocholder, but also stakeholder.The investors are workers, customers, and community tax payers that support the existence of the company.This type of investors are called stakeholder, and they have loan in accounting sense because they invest in large amount through the submission of valuable resources in the form of money, work, career, and sometimes life to the company.
There are a number of variants in ethic theory.Two of them are utilitarianism and deontology.As reflected in the name, the ethic theory of utilitarianism roots in the teaching of "utility".The main proponent of this variant is Jeremy Bentham and John Stuart Mill (Hunt, 1991).Keraf (2002) wrote that the notion of Benthal and then also Mills can be formulated in 3 (three) criteria.The first criterion is utility, that is any policy or action that generates particular benefits.The second criterion is the largest benefits, that is any policy or action that generate the largest benefits of all alternative policies or actions.Or, in the situation where all existing alternatives seem to generate disadvantages, good action is any action that generates the least disadvantages.The third criterion is the largest benefit for the largest amount of people.It means that a policy or action is considered good when the it generates the largest benefits for many people.Alternatively it can be concluded that the principle held by utilitarianisme is "the greatest good for the greatest number" implying the message to act in such a way to generate the largest possible results for the largest number of people.
Deontology theory consists the word "deon" deriving from the Greek word of "obliged".Pieris and Jim (2007) wrote that this ethic school of thought measures the goodness of action not from the result of the action, as it has been argued in utilitarianism.The concept of obligation is not at all related with the concept of goodness.Even a good action does not depend on the good result.Humans are not ethically responsible for the results of the action, because it has been their obligation to conform moral obligation, free of any other considerations.

Conclusion
Because GCG and CSR are closely related, the objective of this study is to describe the articles with the topic of CSR and marketing published in EBSCO database.Totally, there were 84 articles as the sample as specified by the minimum number of study sample.The results showed that the published articles consisted of theoretical and empirical researchs.The articles of theoretical contained the framework of CSR and marketing analysis