LINKED FACTORS AS PREDICTORS OF THE PRACTICE OF PUBLIC RELATIONS ON THE INTERNET : A STUDy OF CROATIAN ORGANIZATIONS

A change from traditional interactions between organizations and the public to interactions via the Internet needs further exploration in order to understand factors that might influence public relations activities of organizations on the Internet. This study investigates responses of 111 Croatian PR professionals to a survey about the influence of organization-linked factors on the choice of public relations models on the Internet. The results confirm that the social orientation level of top management members, availability of resources to operate, a previous negative experience with the public and practitioners’ knowledge of research methods and techniques are influential predictors of the choice of the public relations models on the Internet. The findings also validate the contention that the organizations’ selection of the public relations models depends on practitioners’ formal education in PR and the experience in dealing with conflict, PR unit size, its location and role in the formulation of PR strategies. It is also shown that organizational culture, a management style, a support for and understanding of the PR function by top management, an existence UDK: 659.4:004.738.5>=111 Izvorni znanstveni rad/ Originale scientific article


INTRODUCTION
The online environment is much more complex, diverse and dynamic than the traditional media sphere.The Internet has led to decentralization of information and power: a communicator is no longer in a privileged position, as is the case in the traditional mass communication sphere, neither is the public a passive consumer of pre-selected and arranged contents according to the criterion of media professionals anymore.The public has the opportunity to comment information, express opinions and reactions, discuss with their authors, refer them to new sources, similar and different information and thus participate in their creation (Zgrabljić Rotar, 2011, 49).
In the field of public relations, the use of the Internet makes it possible for public relations professionals to communicate with various stakeholders without the gate-keeping function of other mass media.One of the most distinctive features of the Internet in public relations is that it gives an organization ability to build an interactive relationship with its stakeholders.Now, public relations practitioners have the opportunity to gather information, monitor public opinion, and engage in a direct dialogue with the public (McAllister & Taylor, 2007).While many authors argue that the Internet creates a new communication landscape that facilitates two-way communication between an organization and the public (Holtz, 2002;McAllister & Taylor, 2007;Philips & young, 2009;Grunig, 2009), there is still a lack of empirical data on the extent and ways in which public relations practitioners use the Internet (MacNamara, 2010, 4).The majority of studies, so far, have suggested that most public relations practitioners, regardless of the benefits and opportunities of the Internet, have not been using it in full measure (Kirat, 2007;McAllister & Taylor, 2007;Kelleher, 2008).The Internet, by itself, does not lead to new forms of communication, and public relations units do not exist separately from organizations.In order to improve the efficiency of PR in organizations, it is necessary to understand the factors influencing their performance.Relevant literature is limited regarding the impacts upon the performance of PR in cyberspace.Hence, the purpose of this study is to determine what factors influence and restrict the choice and use of public relations models in Croatian cyberspace.The research question is presented as follows:

RQ: What organization-linked factors influence the performance of PR in Croatian cyberspace?
Although often criticized, the theoretical framework is based on the work of Grunig and Hunt "Managing Public Relations" (1984) describing four models of public relations.There are several reasons for their re-actualization.The Internet, with the development of web technologies, is the only medium that allows the use of all four models of public relations (Ingenhoff & Koelling, 2009, 67).Its dialogic, interactive, and global properties make it perfectly suited for a two-way symmetric paradigm of public relations and to overcome the contextual conditions that limit the application of the models in practice (Grunig, 2009, 6).Studies specifically using Grunig and Hunt's four models to analyze online public relations are seemingly non-existent in Croatia.This research contributes to understanding and possibly prediction of public relations models in cyberspace by documenting the influential factors through a literature review and through surveying public relations practitioners in Croatia.

THEORETICAL BACKGROUND
Over the last few years, public relations has evolved into an independent discipline of theory and practice.A long-term success of organizations and the growth of competitiveness nowadays cannot be envisaged without public relations.More and more organizations have recognized the necessity of PR and are making use of it on a professional basis.Despite of that, there is still not one single generally accepted definition of public relations.Public relations has been defined by different authors in various ways.However, common to all of these definitions is an emphasis on communication.
Model building in public relations has been one of the most pervasive methods for comparing public relations practices.Many studies in the field of public relations have used models based on the original work of J. E. Grunig and T. Hunt (1984).Considering the historical evolution of public relations, they define four models of public relations: press agentry, public information, a two-way asymmetrical model and a two-way symmetrical model.Although these models were originally conceived as a means of explaining the historical evolution of PR, the authors maintain that these models also provide a means of explaining differences in the way in which PR is practised in organisations.For the purposes of this paper, it is appropriate to define these models clearly, since subsequent discussion depends on these definitions.
The first model, press agentry is propaganda where telling the truth is not important.The primary goal is to obtain media attention in almost any possible way (Grunig & Hunt, 1984).Organizations seek favorable publicity from media sources, and attempt either to prevent negative publicity or to respond to negative publicity with the objective of neutralizing it.Communication is always one-way.Organizations neither seek out nor utilize communications and information from the public and have no active intention of learning public's reaction.It is the least effective and least ethical model (Gordon & Kelly, 1998).The second model, public information is also based on a one-way communication stream from an organization to its publics.The aim is to provide the public with live information for a main purpose of informing and a lesser goal of persuading.This model differs from the press agentry model only in that public relations performs as "journalists-in-residence" who create and provide relatively accurate information about the organization through mass media and controlled media such as newsletters, brochures and a direct mail (Grunig, 1992, 18).In practice, however, this model does not volunteer negative information to the public.Communication programs do not use research and strategic planning (Grunig, 1992), but follow a journalistic model of preparing informational material for largely unknown publics.
The other two models employ two-way communicationin other words, communication flows both to and from publics and public relations practitioners are both talking and listening.A two-way asymmetrical model of public relations has scientific persuasion as its goal.In this model, communication is two-way but unbalanced.Organizations put out information and receive feedback from the public but do not necessarily respond to that feedback in the way the public has requested.Under this model, organizations conduct research and gather information from and about publics for identifying messages that are likely to persuade the public to act as the organizations wish (Grunig, 1992, 18).This approach is not about collaborating on goals, plans, or actions, but rather about eliciting and controlling preferred stakeholder attitudes and behaviors.A two-way symmetrical model, in contrast, has its goal in mutual understanding and a beneficial relationship between an organization and its public (Grunig, 2000).It is based on a research and uses communication to manage conflict and improve understanding with strategic publics (Grunig, 1992, 18).The model advocates a free and equal information flow between the organization and its publics -more attention is paid to the use of dialogue and negotiation.This approach reflects organizations' desire to create information and feedback loops and to form interactive, strategic partnerships with key stakeholder groups.
Publics are not merely heard; their interests and views are incorporated in strategic decision-making processes.This may result in either the organization or its public being persuaded to change its position.However, the model is regarded as equally effective if neither group changes, as long as both communicate well enough to understand the position of the other (Grunig & Hunt, 1984, 23).Although all four models, depending on the problem and the situation, can be effective, two-way symmetrical communication, based on the assumption of listening, negotiation and compromise, is considered the best practice.It produces better long-term relationships with the public than other models of public relations (Grunig, Grunig & Dozier, 2002, 15).
The practice of PR is influenced by needs and idiosyncrasies of an organization.PR practitioners not only play different roles in different organizational environments (Hogg and Doolan, 1999), but also practice public relations differently depending on a number of organization-linked factors.Public relations literature includes many variables that make good candidates as factors influencing public relations practitioners and their practice.However, there is no list of universally acceptable factors.
Top management is in the focus of many theoretical discussions and a public relations research.The perceived importance of the public relations function by top managers will determine their position in the organizational hierarchy and the type of program that will dominate in practice.Understanding of what public relations can do, will dictate their purpose.Grunig J. (1992, 493) claims that the top management support for and understanding of the public relations function coincides with the use of sophisticated two-way models of public relations.Pollack R. A. (1986, 87) is of the same opinion when she claims that the top management support creates a suitable atmosphere for the practice of two-way symmetrical public relations.According to Rhee (2002, 165), the level of the top management support for public relations impacts the way in which they will be organized, the extent to which they will be involved in the strategic decisionmaking, and the choice of public relations models.
Researches on the influence of top management philosophy suggest that top executives with a strong social orientation would place greater importance on communication and be more concerned about the views held by their primary publics (Miles, 1987).Given that the social orientation level of top management members influences organization's strategies towards its publics, it is conceivable that it influences the way an organization practices public relations.Top managers with a strong social orientation try to achieve legitimacy through collaboration with multiple stakeholders (Meznar and Nigh, 1995).This philosophy is based on trust and open communication with various publics.A management style (democratic vs. autocratic) of top executives can also greatly influence the public relations practice of organizations.Hage (1980, 143) indicates that organizations with liberal ideology (democratic style) are very sensitive to the external environment.McMillan (1984) found a link between conservative managers and the use of the two-way asymmetric model while R. Pollack (1986) found a positive correlation between valuing a strong central authority (autocratic style) and the use of public information and press agentry models.These results suggest that organizations with top management made up of individuals who encourage liberal values (democratic management style) are more likely to be open to the external publics.A democratic management style involves two-way communication that motivates both, groups and individuals (Sikavica and Bahtijarević -Siber, 2004, 352).Dant and Schul (1992) suggest that an organization's history can influence its interactions with publics.Organizations that have had a negative experience with the public in the past often create a perception of the situation based on these experiences.The negative experience often leads to fear, lack of trust and risk avoidance.Engaging in a dialogue and negotiations is something one approaches with fear and hesitation.As the result of the negative experience, organizations will either try to avoid any form of two-way communication, or endeavor to establish control over the discussion using an aggressive communication style.The existence of an organization's legal department could be another potential influencing variable.When organizations are under threat or handling crises there is often a tendency to turn to a legal counsel.Legal concerns normally revolve around liability.
Fearing that what is said might be misunderstood, the instinct is to close down communication with no comment (Tomić, 2008, 203).Although public relations practitioners are aware that no comment means that an organization is hiding something and that expressions and demonstrable actions of sensitivity, concern and responsibility enhance reputation (Gregory, 2004, 64), public relations professionals might avoid dialogue and openness with the public due to the lack of getting legal advice.Sriramesh et al. (1996), Vasquez and Taylor (1999) and Grunig et al. (2002) wrote that organizational culture influences all decisions made in an organization, including how the organization practices public relations.Reducing typologies of organizational culture into a continuum between authoritarian and participative cultures, Sriramesh (1996) found that authoritarian cultures usually use a closed -system approach to management while participative cultures tend to use an open-systems approach.In authoritative cultures, communication is mostly one-way.By contrast, organizations with participatory culture demonstrate greater use of two-way communication (Grunig et al, 2002, 496).The existence of effective risk and issue management programs or staff in an organization may help to make the top management members and public relations staff well aware of the status and importance of organization's publics.An organization that recognizes the value of utilizing issue management is also likely to recognize the value of positive relationships with publics and thus will be more willing than other organizations to strategically adapt in some way to changes in its external environment instead of reacting to changes (Benoit and Brinson, 1994;Heath and Nelson, 1986) Several studies have suggested that public relations activities vary depending on PR practitioners' formal education in public relations.J. Grunig and L. Grunig (1989) and Zoch et al. (1997) indicated that the practice of the two-way symmetrical model is significantly related to practitioners' educational background in PR.A positive correlation was also found between the choice of public relations models and a number of practitioners who have a college degree in a unit (Grunig et al., 1989, 50).The involvement of the PR unit in major decisions made by an organization and its role in the formulation of PR strategies are also factors that have an impact on the use of two-way symmetrical communication and define excellence of the public relations unit (Simcic Brønn, 2001;Steyn, 2007, 167;yang, 2007;Grunig, 2009;Seitel, 2011, 51).Participation of public relations in the strategic decision -making and strategy formulation depends on the unit's ability to do research.Results of studies show that there is a strong link between the practitioners' knowledge in applying research methods and techniques and the strategic position of the public relations unit (Dozier, 1990, 78).
A practitioners' experience level in dealing with conflict is another factor that can influence the public relations practice.An experience in dealing with conflict can empower public relations managers to become an effective part of the communication process in the management decision-making group (Grunig and Grunig, 1992;Plowman, 1998).The power of the PR unit is represented by the value and support that the unit receives from the top management.Much of this value and support for public relations managers can come from the use of skills attained from experience to resolve conflicts or problems with the organization's environment (Plowman, 1998).
The amount of available resources to operate (such as time, funds and staff) also has a significant impact on organization's communication behavior.L. Grunig's (1992, 524) research showed that many practitioners, due to lack of time, funds and staff members, often ignored public pressure.More resources, according to Davis (2009, 34), means "... more media contacts, greater output of information subsidies, multiple modes of communication and continuous media operations".Several studies have suggested that public relations activities vary depending on the size of the public relations unit.Schneider (1985) found a correlation between the size of the public relations unit with the use of a two-way asymmetrical model.yang and Chiou (2010, 225), in addition, have found that public relations experts in smaller units are more likely to act as a communications technician, while experts in larger units will be more inclined to act as a communications manager.
Based on a research documenting the negative influences on the public relations unit placed under a corporation's marketing umbrella, the location of the public relations unit in the corporate hierarchy may influence the unit's potential to influence an organization to practice open communications and to accommodate certain publics.Sublimation to another management function typically results only in attention to the stakeholder category of an interest in that function.Grunig and Grunig, (2006, 12) point out that sublimation to marketing usually resulted in asymmetrical communication.According to Kristensen (2010, 144) and Hon (2007) the PR unit's influence in institution-wide decisions in relation to other units might be another potential factor influencing organizations' PR activities.

RESEARCH HyPOTHESES
On the basis of the literature review, the present study has investigated the influence of the following organization-linked factors (independent variables): (1) organizational culture, ( 2

) management style, (3) social orientation level of top management members (4) support for and understanding of the PR function by top management (5) organization's previous negative experience with public, (6) legal department (counselor) support, (7) existence of risk and issue management programs or staff, (8) number of higher educated practitioners, (9) practitioners' research methods and knowledge techniques, (10) practitioners' experience level in dealing with conflict, (11) amount of available resources to operate, (12) involvement of PR in major decisions, (13) PR unit 's influence in relation to other units, (14) formal education in public relations, (15) PR unit's role in the formulation of PR strategies, (16) location of public relations unit and (17) PR unit size.
This research assumes that organization-linked factors (independent variables), derived from the literature review, will to a certain extent predict the choice of the models of public relations on the Internet (dependents variables).

Sample
Survey participants were invited to participate via e-mail.An invitation was sent to 419 e-mail addresses (selected from the Croatian PR association directory) on January 15, 2011.One hundred and eleven practitioners responded, yielding a 26.49% valid response rate.The survey respondents were essentially defined as top public relations managers (i.e., head of the public relations unit).Members of the public relations units designated by top public relations managers were also considered as acceptable respondents, because it is assumed that a top PR manager will designate individuals who can provide the appropriate answers.Given the structure of the questionnaire, public relations practitioners and members of PR agencies (counselling firms) were excluded.
The largest group of practitioners (52%) falls in the age group 30-39, followed by those in the age group 40-49 (28%).Of the remaining 20%, only 8% were 50 years old or older and 12% were in the 20-29 age group.Of 111 respondents (84 women and 27 men), 75% of them have a tertiary university qualification.The majority of respondents (82) come from organisations with more than 100 employees (34% with 100-500, 14% with 501-1000 and 26 % with more than 1000 employees).Almost half of the respondents (49%) were employed in private organizations, 36% in public and 15% in organizations without ownership (associations, civic organisations, etc).

Instruments
The study employed a web survey of public relations practitioners.The questionnaire was constructed on the first author's website (http://bernardmi.netne.net/questionnaire.html).A builtin control mechanism warned respondents about unanswered questions and prevented them to press the submit button.Before the survey was administered, a pilot test of the survey instrument was conducted across a sample of eight respondents from the same population to ensure that questions were easy to understand and answer, and clear in conveying meaning.

Structural measures of the models of PR in Croatian cyberspace
Public relations models were measured using 18 statements (items) borrowed and adapted from previous studies (Kim &Hon, 1998, 166 andLim, Goh andSriramesh, 2005, 331).The statements were translated by the first author of this paper, and then reverse-translated by a Croatian English-teacher and an English-speaking person to achieve maximum veracity to the English original.Respondents were asked to indicate their attitude and perceptions on a 5-point Likert type scale toward the current activities (elements of PR models) performed by their PR unit on the Internet, where 1 represented strongly disagree and 5 represented strongly agree.The statements addressed the press agentry (items 1-3), public information (items 4-7.), two-way asymmetrical (items 8-14), and two-way symmetrical (items 15-18) models.The public relations models were measured by asking the respondents to indicate on a five-point scale (from 1=completely disagree to 5=completely agree) the extent to which they agree with the statement that the current purpose of the Internet's use in their organization is to...: (1) get publicity for their organization, (2) determine how effective the PR programme has been measuring the amount of registered 'hits' on the organization's website; (3) prevent /neutralize unfavourable publicity; (4) disseminate accurate information but not to volunteer unfavourable information; (5) design an easy to use and functional website (i.e.navigational capability); ( 6) keep a news-clipping file /to publish news stories; (7) disseminate neutral of information (rather than those in favour of the organization); ( 8) change public opinion about the organization; (9) conduct research in order to design a website which could be used to persuade the public; (10) change the public's attitudes towards the organization; (11) conduct or study attitude surveys to make sure the organization and its policies are described in ways their publics would be most likely to accept; (12) conduct research to determine public attitudes before starting PR programmes; (13) persuade publics to behave in the way the organization wants them to behave; (14) determine how effective the PR programme has been in changing public's attitudes towards the organisation (after it has been completed); (15) create dialogue (interactive communication) between the organisation and its public; (16) solicit feedback from the public in order to change the behaviour of the organization; (17) negotiate conflicts with theeir publics (or vice versa); (18) do surveys to develop mutual understanding between their management and the public that the organisation affects.
Factor analysis using principal components analysis and varimax rotation was performed.The scale had good internal consistency (Cronbach's alpha = 0.86).The Kaiser-Meyer-Olkin test (k = 0.824) and Bartlett's test of sphericity (c 2 = 933.090,df = 153, p < 0.01) were both satisfactory.The analysis was conducted by respecting the rule of a minimum of 3 variables per factor and a factor loading of 0.50 or better for an item to be considered a part of a factor (be to retain).Five factors met the Kaiser-Guttman retention criterion of eigenvalues greater than 1.0.Cattell's scree test also suggested a five-factor solution.The total item variance explained by the five-factor solution was 69.60%.One of the five factors had only two variables that loaded on it suggesting that it should not be considered a stable factor.When a four factor model with varimax rotation was completed, only one variable on factor fourth met the criteria used for an item to be retained.Therefore, we decided, to carry out a factor analysis restricting the number of factors to three.Three items (5, 6 and 17) loaded on more than one factor and two (2 and 7) did not have a significant loading on any factor, therefore they were dropped.When completing the three-factor model without those variables once again, item 8 loaded highly on two factors.Due to its failure to load reliably on a single factor, the analysis was repeated after the variable was eliminated.Finally, a satisfying three-component factor solution was obtained with all items having factor loadings equal or greater than 0.50 (see table 1).

Table 1. Factor structure matrix after the varimax rotation -3 factor loading (cleaned)
The current purpose of the Internet's use in our organisation is to… Component 1 2 3 (12) conduct research to determine public attitudes before starting PR programmes 0.826 (11) conduct or study attitude surveys to make sure we describe our organization and its policies in ways our publics would be most likely to accept 0.744 (14) determine how effective the PR programme has been in changing public's attitudes towards the organisation (after it has been completed) 0.734 Out of four possible models from literature, only three factors (models) were identified.The first factor was labelled as the two-way asymmetrical model because it consisted of 5 variables (items 9, 11, 12, 13 14) defining the two way asymmetrical model.Factor 2 consists of variables (items 1, 3, 4 and 10) which are in general related to the public information model and the press agentry model.This factor was labelled as the news bureau model.The third factor is defined by three variables (items 15, 16 and 18) and was labelled as the two-way symmetrical model, because all these variables related to the two-way symmetrical model.Cronbach's coefficient alphas were calculated to examine the internal (9) conduct research in order to design a website to persuade the public 0.694 (13) persuade publics to behave in the way our organisation wants them to behave 0.689 (4) disseminate accurate information but not to volunteer unfavourable information 0.825 (1) get publicity for our organization 0.786 (10) change our public's attitudes towards the organisation 0.704 (3) prevent/neutralize unfavourable publicity 0.592 (15) create dialogue (interactive communication) between the organisation and its public 0.884 ( 16) solicit feedback from the public in order to change the behaviour of the organisation 0.864 (18) do surveys to develop mutual understanding between our management and the public that the organisation affects 0.615 Note: Rotation converged in 4 iterations consistency of the two-way asymmetrical subscale (factor 1 -α = 0.824), the news bureau subscale (factor2 -α = 0.764), and the twoway symmetrical subscale (factor 3 -α = 0.794).Acceptable coefficient alphas were obtained for all factors, supporting their internal consistency.

Structural measures of factors influencing PR practice
The index of collectivistic organizational culture includes the following items: a) group welfare is more important than individual reward, b) group success is more important than individual success, c) being accepted by members of the work group is very important, d) managers should encourage group loyalty even if individual goals suffer and e) individuals may be expected to give up their goals in order to benefit group success.
The items have been taken and adapted from Dorfman and Howell (1988).All the items use a standard 5-point agreement scale.Possible total scores range from 4 to 20 with higher scores indicating greater level of collectivism.The democratic management style index consists of 4 questions taken from Sriramesh et al. (1996): In my organization…: a) most decisions in my company are made after thorough discussion between all people who will be affected in a major way; b) senior managers believe in the sharing of power and responsibility with lower level employees; c) most employees in my company share a common sense of the company's mission that most think is worth striving to achieve and d) senior managers care deeply about other employees.A 5-item scale (from 1=completely disagree to 5=completely agree) was used for recording answers.The index is obtained by summing the values.Larger score indicates a higher degree of democratic management style.
The social orientation level of top management members was measured by a 5-item scale ranging from 1=completely disagree to 5=completely agree.The scale, constructed by J. Choi (2007), consists of 6 statements: Our top management thinks that ... a) society has the right to make certain demands from a private corporation; b) a corporation has a responsibility to respond to social change; c) a corporation has a duty to adapt it policies and practices in response to social expectation; d) corporations should bring social responsibility into their day to day operations; e) corporations should make social responsibility a part of business decisions and f) corporations must set social goals just as they set business goals.
Possible index values range from 6 to 30 with higher scores indicating greater level of social orientation.The level of support for and understanding of the PR function by top management is obtained by summing the value of the rating for each of the following statements (items): the top executives (managers) in my organization..

. a) have an understanding of the public relations function, b) support the public relations function, c) fully understands the role of public relations department and d) consider public relations function important to the success of our organization.
Respondents were given a 5 point scale with 1=completely disagree and 5=completely agree.The index is obtained by summing the values (possible total scores range from 4 to 20).Larger scores indicate a higher level of support and understanding of the PR function by top management.
The organization's previous negative experience with the public was measured by asking the respondents to indicate on a five-point scale whether they agree or disagree with the statement: "My organization didn't have negative experiences with the public in the past".As an indicator of the support of the legal department (counsellor), respondents were asked to indicate the frequency with which the organization's legal department provides legal support in complex situations, with 3 offered answers: never, sometime and always.The existence of risk and issue management programs or staff was dichotomized into either yes or no.As an indicator of the role of the PR unit in the formulation of public relations strategies, the following question was used "In your organization, PR strategy is formulated by..." with four offered answers: top-management and then implemented by the public relations staff, top management after consulting with the public relations staff, public relations staff but reviewed by management and solely the public relations staff.
The share of employees with formal education in the public relations unit was assessed with a question with 5 offered answers: none, a few, about the half, the majority, all.To determine the location of the PR unit in the organizational structure, a question with the following answers was used: stand-alone unit, within the marketing unit, within another unit and the unit is not formally established.The size of the PR unit was measured by the number of workers in the unit.As an indicator of the influence of the PR unit in institution-wide decisions in relation to other units a question was used with the following 5 answers: much less, somewhat less, about the same, somewhat more, much more influence.For the last six questions in this part of the questionnaire the respondents were asked to assess on a scale from 1 (strongly disagree) to 5 (strongly agree) the extent to which they agree with the statement that their unit had enough higher educated workers, enough experience in conflict resolution, sufficient trained staff in applying research methods, sufficient resources to operate and that their unit is involved in all major decisions made by their organization.

Data analysis procedure
The researchers used the IBM SPSS 19 to analyze data.Descriptive statistics were used to summarize the demographic data.Regression analysis with optimal scaling (also known by the acronym CATREG) was adopted to examine which factors predict the choice of the PR models of Croatian organizations on the Internet.CATREG extends the standard approach by simultaneously scaling nominal, ordinal, and numerical variables.The procedure quantifies categorical variables so that the quantifications reflect characteristics of the original categories.The procedure treats quantified categorical variables in the same way as numerical variables.The contribution of each predictor variable is expressed by a regression coefficient (β) which measures the average amount the dependent variable changes for a one-unit change in the independent variable (Kockla uner, 2000, 125;Meulman, 1997).

RESULTS
Before entering all independent variables for regression analysis, correlations among them were analyzed to check multicollinearity.The correlation coefficient greater than 0.8 between any two variables is an indication of multicollinearity (Field, 2005).Since the highest correlation coefficient is r=0.72 which is less than 0.8, there is no evidence of multicollinearity among the independent variables in this research (see table 2).

Table 2. Correlation Matrix (Pearson's correlation coefficients between independent variables)
* Significant at the 0.01 level ** Significant at the 0.05 level The regression model in table 3 shows the predictors of the choice of the two-way symmetrical model.The total amount of variance explained in the criterion variable was 64.4%.The overall relationship proved to be statistically significant (R = 0.803, F = 2.021, p = 0.005).In general, the results show that the choice of the two-way symmetrical model of public relations on the Internet can be predicted on the basis of certain organizational characteristics.
In the following analysis, we tried to identify the variables Finally, in our third analysis, we wanted to find out the most significant variables in explaining the choice of the two-way asymmetrical model of public relations on the Internet.According to data in table 5, the most significant predictors of the summative measure of the two-way asymmetric model are: formal education in public relations (b=0.384,p=0.002), location (organization) of public relations unit (b=0.312,p=0.023) and PR unit size (average number of employee) (b=0.282,p=0.008).The set of independent (predictor) variables explains 61.4% of variance in the criterion variable.The multiple correlation coefficient is 0.783 and is statistically significant at the 0.01 level (F = 1.908, p = 0.009).The results show that the choice of the two-way asymmetrical model of public relations on the Internet can be predicted solely on the basis of some tested variables.Education in public relations has the highest value of beta weight.The other two predictor variables have a lower value, particularly in the case for the number of employees in the unit.

DISCUSSION
The main purpose of this study was to verify the role of some factors in explaining the choice of public relations models in Croatian cyberspace.In order to determine which factors contribute to the choice of each of the three models separately, regression analysis with optimal scaling was conducted.The models were treated as criterion, and the organization-linked factors as predictor variables.Thus we got the answer which factors and to what extent they explain the choice of public relations models and how much of variance of choice can be explained by differences in organization-linked characteristics.
Regression analysis of the two-way symmetric model shows that all the variables explained about 64% of the variance, which is a sufficiently large percentage to claim that at least some of our variables are strong predictors of the choice of the two-way symmetrical model.The findings reveal that the social orientation level of top management members is the most potent contributor to the prediction of the two-way symmetric model.A good prediction also makes the organization's previous negative experience with the public.It is obvious that organizations that seek harmonious relations and relate responsibly to their environment are more likely to use the two-way symmetrical model.Furthermore, those who have experienced a negative experience with the public in the past are willing to engage in dialogue, which will allow them to put similar trouble aside.This suggests that negative experiences help increase awareness of the positive implications of the use of the two-way symmetric public relations.
PR unit's role in the formulation of PR strategies is also identified as a predictor of the two-way symmetrical model.This finding agrees with the works of Broom and Dozier (1985), Grunig and Schneider Grunig (1989) and White and Dozier (1992) that discovered that a top management team is a key to success in establishing two-way symmetrical communication programs and in defining just how broad-based PR is going to be.They argued that the top management team is instrumental in determining public relations strategies and therefore is instrumental in defining the focus, scope and structure of PR.Isolating public relations from the top management team limits the profession to developing 'reactive' PR strategies because PR units rarely rise above the role of communication technicians.Training in applying research methods and techniques is also found to be a significant predictor of the choice, which is not suprising.The two-way symmetrical model is a more complex form of public relations that seeks to balance the interests of the organization and public.In the model, the emphasis is on encouraging and creating understanding, which involves changing attitudes, awareness and behavior.Although these changes can sometimes be easily recognized, achieving these goals and quantifying the level of change usually require research.
Also consistent with Grunig and Grunig (1992) and Plowman (1998), the experience level in dealing with conflict is a statistically significant predictor of the two-way symmetrical model.Overall balance between the organization and the public rarely exists, especially when their interests are different.A conflict is an integral part of an organization's life.A resolution can be reached gradually, by learning how to resolve issues through negotiations in the spirit of mutual respect.The conflict cannot be resolved constructively without positive interactions and the avoidance of creating winners and losers.Therefore, it seems logical that organizations, whose PR units have an experience in conflict resolution, tend to use the Internet to create the atmosphere of an open dialogue with the public, which can lead to better relationships.
Resources to operate also make a significant contribution to the prediction of the two-way symmetric model choice, even slightly better than formal education in public relations.This positive relationship indicates that the establishment of open, two-way communication with the public is under the direct influence of the available amount of financial resources, time, and/or staff that the PR unit can mobilize to engage in such activities, as well as the formal education of the PR staff, which is not surprising because the two-way symmetrical concept, as an advanced stage of public relations, requires the use of strong analytical and communication skills.Without an adequate level of knowledge of public relations theory, public relations can hardly be understood as a two-way communication process.
Similar results were obtained in explaining the choice of the news bureau model.The total percentage of explained variance is somewhat smaller than in the two-way symmetrical model (about 62%).Here too, the social orientation level of top management members is a significant predictor (it also performs the best prediction in the regression model -beta 0.578).The previous negative experience with the public also makes a significant contribution to the prediction of the news bureau model use, but lower when compared to the two-way symmetrical model.
The relation between the news bureau model and the social orientation level of top management members as well as the previous negative experience with the public is somehow surprising given that both variables are also predictors of the two-way symmetrical model.A possible explanation for a greater contribution of the social orientation level of top management members and of the previous negative experience with the public in choosing the news bureau model could be that the model, among others, includes the consistent release of objective and accurate information.Namely, socially orientated top management members, although prone to cooperation and to establish harmony between the organization and society (the principle of mutual benefit), will sometimes hesitate between a reactive or proactive attitude because of the previously negative experience.Fearful of a re-failure and aware of the dangers that a dialogue (proactive stance) can bring, they will decide to publish only objective and accurate information (reactive stance), in order to respond to the request, pressure, etc.
The PR unit's role in the formulation of PR strategies, knowledge (training) in applying research methods and techniques and experience level in dealing with conflict are also found to be significant predictors of the choice of the news bureau model.It is evident that public relations units when dealing with different conflict situations, and based on their own prior experience, are also using the news bureau model besides the two-way symmetric model.That leads us to the conclusion that depending on the situation, it is desirable to combine different communication styles, but also that two-way communication does not always ensure successful conflict resolution which is in line with the results of previous studies that have shown that in some cases, two-way communication can infact worsen the outcome (Krauss and Morselli, 2000).Depending on the situation, the public relations unit will also disseminate information (with or without an emotional charge) and depending on the structured message, will rely either on the rational interpretation of information by the public or on the emotions of the public.
At first sight, based on a quick and superficial judgement, the relation between training in applying research methods and techniques and the news bureau model can appear unacceptable.However, if we start from the basic assumption that conducting evaluation researches with identical methodologies as in traditional media can cause larger errors in the measurements, then the connection between research methods and the news bureau model is more than clear.The news bureau model is a one-way model in which, a research is usually limited to descriptive summation (for example, to determine the number of people who visit the Web site) or to determine the degree of intelligibility of information (press releases on the internet).However, many practitioners in public relations began their professional career before the digital age and are not comfortable with the internet environment.As digital immigrants they are forced to adapt to the rapid changes in communications and information technologies.Monitoring an on-line media release requires a shift away from the traditional metrics.
If the goal of on-line public relations is to increase the number of visits to a website, using web analytics tools it is possible to monitor what is happening and where the traffic is coming from.For example, on a blog, using frequency of comments, it is possible to determine how much an organization is interesting to the public, which may be an effective indicator of attention attraction (publicity) or changing attitudes.If the number of commentators is large, it is obvious that something must be happening.Using Google Trends, for example, it is possible to monitor trends and evaluate the popularity of search terms, while blogpulse.commonitors what people are talking about (blog trends).Although they will not reveal whether the exchange is positive or negative, they will reveal the frequency of conversation.
The results of the analysis also indicate that the PR unit's role in the formulation of PR strategies has a significant impact on the choice of the news bureau model.However, the impact is mostly indirect.In organizations where PR units are not involved or do not cooperate with the top management in the formulation of PR strategies, public relations activities are of a lower order and are practiced as a reactive activity (reacting to events, usually defensively).The public relations unit's assignment is to make public what has already happened.It is a one-way activity.Public relations units are telling the world what the organization has done or is doing, without taking into account the views of others.They perform routine daily tasks: creating publications, press releases, news stories, reports, speeches, etc. Public relations are just a tactical communication tool used when adding ''gloss'' to information.
Finally, regression analysis of the two-way asymmetric model shows that the percentage of explained variance is lower than in the first two models (0.61).In this model, only three variables significantly predict its choice.The best predictor of the twoway asymmetrical model is formal education in public relations.The location of public relations unit also makes significant contribution to the prediction of the choice of the two-way symmetric model choice as well as the size of public relations unit.
This finding could be attributed to the fact that the purpose of the two-way asymmetrical model is scientific persuasion.In the model, communication flows in two directions, to and from the public, and PR practitioners speak and listen.Knowledge based on formal education in public relations is obviously an element that helps public relations professionals in persuading the public to accept the point of view of the organization and to behave in a way that supports the organization.This is in line with Grunig and Grunig's (1992) claim that public relations can be the professional and management function only if public relations practitioners have necessary theoretical knowledge, which includes education for that profession.The two way models, according to the authors, are the opposite ends of a continuum which they define as professional public relations.
As for the location of the public relations unit, the choice of the two-way asymmetric model is stronger in those organizations where public relations are not organized as an independent organizational unit.When structured under other organizational units (marketing, legal, services ...), public relations are often only a support tool of those organizational units, and they deal only with the public of interest for the unit in which they are located.In such a paradigm, according to Grunig (2009, 4), it is often believed that the public can be persuaded, that is, that it is possible to influence their knowledge, attitudes and behaviours by means of communication that is designed to promote the interests of the organization with little or no concern for the public interest.The size of the public relations unit is also a factor that influences the decision of the choice of the two-way asymmetric model.This is not surprising since the rise in the number of employees (practitioners) leads to specialization and professionalization of the staff, which are prerequisites for the advanced two-way practice of public relations.This result is similar to results found by Schneider (1985, 573) of a positive correlation between the size of public relations unit and the use of the two-way models (asymmetric and symmetric) in mixed mechanical -organic organizations.

LIMITATIONS
In order to fully consider the potential of this study, several limitations must be addressed.The size and convenience of the sample limits generalizability of these findings.Organizations of different sizes, types of ownership and sectors were not equally covered.There are likely some limitations in the selection of the independent variables (factors) too.The practice of public relations is complex and is influenced by many factors.In this study, only a limited number of independent variables have been tested.Other important variables, and their relationships to cyberspace public relations models, may need to be tested (i.e. the personal characteristics and abilities of practitioners themselves).
The additional limitation is the fact that public relations is still a young industry in Croatia.Not all the participants in this study have an appropriate concept of PR.In other words, some answers from certain items of the questionnaire could be limited because of the participants' limited perceptions of PR.Furthermore, the collected responses represent public relations practitioners' subjective assessments that do not necessarily reflect reality.The assessment of non-public relations members (top managers or other members of the organization), for certain variables could be more objective and closer to reality.
The lack of a previous research with which the results of this study could be compared is another limitation of this study.During the preparation of the questionnaire, numerous studies were consulted, mostly conducted in the U.S. or in other foreign countries that are culturally and developmentally different from us.For this reason, the interpretation of the research results is somewhat difficult due to inability to compare the results and to draw conclusions by comparing them.Using online surveys to collect information also entails certain limitations.It is not possible to check the understanding of instructions or monitor the behaviour of participants in completing the survey.It is difficult to verify the identity of participants, their age and gender.In addition, the anonymous nature of the Internet allows people to participate negligent or malevolent.It should be noted that the response rate to the questionnaire was low, which makes the interpretation of the data insufficiently precise, but still indicative.In the future, other survey methods, or qualitative interviews, should be considered as additional methods.

CONCLUSION
This research studies what organization-linked factors are associated with (predict) the choice of the public relations models in cyberspace by Croatian organizations.The results show that the tested set of variables is significant in predicting the choice of each model.The influence on the choice of the two-way symmetric model and the news bureau model exert: an experience in dealing with conflict, PR's role in the formulation of PR strategies, research methods and knowledge techniques, the social orientation level of top management members and a previous negative experience of the organization with the public.The amount of resources to operate and staff formal education in public relations also exerts a direct influence on the choice of the two-way symmetrical model.Only three variables emerged as predictors of the choice of the two-way asymmetric model: formal education in public relations, the size and location (organization) of the PR unit.
Although satisfactory models were obtained in which some clear relationships between the factors and the models were found, and even predictive power of certain factors, the results of this study revealed a weak correlation between the models and the tested factors.Moreover, the results show that most of the factors can be considered to be universal.Nevertheless, on the basis of the obtained research results, we can conclude with a measure of caution, that there are at least some variables that influence their choice.The relationship between the models of public relations and factors influencing their choice on the Internet in this study is only partially resolved.

Table 4 .
the choice of the news bureau model.The steps were identical to those in the previous analysis.Table4shows the predictors of the choice of the news bureau model.Together they explain 62% of the variance.The multiple correlation coeffi-Regression of independent variables on a news bureau model