Frame Contention between News Sources and News Media : Framing the Dispute of Teaching Mathematics and Science in English

The studies of framing have been so far focusing on the analysis of media effects or the frame-setting process. Comparatively, the process of building media frames that takes place in the newsrooms has received meagre attention from the scholars. And traditionally, most of the studies that investigate frame-building process have their attention centred on the internal factors, including newsroom practices and characteristics of the journalist. External factors, especially the interaction between media and news sources, are thus far being overlooked. On that account, this study seeks to examine the influences of the above two variables on frame-building process through examination of a contemporary issue in Malaysia – teaching Mathematics and Science in English. The scrutiny of the role of news sources in the process of frame-building may contribute significantly to the knowledge of framing studies. Based on the 1,156 units of analysis collected from five Malaysian Chinese dailies, the study’s findings show that news sources or frame sponsors are comparatively more influential in the frame-building process.


Introduction
Like it or not, we are now living in the era of information overload.Each day, mass audiences are bombarded with thousands of information in various media formats and forms.With a range of mass media now joining the ranks of competition to capture audience's attention, the power of media in influencing audience's perceptions and actions has drawn greater interest from not only the business and political elites, but media and communication scholars as well.As noted by McCombs & Shaw (1972) in the Chapel Hill study (during the 1968 United States presidential election), news media are capable of telling the audience "what to think about" by highlighting on the salient issues throughout a period of time.This discovery has given birth to the Agenda Setting Theory and set the milestone for the mass communication studies.This theory however does not support its precedent that advocates for powerful media effects, instead, it is dancing to the tune of limited media effect paradigm that emphasises on media pervasiveness rather than persuasiveness (Baran & Davis, 2003;Dearing & Rogers, 1996;Littlejohn, 2002;Miller, 2002).In Scheufele & Tewksbury's (2007) terms, agenda setting is an accessibility effect, that is, pervasiveness of information can make certain issues more accessible for people to recall and thereby influence the manners they form perceptions about a particular issue.Since its inception in the 1970s, the McCombs & Shaw's theory has been constantly an attraction to many communication scholars.
Further studies of agenda-setting theory has led to the emergence of priming and framing, with the latter outshined the prior and its precedent.Weaver (2007), in his meta-analysis to detect the trend of media effects studies, finds that the growth of framing research has escalated more than two folds in just ten years.Not only framing is gaining ground in communication science, it is believed to be the most popular and widely applied approach in the field of communication studies (Bryant & Miron, 2004).
The rise of framing and the advent of priming have marked the shift of limited effects paradigm to the powerful effects in political communication research (Scheufele & Tewksbury, 2007).While agenda setting theory places the stress on the transfer of news salience, which largely aims at telling the audience what they should be thinking about, priming looks at the influences of media agenda on audience's evaluation process (Brewer et al., 2003;Domke, 2001;Holbert & Hansen, 2006), and framing takes a step further to guide the audience on how to think about an issue (Baran & Davis, 2003;Miller, 2002).This approach explores the potential influences of media by examining its effects on audience's schemata or the human frames.Framing, in particular, pays special attention to the organisation and presentation of news attributes as mean to impact attitudinal changes among its audience.
Although a lion's share of framing studies is on media effects, it is important to pay a closer look at how these message frames are created and built into the media frames -the frame-building process.As specified in the three important questions about framing studies raised by Scheufele & Tewksbury (2007), besides examining the effects of framing, ones should also explore the process of how a frame is built.Also noted by Fahmy et al. (2010) in their study of the U.S president's media framing, there are rooms in the scientific research to investigate the influences of important frame sponsors.To further explicate the process of frame-building, de Vreese (2005) develops a model to distinguish the process of frame-building from frame-setting.
In the process of frame-building, the internal factors that influence the outcome of a frame take the centre stage in framing analysis.Tuchman (1978) in his study of daily newsroom routines has noticed that media workers are bounded by their workplace cultures in reporting an issue.Zhou & Moy (2007) also noted that, traditionally, scholars are more interested in figuring out the influence of internal factors, including the individual characteristics, ideological and political orientations, professional values, journalistic routines, and organizational constraints, on journalists' presentation of a frame.However, influences of external factors must not be excluded from the scene; they do play a role in the frame-building process.As noted by Zhou & Mou (2007), news framing are influenced externally by the political system, political culture, and prevailing notions among the public (Zhou & Mou, 2007).Carragee & Roefs (2004) also point out that media frames do not develop in vacuum, but are shaped by actors or claim-makers through series of competition.These actors may include politicians, organizations, and social movements.Hence, recognizing the influences of these actors in building frames of certain issue is of the essence in framing analysis (Olausson, 2009).This paper classifies the actors into two general categories -journalists and news sources, and attempts to examine the interaction between the two in determining their influences on frame-building, particularly on the education issue in Malaysia.

Literature Review
While agenda setting attempts to explore the ways media highlighting salience issues that aim to notify audiences about what make the important news of the day, framing, as Matthes (2009) puts it, "examines the selection and salience of certain aspects of an issue by exploring images, stereotypes, metaphors, actors, and messages" (p.349).The distinction between agenda setting and framing is thus clear, at least according to this definition, the prior is looking on a bigger picture which take on the overall presentation of a mass media, whereas the latter has a comparatively narrower focus that pinpoints on the presentation of the issue itself.
To depict a clearer idea of what framing is about, Gamson (2004) puts forward an analogy liking the intangible media frame to a tangible picture frame.As what a picture frame does, media frame too places borders around something (issues highlighted in media in our case), distinguishing it from what is around it (Gamson, 2004).Also through this media frame, journalists are able to highlight on a particular aspect of the issue, have audience's attention focuses on it and ignore the others.It is through the combination of various framing devices as noted by Matthes (2009) in his above-mentioned quote, frame sponsors actively pack their messages with the aims to influence the audience's perception of the issue.Reese (2001) considers framing in this capacity as the "symbolic forms of expression" by ways of using words or visuals to entail patterns or categorisations of pictures in the audiences' heads (p.12).Reese (2001) also refers framing as "organising principles that are socially shared and persistent over time, that work symbolically to meaningfully structure the social world" (p.11).Simply put, framing targets the cognitive aspect of its audiences in order to bring about an impact in their perceptions (see Carragee & Roefs, 2004;Endres, 2004;Gevorgyan, 2010;Scheff, 2005;Goftman, 1974;Simon & Xenos, 2000).
According to Goffman (1974), human frame is defined as the cognitive structure of an audience that guides both his/her perception and representation of reality.As such, "how issues are constructed, discourse structured, and meanings developed?"are the few questions that constantly draw interest from the framing researchers (Gamson 1989(Gamson , 1992)).The structured messages framed by frame sponsors, as Lorio & Huxman (in Reese 2001) point out, carry impact on human cognitive processing.Both the scholars observe that audiences derive meaning of a textual description in the process of interaction between theirs schemata and the text.In other words, "to frame is to actively construct the meaning of the reality in question" (Hanggli & Kriesi, 2010) and "make the world beyond direct experience look natural" (Gitlin 1980, p. 6).According to Kinder (2007), "…in modern society, ordinary citizens must rely on others for their news of national and world affairs" (p.155).This phenomenon is in line with the famous quote of "pseudo-environment" carrying the meaning of "the world out there and the picture in our head" as expounded by Lippmann (1997, p.3) in the understanding of the formation of public opinion since the 1920s.
With that in mind, framing as part of the communication process in general, it is not static.Framing, by itself, can take place at various levels and by a range of groups of people (Miller & Riechert, 2001;Tankard, 2001).As noted by de Vreese (2005), framing can basically be divided into two distinctive but inter-connected sectionsframe-building that explains on how a frame emerges in the media, and frame-setting which look into the interplay between media frames and the audience schemata in generating an impact on meaning construction.In the process of frame-building, scholars pay close attention to the factors that determine the qualities of a frame.These factors, as de Vreese (2005) identifies, can be basically divided into two general categories -internal factors that refer to the practices in the newsroom and the personalities of the journalists, and external factors that have close interaction with the editorial personnel in such a way influencing the outcome of a frame.
To further explore the dynamic feature of framing, Entman et al (2008) assert that framing process takes place at four different levels: in the culture; in the minds of elites and professional political communicators; in the texts of communications; and in the minds of individual citizens.Out of these four levels, culture sets the boundaries and limits of framing in terms of stock of schemas commonly found in the minds of a society's individuals, and stock of frames present in the system's communications.These may include literature, entertainment, news, conversations and other political discourse (Entman et.al, 2008, p.176). Van Gorp (2007) has argued that "a shared repertoire of frames in culture provides the linkage between news production and news consumption" (p.61).
Within the cultural boundaries, frame sponsors are competing among themselves to exert influence on the media frames.Frame sponsors are described by Gamson (1992) as news sources who make efforts to frame information in news stories.They will make sure that the messages are framed in a way that is consistent with their preferred framing (Hallahan, 1999).Those messages are carefully designed during the competition (strategic framing) in such a way to not only capture editors and journalists' attentions but also to induce target audiences' acceptance of the interpretations that favour their interests or goals.Meanwhile, Pan & Kosicki (1993) regard the contention of frames as a strategic action in which participants manoeuvre strategically to achieve their political and communicative objectives.They see the contest as successfully judged by the media workers who choose to accept one set of terms over the other.These frame sponsors may include politicians, pundits, editorial writers, and bloggers.Journalists and editors, however, are not perceived as part of the contenders, as Entman et al. (2008) justified, though they play a major role in the frame-building process, they rarely push any political goal.These two key players of the newsroom are somehow seemed passive to many framing scholars (Hanggli & Kriesi 2010, Sheafer & Gabay 2009, Stromback et al. 2008).
One explanation for the "passive nature" of the journalists and editors is that they keen to turn to official and established news sources.These open up a world of opportunity for these sources to dominate the news frames (Stromback et al. 2008).In this account, journalists are merely deemed to be the transmitters of the official viewpoints (Campell 2004).It is true that media is not completely dependent on a particular source, but still, it is highly restricted with their range of choices among whom the legitimate sources (Stromback et al. 2008).This assertion, however, invites challenges from some communication researchers (Benett 1990, Gans 1979, Shoemaker & Reese, 1996).They see the role of media in the frame-building process can be more active.Media, for at least, have the liberty to select its news sources, choose from whom to quote and organise the information, rather than just repeating it.In the process of framing a news story, journalists "select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described" (Entman 1993, p. 52).Although this famous quote from Entman does not specifically designate to journalist, but it has been widely used to exhibit the importance of journalists' organisational skills in representing a perceived reality.These organisational skills involve a process of selection, emphasis, interpretation, and exclusion.As Reese (2007) puts it, the appearance of a news item in media is through "active forces of order that bracket out certain happening via routinised, legitimised and institutionalised structure that favour certain ways of seeing"(p.149).This indeed denotes frame as representing both the psychological and sociological constructs as explained by Iyengar & Simon (1993).In this study, the level of influence for both journalists and sources of information will be put to test through the examination of the education issue in Malaysia.

Background of the Issue
Being a multi-racial society, education issues are constantly drawing attention of various ethnic groups in Malaysia, especially when it touches on mother tongue education.Teaching and learning of Mathematics and Science in English (PPSMI: pengajian dan pembelajaran sains dan matematik dalam Bahasa Ingeris in Malay), is one of such issues that has getting on the nerves of many Malaysians from different ethnicity, including the Chinese community.After seven years of implementation, the language policy which initiated in 2002 was eventually announced to be nullified in stages starting 2012.It is to be replaced with the new policy of upholding the Malay language and strengthening English (MBMMBI: memartabatkan Bahasa Malaysia, memperkukuhkan Bahasa Inggeris in Malay).The issue of PPSMI emerged from a resolution proposed by the United Malay National Organisation (UMNO) Kubang Pasu branch on the importance of English language in May 2002 (Jiao Zong, 2003).UMNO is the political party that dominates the ruling coalition of Barisan Nasional (BN).It took less than two weeks for the government to announce a rather disputed education policy that required all primary schools in Malaysia to teach Mathematics and Science subjects purely in English starting 2003.Announcement of the policy immediately sparked uproar among the Chinese community.To most Malaysian Chinese, primary education is the baseline to safeguard their mother tongue education, any attempt to make changes on the medium of instruction is perceived unacceptable.Seeing the strong objection from the Chinese community and the appeal of the BN Chinese-based component parties, Malaysian government had agreed for Chinese national-type primary schools to use Mandarin along with English in teaching Mathematics and Science.The government's trade-off, however, did not calm down the unrest of the Chinese community.There was no solution in sight and this education issue continued to be debated and it got inflamed in mid-2007.The Malaysian Chinese education movement, Dong Jiao Zong had taken the lead on 8 July, 2007 to launch a campaign asking government to revert the teaching of Science and Mathematics to mother-tongue.Dong Jiao Zong, the combination of United Chinese School Committees' Association of Malaysia (UCSCAM or Dong Zong) and United Chinese School Teachers' Association of Malaysia (UCSTAM or Jiao Zong), is widely recognised as the guardian of Chinese vernacular education in Malaysia (Tan, Ho, & Tan, 2005;Tan, 2005).Subsequent to the above campaign and the pressure from various non-governmental organisations, Malaysian Chinese Association (MCA) as the leading Chinese-based ruling political party was forced to react and it had to respond to those voices that were reflected in the news coverage.Consequently, a motion was passed by its youth wing to urge the government to continue using Chinese language as the medium of instruction in the 2008 Primary School Achievement Test for Mathematics and Science subjects.Nonetheless, their pleas were not accepted by the government.On 27 October, 2007, the then Prime Minister Tun Abdullah Ahmad Badawi made a surprise announcement on retaining the exam format.He declared that all primary schools could use two languages, i.e.English and their respective mother-tongue in answering these two subjects.
The hotly debated issue had captured the attention of various news media, especially the Chinese ethnic media.A content analysis conducted on the news coverage revealed that different actors in their respective organisations attempted to portray different slants and frames on this controversial issue.Subsequent to the announcement made by the then Prime Minister, many players attempted to frame their respective side of story through news coverage in mass media.Hence, the authors would base on this important announcement to make a closer examination on news sources in frame-building process besides the normal scrutiny of media framing in news coverage.

Methodology
Due to the multi-ethnic nature of Malaysian society, its mass media are designed to serve a particular ethnicity.Most of these media are reporting news in the community's mother tongue and as such have their target audience highly fragmentized.In this respect, only Chinese daily newspapers would be purposely selected for content analysis.Out of the six Chinese dailies being circulated in Peninsular Malaysia, i.e.China Press, Guang Ming Daily, Kwong Hwa Yit Poh, Nanyang Siang Pau, Oriental Daily News, and Sin Chew Daily, five were selected.Kwong Hwa Yit Poh was excluded due to its difficult availability in the location of study.
In terms of item collection period, this study took on a total of 21days starting from 28 October to 17 November, 2007, starting from the day after the then Prime Minister's announcement on the examination format.It is worth noting that previous studies on framing analysis do not indicate standardized criteria for sampling in terms of days of newspapers.For instance, Semetko & Valkenburg (2000) in their study of the "Eurotop" meeting covered 46 days before and five days on and after the event.Chyi & McCombs (2004) took one month to study the framing of tragic school shooting.Meanwhile, de Vreese & Boomgaarden (2006) took three weeks for the study of news framing on 2002 EU enlargement for two countries.
With the target newspapers and item collection period fixed, three coders, all proficient in Mandarin and familiar with the issue under studied, were engaged to carry out the content analysis.A background check was done on the coders to make sure none of them affiliated to any of the news sources.All three coders also had been trained on the procedure of content analysis, including identification of related news items and unit of analysis.Many scholars (e.g.Berelson, 1952;Weber, 1990) have listed some common choices of units, i.e. word, word sense, sentence, theme, paragraph, whole text, character, item, space, and time.This study, however, chooses to take paragraphs contained in the relevant news items as unit of analysis.
To begin with the coding process, coders were first asked to identify the relevant news items with the assistance of a list of possible wordings and expression used by the news sources.They were then provided a code book for coding instructions and reference, and a code sheet to record their observation.This study groups news sources into six categories as listed below:  Government: Policymakers and executors, which include politicians holding the governmental positions and civil servants (e.g.ministers, directors general, etc.). Barisan Nasional: Politicians from ruling political parties (e.g.UMNO, MCA, etc.) who are not holding governmental positions nor speaking on behalf of government. Movement: Education movement (e.g.Dong Jiao Zong, etc.) and non-governmental organisations (NGOs). Opposition: Opposing political parties.
 Editorial: Editorial write-ups and spot news.
 Reader: Letters to editors and comments or views from readers.
The coding scheme for news sources in this empirical study is determined by the players involved in the issue.As outlined above, there are six categories under this construct.Generally, office-bearers in the government and the politicians in the ruling political parties are those with authority to decide on the execution of a public policy.Meanwhile, activist groups or NGOs, especially Dong Jiao Zong, who are deemed as an interested party would give their views and comments on the policy executed by the authority.There are also some other news sources, such as the opposing political parties, readers, and media workers.Entman (2008) notes, scholars are far from a unified agreement on the definition of frame and framing.Hence, many media studies have examined news frames from various standpoints.Some researchers interpret the frames from the issue-specific perspectives (e.g.Entman, 1991;Norris, 1995), while others measure the frames in the generic manners (e.g.Hallahan, 1999;Iyengar, 1987Iyengar, , 1991;;Semetko & Valkenburg, 2000).This research study applies the generic-frame measurements developed by Semetko & Valkenburg (2000).The two scholars proposed five generic frames which are called Responsibility, Conflict, Morality, Economic Consequences, and Human Interest frames.
Each of these news frames comprises several attribute statements with a simple "yes: no" answer selection scale.The attribute statements and its measurement scale are adopted from the researches of Semetko & Valkenburg (2000), and Valkenburg, Semetko, & de Vreese (1999).Each "yes" in corresponding to a statement will carry a score of 1 mark while 0 mark for "no".The total marks collected for a particular news frame will be averaged as according to the number of attribute statements it contained.The values ranging from "0" indicating a particular frame is not in existence to a perfect "1" indicating the frame is fully present.
Human Interest frame has five attribute statements, in which each affirmative answer to the attribute statement contributes 0.2 score to the frame.If all five statements are ticked "yes", it indicates the Human Interest frame is fully present.The four attribute statements in Responsibility frame carry a weight of 0.25 score each, while three attribute statements in Conflict, Morality, and Economic Consequences frame give 0.33 score each.
In the process of categorising and measuring a news item's contents, there is always a risk of subjectivity in interpretation.The defect could result in low quality research outcome.To prevail over the issue of reliability, inter-coding between the three coders were done as an attempt to identify and correct the semantic problems.The coders' decisions in determining the existence of news frame were checked against each other using Holsti's (1969) agreement index.The results yielded an agreement level of .710 to .833among the coders, which has passed the obligatory .70mark.

Research Findings
A total of 220 news articles with 1,156 units of analysis had been identified from the three-week publication during the study period.As shown in Table 1, the peak with the most number of news articles and units of analysis being published on the second day after Badawi's announcement.
<Insert Table 1 Here> Oriental Daily News contributed the most units with 333 counts (or 28.8 percent of total), followed by Sin Chew Daily with 330 units (28.5 percent), China Press with 218 units (18.9 percent), Nanyang Siang Pau with 186 units (16.1 percent), and Guang Ming Daily with 89 units (7.7 percent).As for news sources, the sequence of ranking began with Government (362 units or 31.3 percent), followed by Movement (296 units or 25.6 percent), Barisan Nasional (290 units or 25.1 percent), Editorial (100 units or 8.7 percent), Reader (96 units or 8.3 percent), and lastly Opposition (12 units or 1.0 percent).
<Insert Table 2 Here> As shown in Table 2, a 6X5 Factorial Multivariate Analysis of Variance (MANOVA) was applied in testing the effects of news sources and media on frame building.The General Linear Model is executed on SPSS and the Box's Test of equality of covariance matrices is performed.The test result is significant and hence rejects the null hypothesis of equality of covariance matrices, meaning it has violated the assumption of similarity of covariance matrices for MANOVA test.In other words, the various frames differ in their covariance matrices.However, the F test is quite robust even when there are departures from this assumption.
MANOVA also assumes that each dependent variable has similar variances for all groups.The Levene's test statistic for Human Interest frame was not significant and hence meets the assumption.However, the test statistics for other frames reject the null hypotheses that the groups have equal variances or homogeneity of variances.Failure to meet the assumption is not fatal as this procedure of analysis uses the Enter method.
Multivariate tests are conducted in testing both main and interaction effects.The result of Pillai's Trace for "News Sources" [F (25, 5645) = 6.541, p < .05]shows that there is an overall main effect of news sources on the dependent variables -the level of visibility of the five generic frames in news reporting.The result of Pillai's Trace for "Media" [F (20, 4512) = 3.555, p < .05]also shows that there is an overall main effect of media on the above-mentioned dependent variables.Likewise, the result of Pillai's Trace for "News Sources*Media" [F (85, 5645) = 1.913, p < .05]shows that there is an overall interaction effect on the dependent variables too.
In measuring the portrayal of these frames in the study, Table 2 demonstrates the overall observed mean scores of the visibility of frames according to news sources, media, and their interaction.These values for Responsibility, Conflict, Morality, Economic Consequences, and Human Interest frames are to be elaborated with the inferential statistics as shown in Table 3.
<Insert Table 3 Here> Overall, Responsibility gains the most attention and became the most outstanding frame with M = .355(s.d.= .280).Both Human Interest and Conflict frames are subsequently second and third most prominent frames with M = .212(s.d.= .203)and M = .208(s.d.= .257)respectively.Meanwhile, Morality and Economic Consequences frames capture the least attention with the recorded mean value of M = .048(s.d.= .120)and M = .015(s.d.= .087).
For the F tests of significant differences in the visibility of frames for news sources, the test results show that four out of the five frames encounter significant differences.The Responsibility [F (5, 1129) = 13.857,p = .000],Human Interest [F (5, 1129) = 8.068, p = .000],Conflict [F (5, 1129) = 10.145,p = .000],and Economic Consequences [F (5, 1129) = 6.451, p = .000]have been ascertained for their significant differences in the visibility of frame.As for media, only three frames are found to be significantly different and they are Responsibility [F (4, 1129) = 3.170, p = .013],Conflict [F (4, 1129) = 5.285, p = .000],and Economic Consequences [F (4, 1129) = 6.290, p = .000].Meanwhile, the interaction between news sources and media produces significant differences for all the five frames.From the Post Hoc test and of close examination on the Responsibility frame, Government (M = .434)and Barisan Nasional (M = .388)are found to highlight the accountability aspect in solving the problem.Relatively, Movement (M = .326)tends to be more moderate as compared to the former news sources.The other news sources show lesser degree of visibility in this aspect.In terms of media, Nanyang Siang Pau (M = .406)and China Press (M = .396)show more emphasis on Responsibility frame while Guang Ming Daily (M = .362)and Sin Chew Daily (M = .361)are slightly lower.Orient Daily News (M = .293)is far behind the others.From the interaction of both, the highest visibility is found on Government in Nanyang Siang Pau (M = .500)and the lowest is Opposition in China Press (M = .083).
The F tests have ascertained the main effect of news source and interaction effect on Human Interest frame, but no main effect of media.The Editorial (M = .308)and Reader (M = .273)are seen as trying to touch on the emotion of the audience and their visibility of this frame is highest among all.Barisan Nasional (M = .210)and Movement (M = .208)are positioned at the middle level.The other two sources were relatively lower.The subgroup of Reader in Oriental Daily News (M = .450)is found to have projected the most Human Interest frame while Opposition quoted in Sin Chew Daily (M = .000)seems to have neglected the sentimental aspect of the news story.
In terms of Conflict frame, Opposition (M = .389)and Reader (M = .319)are found to have sustained their different opinions as distinct from the Government and Barisan Nasional.The two former sources are positioned at the highest level while Government (M = .147)and BN (M = .168)are on the other extreme located at the lowest level for the visibility of this frame.Meanwhile, Editorial (M = .280)and Movement (M = .255)are moderately placed at the middle level.In the media aspect, Oriental Daily News (M = .274)and China Press (M = .214)have demonstrated a significant greater visibility of the Conflict frame from the other media.As there was no source of Opposition attributed in Oriental Daily News and hence this news source cited in China Press (M = .556)portrayed to be the most vocal player, vis-avis the least quarrelsome Government in Guang Ming Daily (M = .043).
In spite of being the most argumentative player in this issue, Opposition (M = .139),as compared to the other news sources, paid the highest attention to Economic Consequences frame.
The mean scores of visibility for all other news sources were below .1 level.Meanwhile, Guang Ming Daily (M = .034)is found to have attained the highest visibility but the score is rather low and fails to establish significant differences from China Press (M = .021),Sin Chew Daily (M = .018),and Oriental Daily News (M = .010).Opposition in Sin Chew Daily (M = .267)and in China Press (M = .111)are the subgroups that promoted the most for the Economic Consequences frame while many others do not even show any visibility of this frame, i.e.M = .000.
Although there are no main effects of news sources and media on the Morality frame, an interaction effect of both is established.Looking for the mean scores for above .1 level of visibility, the portrayal of three subgroups, i.e.Movement in Guang Ming Daily (M = .167),Opposition in China Press (M = .111),and Editorial in China Press (M = .111)tend to be more visible than others.

Discussion and conclusion
From the above findings, this research has enabled us to understand 'the occurrence of communication within communication' -how news sources and media workers play their roles in framing news stories.Looking at the values of Partial Eta Squared or in Table 3 for the effect sizes, news sources are noticed to have greater impact on Responsibility, Conflict, and Human Interest frames.Meanwhile, its interaction with media produced more effect sizes on Economic Consequences and Morality frames.It is worth noting that media have the least size of effect for all frames, including those with significant differences.
The findings of this study suggest that news sources are more influential in the frame-building process in comparison to media workers.The prevalent frames promoted by the respective news sources quoted in the daily newspapers have also been recognised.As illustrated in Figure 1 <Insert Figure 1 Here> From the above illustration, all players in the issue have obviously established predominant frames in which to promote certain angle of the story.Their strategic actions in gaining public support through media is in line with the function of framing as defined by Pan & Kosicki (1993) in achieving political and communicative objectives.The news sources are to either align or contest with each other in trying to win this opinion game.These frame sponsors as described by Gamson (1992) have clearly and certainly slanted the stories towards their preferences.In other words, they framed the news stories in a way that is consistent with their preferred framing as expounded by Hallahan (1999) in his public relations research.
Although news sources are comparatively more influential in the frame-building process, it doesn't mean journalists are merely the 'message transmitter' for new sources.Based on the research findings of Tuchman (1978) and Shoemaker & Reese (1996), journalists do not simply repeat whatever information put forward by the news sources but rather select and organise this newsworthy information before getting it on print.The significant interaction effects on all frames have proven this journalistic practice in framing the news stories.On that account, neither news sources nor media workers can claim sole ownership to the news frames portrayed on media.It is clear that the portrayed news frames are the results of interaction between news sources and media workers.
Overall, it is hoped that this research has resolved the problem as stated in the earlier section.The data collected from the field has met the research objective in understanding the frame-building process performed by news sources in promoting their preferred frames.This research has also reaffirmed the findings of framing by media as encountered by most communication scholars (e.g.Semetko & Valkenburg, 2000;Valkenburg, Semetko, & de Vreese, 1999).Nevertheless, the authors wish to contribute further details on the frame-building process that the strength in terms of influences in the frame contention between news sources and media, the prior is greater.
, Government and Barisan Nasional have prominently promoted the Responsibility frame.Meanwhile, both Reader and Editorial are found to have significantly promoted the Human Interest frame.As for the Conflict frame, Reader joins hand with Opposition in objecting this rather disputable issue.Except for Government and Barisan Nasional, Movement and Editorial also portray this Conflict frame moderately.Movement's stand is not clear in this study as it has promoted Responsibility and Human Interest frames moderately too.Besides prominently highlighting Responsibility frame, Barisan Nasional has portrayed a moderate Human Interest frame.Other than the above, Editorial has promoted Responsibility frame moderately and Opposition has spread its focus to promote Economic Consequences frame subtly as well.

Table 1 .
News Articles and Units of Analysis Note: There was no publication on November 8, 2007 due to Deepavali (Festival of Light).

Table 2 .
Mean Scores for Visibility of Frames Note: Values in parentheses represent standard deviations.

Table 3 .
Tests of Mean Differences