Introduction……………………………………………………………………… 3
1. Overview…………………………………………………………………………… 5
1.1 Taiwan
1.2 Women Migration in Asia
1.3 ‘Foreign Bride’ Phenomenon in Taiwan
1.4 Don’t Call Me Foreign Bride Movement
2. Theoretical Framework……………………………………………………… 10
2.1 Social Representation
2.2 Mass Media
2.3 Representation of Ethnic Minority in Media
2.4 Image of New Immigrant Women in Taiwanese Media
3. Research Method………………………………………………………………
3.1 Content Analysis
3.2 Longitudinal Research
3.3 Data Collection
3.4 Categories and Coding
3.5 Reliability
4. Analysis and Discussion ……………………………………………….…..
4.1 Don’t Call Me Foreign Bride?
4.2 Data Description
4.3 Positive and Negative Valuations
4.4 Images of Immigrants
4.5 Who Talks About Immigrants?
Limitations and Implication for Future Research




The other is in no way another myself, participating with me in a common existence. The relationship with the Other is not an idyllic and harmonious relationship of communion or a sympathy through which we put ourselves in the Other’s place; we recognize the other as resembling us, but as exterior to us; the relationship with the Other is a relationship with a Mystery.

Emmanuel Levinas, Time and the Other

‘Foreign Bride’ is an unbearable label attached to new immigrants in Taiwan, marking the fact that they are perpetually the ‘Other’ in their beloved homeland. A generation of ‘Foreign Brides’ from Southeastern Asia have settled in Taiwan for more than two decades. They speak the Taiwanese language, eat Taiwanese food and raise their children in Taiwan, but due to their cultural and ethnical backgrounds, new immigrant women are continually fighting against being ‘othered’ in the society.This research seeks to examine the attitude of Taiwan print media in reporting immigrant women issues. The objective is to analyze whether or not after more than 20 years since Southeastern Asian women married Taiwanese men and settled in Taiwan, they are still discredited by the mass media as ‘uncivilized’ , ‘burden of society’ or ‘social problems’.

Mass media performs a crucial role in constructing social representations and it is the key to the process of the development of the public sphere in modern society. It distributes information citizens need to make choices of public affairs and provides an independent forum to facilitate the formation of public opinion (Curran 1991); as well as reflects and enhances pre-existing values. Studies on how newspapers and television news portray ethnical minority groups have increased significantly in recent years. Media and sociology scholars in Taiwan have published several outstanding researches on the representation of Southeastern Asian new immigrants in mass media. However, despite the amount of studies with a qualitative analysis (see Hsia, 2001; Huang &Huang, 2002; Lin 2005), far too few efforts have been done with a larger scale of data to better illustrate the overall changing of immigrants’ image in print media.

As a result, this research employs the content analysis method to analyze newspaper articles published in a two – year period of time – 2003, the year before the major social movement of immigrants began, and 2012, the most recent year when the data of a full year could be obtained.

Despite the fact that the circulation of printed newspapers is dropping year by year, studying the contents of newspapers is pivotal to this research since print media still plays the role of ‘leading the issue’ and ‘setting the agenda’ in Taiwan. By analyzing 707 newspaper articles, this research presents shifting trends of representations of new immigrant womens’ image in the mainstream print media. The frequency of occurrence of positive/negative images and the valuation tones of news reports will be compared; and the main factors contributing to the construction of immigrants’ image of the two periods of time will be identified.

Unlike most developed countries with a long history of cross-ethnic migration, Taiwan has just started to transform into a multi-ethnic country in the past two decades. From the results of media content analysis, this paper concludes that although Taiwanese citizens have gone through a period of ‘anxiety’ facing the unfamiliar new family members, it is statistically significant that after ten years of immigrants’ and NGO’s hard work, the media tend to represent the image of new immigrants friendlier than before. However, it remains questionable whether new immigrant women are being less ‘othered’ in the society.



1.1 Taiwan

Taiwan, officially named the Republic of China, is an island located off the southeastern coast of China, lying across the Taiwan Strait, with a total area of only 36,193 square kilometers. The population of Taiwan in 2012 was 23,268,087. This makes Taiwan the 50th largest country in the world in terms of population, and the 16th most densely populated country in the world. The vast majority (94%) of people living in Taiwan are Han Chinese, including around 12% of the population who are classified as Chinese mainlanders, who represent the people who fled from mainland China after the Chinese Civil War (and their descendants). The remaining six percent are Taiwanese Aborigines and new immigrants.


1.2 Women Migration in Asia

The migration experiences of women are becoming increasingly diversified due to forces such as globalisation in the Asia Pacific region and elsewhere. Asian women migrants have been largely perceived as either overseas workers or married though an introduction agency.The marriage migrant phenomenon started in the 1970s with mail-order brides from the Philippines marrying Australian men, and then in Japan, in the 1980s, where married migrant women from other Asian countries are called the Ajia no hanayome (Asian Brides).

1.3 Foreign Bride Phenomenon in Taiwan

The so-called ‘foreign bride phenomenon’ (Hsia, 2003) in Taiwan first appeared in the early 1980s. The migrant womenmarried to Taiwanese men through introduction agencies are mostly from China and Southeast Asiaand were seeking a more financially stable life. The number of immigrant wives has increased dramatically in recent years. According to the Ministry of Interior, 93% of the foreign spouses in Taiwan are women, and among those from Southeast Asia, 57.5% are from Vietnam, 23.2% from Indonesia, 5.3% from Thailand and another 5.3% from the Philippines. There were 4,899 registered immigrant wives in Taiwan in 1994, and the number has now reached 400,000 in 2011. Every 1 out of 7 new born babies in Taiwan have new immigrant mothers now, and the first generation of ‘new children of Taiwan’ is now turning 18, and so legally adults.

Taiwan has a long historical tradition of immigration since the 16th century when the first wave of Minnan Chinese from South-East China crossed the strait to find arable farmland in Taiwan, and the Spanish, Dutch and Japanese merchants occupied parts of Taiwan as their trading base. In the 1950s, several intense ethical conflicts occurred between the Minnan Taiwanese and Chinese mainlanders. However, this ‘foreign bride phenomenon’ is the most considerable ethnic issue since Taiwan abolished Martial Law and became democratized in the 1980s. Taiwan is now an open society, but has never truly opened its door to non-Chinese ethnic groups.

The Taiwanese society constructed the ‘foreign bride phenomenon’ as a social problem. The brides are portrayed either as passive victims or materialistic blood-suckers, and prone to committing crimes, while the bridegrooms are portrayed as ‘socially undesirable,’ that is, physically or mentally disabled, and deceivers and sexists; and their children are seen as ‘less competitive’ in schools and have often been accused of bringing a negative cultural impact to other children. (Hsia, 2001)

1.4 Don’t Call Me Foreign Bride Movement

‘A woman can only be a bride once in her life. Why are we still called a foreign bride? I am an old lady already!’ Lina Peng, originally from Indonesia, said with bit exasperation in fluent Mandarin, to the camera in the first ‘Don’t Call Me Foreign Bride’ press event in 2003. It was the beginning of the ‘Don’t Call Me Foreign Bride’ movement, which was organized by TransAsia Sisters Association, Taiwan (TASAT), the NGO formally established in the end of 2003 by a group of new immigrant women. A series of press events fighting against all kinds of discrimination were then organized within the next few months. Furthermore, TASAT has endeavored to change the public perception of immigrant women through more long-term projects such as seminars, films, and blogs. The first book collecting immigrant women’s writings, pictures and art works entitled ‘Don’t Call Me Foreign Bride’ was published in September 2005. The book, according to Hsia (2006), has successfully drawn awareness of Taiwanese reader’s prejudices toward immigrants and enabled readers to appreciate the multicultural influence brought by immigrants.



2.1 Social Representation

Social representation is an organized corpus of knowledge and one of the psychical activities that allow human beings to make physical and social reality intelligible, to insert themselves into groups or day-to-day relations of exchange and to free the powers of their imagination.(Moscovici, 1961/2008, p.xxxi)

As Moscovici (1973) suggests, social representation is a series of values, ideas and practices that constructs meaning through language in a shared context with a two-fold function: one is to establish an order that enables individuals in a social world to orient themselves; the other is to provide a code for naming and for social exchange that makes communication possible. Philogene (2001) contends that people communicate new situations and adjust to new situations by the medium of social representation.

Wagner & Hayes (2005) assert that social representation emerges when the society faces an unfamiliar or unknown situation. Novelty, as scientific theories and new technology is often regarded as threatening. It could be both an epistemic threat when encountering something unfamiliar and difficult to understand, and an ideological threat when contradict established beliefs or practices took place. To transform the strange to familiar, Wagner & Hayers (2005) further illustrate that a group needs to ‘make sense’ or ‘symbolically cope’ with the object through various forms of communication. The role of mass media in transforming and anchoring the unfamiliar to the familiar will be further discussed in a later section.

The study of representation is pivotal because it constructs and reflects existing knowledge. While meaning is never a fixed set of values, it floats as the social structures shift and are always in a process of transformation (Hall, 1980) , Lewis (1994:25) recognizes that ‘meaning becomes a battle ground between and among folk cultures, class subcultures, ethnic cultures, and national cultures. […] The sign is no longer inscribed within a fixed cultural order. The meaning of things seems less predictable and less certain.’


2.2 Mass Media

Through media, meanings circulate in a social group. It goes without saying that mass media is one of the most powerful tools to carry and transform information in modern society. The role media plays in influencing public perception about ethnic minority groups has been widely studied (Hall, 1995; Van Dijk, 2000), and it is particularly important to study the mass media effect because news plays the crucial social function of constructing a sense of reality for people who are absent from the news event. (Schlesinger, 1988) From the minority group’s point of view, the images and stories presented in the media create a so-called ‘multi-ethnic public sphere’ (Husband, 2000), where minority groups can have a sense that they are included in and contribute to the society.

Three models are often employed to describe the effect of the media in communicating representations and constructing meanings among the general public (See Scheufele &Tewksbury, 2007; Nisbet&Feldman, 2011): Agenda setting, priming and framing.

s   Agenda Setting-Shaping public priorities

Media agenda setting refers to the notion that there is a direct correlation between the emphasis that the mass media places on certain issues and the importance attributed to these issues by the general public. In other words, a news story’s prominence and the attention paid to by audiences is influenced by the amount of information provided in the media and its relative position and placement in the media, as arranged by editors. (McCombs & Shaw, 1972).Lee’s (2007) research on the production of news regarding foreign spouses in Taiwan concludes that there are two main reasons influencing the representation of new immigrant women: First, issues on minority groups are considered less important so journalists are not allowed to work only on immigrant stories. Therefore most of the journalists he interviewed stated that although they might be interested in writing something in-depth, they would not have time to do so. Secondly, editors tend to assume that the majority of readers may pay more attention to ‘Taiwanese’ stories instead of ‘foreigners’ affairs that they often purposely move articles on immigrants to later pages; which leads to the result that immigrants have long been underrepresented in mainstream media.

s   Priming-Why the news focus matters.

The concept of priming originates in social psychology studies, which are referred to as ‘activation of accessible categories or schemas in memory that influence how we process new information.’ (Hogg &Vaughan, 2008:62) Human beings are often automatically primed without self-consciousness and tend to depend on the most accessible information; for instance people who are extremely concerned about racial discrimination may find themselves encountering discriminating discourses everywhere. Applying the priming theory to mass communication study, as Mendelsohn (1996) concludes, ‘media may provoke opinion or behavioural change not because individuals alter their beliefs or evaluations of objects, but because they alter the relative weight they give to various considerations that make up the ultimate evaluation.’(p.113) That is to say, after being exposed to messages presented by the mass media, audiences might think, judge and behave accordingly.

s   Framing-Defining meaning and solutions.

Goffman (1974) contends that frames are the schemata of interpretation that individuals and groups use to make sense of the overly complex world and to process new information efficiently. Goffman suggests that people apply ‘primary frameworks’ (ibid, p.24) to interpret and objectify information meaningfully. Framing is described by McCombs et al (1997) as the second-level agenda-setting, for it is in fact an extension of agenda setting. According to Entman (1993:52), ‘to frame is to select some aspects of a perceived reality and make them more salient in a communicating context, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described.’ However, as Nisbet and Feldman(2011) put it, media frames rarely determine public opinion, they only help set the terms of the debate among citizens, and are effective only when it is applicable or relevant ‘to a specific interpretive schema acquired through socialisation processes or other types of social learning.’ (p.288)


2.3 Representation of Ethnic Minority in Media

As Said (1978) puts it in , while representations are inseparable from the representer’s culture and power structure, ‘we must be prepared to accept the fact that representation is eo ipso implicated, intertwined, embedded, inter-woven with a great many other things besides the ‘truth’ which is itself a representation.’(p.272) Since the power relation of the representer and the medium of transmitting plays a crucial role in constructing meaning and knowledge, it then becomes inevitable to identify the powerful ‘actors’ in a group that shifts the representation of the ‘others’. Cottle (2000:2) suggests that through the representations in the media, audiences are co-constructing the sense of ‘we’ and ‘not we’, ‘us’ and ‘them’, ‘insider’ and ‘outsider’, ‘citizen’ and ‘foreigner’, as well as ‘normal’ and ‘deviant.’

The reasons why the mass media is the key to the reproduction of New Racism (Barker, 1981) in Western societies, as suggests by Van Dijk (2000:36) are as follows:

1)         Most white readers have few daily experiences with minorities.

2)         Most white readers have few alternative sources for information about minorities.

3)         Negative attitudes about minorities are in the interest of most white readers.

4)         More than most other topics, ethnic issues provide positive but polarized identification for most white readers, in terms of us and them.

5)         The media emphasize such group polarization by focusing on various problems and threats for us, thus actively involving most white readers.

6)         Minority groups do not have enough power to publicly oppose biased reporting.

7)         The dominant (media) discourse on ethnic issues is virtually consensual.

8)         In particular there is little debate on the new racism.

9)         ‘Anti-racist’ dissidents have little access to the media.

In the work of ‘Policing the Crisis’, Hall (1978) asserts that mass media tend to over stress the criminalisationof minority ethnic communities, and this biased portraying originates from the pre-existing beliefs about the connection between non-white populations and criminality.

On analyzing how racism is reproduced through everydaytalk, Van Dijk(1987) concluded that the media are credited or criticized for bringingup negative stories about such groups. He suggests that the message about ethnic groups which the mass media presents shows great impact on the topical agenda of people’s everyday talk, especially in low-contact areas where people have no immediate contact with minority groups and therefore their source of information is heavily dependent on media information. His later study (1988) on the Western European and American media shows that while responding to immigration and multiculturalism, immigration is often defined as a serious problem, a threat or an invasion. And the positive contributions of immigrants in the form of cultural diversity, economic growth are rarely mentioned. Furthermore, he asserts that not only is the quality of the representation of ethnic minority groups biased; but the frequency of minorities being seen in major stories is disproportionally low, since minority organizations, leaders and other actors have less access to the media than their mainstream counterparts.

Mastro & Susannah (2003) studied the representation of ethnical minority groups in television commercials in the US and conclude that in terms of frequency, minority groups are chronically underrepresented, and in terms of presentation quality, minorities are often seen as background roles and are mostly with lower autonomy.


2.4 Image of New Immigrant Women in Taiwanese Media

The quality of news about contemporary society is important because it reflects the quality of civil society. As Lippmann (2007:172) contends, ‘the better the institutions, the more all interests concerned are formally represented, the more issues are disentangled, the more objective criteria are introduced, and the more perfectly an affair can be presented as news.’ The media freedom in Taiwan is ranked among one of the highest in Asian nations, but it is still being criticized as lacking quality content.

The first noteworthy study on the media construction of ‘foreign brides’ phenomenon is Hsia’s (2001) work with qualitative discourse analysis on 33 print media articles published between 1988 and 1996. Her study focuses on the construction of ‘social problem’ and the inferiority of immigrant groups through ‘double-bind structure’ (Hsia, 1997) She suggests that the content of selected articles from both independent and mainstream media shows a surprising homogeneity with new immigrant women in Taiwan portrayed as ‘passive victims or materialistic blood-suckers, and prone to committing crimes’ (p.154) During her interview with several journalists and editors, she concludes that in the reproduction of ‘foreign brides’ image, the process of making the news is instead the process of silencing a channel for diverse voices. Hsia’s work opened the door for studies of new immigrant women, however, her study has been criticized by Lin (2005) for the thirty three articles she analyzed were purposely selected to show the negative reproduction of foreign brides and has overlooked the positive reports.

Huang & Huang (2002) analyzed the representation of ‘foreign brides’ in the print media in 2001 and concludes that new immigrant women in Taiwan, as is the case for other minority groups, are mostly under-represented as threatening and problem-causing. However, through small scale quantitative research, they suggest that there are certain positive images of ‘foreign brides’ shown in the news stories. Typical images are as hard-working, gentle and caring. Feng (2003) contends that while covering the story of immigrants in Taiwan, the media tend to stress on their negative traits. ‘Foreign brides’ are commonly portrayed as exotic, mysterious or even strange; and they are often seen in incidents as having sexual affairs, committing crimes, and engaging in prostitution.



3.1 Content Analysis

‘It would be very useful to devise some system for monitoring the extent to which certain newspapers and chains of newspapers distort news concerning disarmament in the world.’

-MacBride (1980: ii)

To construct a broader view of the representation of new immigrant women in the print media, this research adapts content analysis as the main method. Content analysis is one of the most common methods to study mass media content, and it is defined by Berelson (1952:18) as ‘a research technique for the objective, systematic and quantitative description of the manifest content of communication.’ Furthermore, the importance of content analysis procedure as being reliable and its result being replicable is stressed. (Krippendorf, 2004) As Wimmer and Dominick (1983) concluded, there are five main purposes for content analysis:

1)     Describing communication content

2)     Testing hypotheses of message characteristic

3)     Comparing media content to the ‘real world’

4)     Assessing the image of particular groups in society

5)     Establishing the study of media effect

With the advantage of being transparent, relatively more objective and capable of handling massive amounts of data, content analysis is widely employed to analyse news articles, visual images as well as audio data. (Bryman, 2001) Content analysis is regarded as the method that allows researchers to construct indicators of ’world views, values, attitudes, opinions, prejudices and stereotypes, and compare these across communities.’ (Bauer, 2000:134)’


3.2 Longitudinal Research

In order to compare the image of new immigrant women in print media before and after the ‘don’t call me foreign bride’ movement, this research also employs the longitudinal method. The two main purposes of longitudinal research, according to Menard (2002), are to present the patterns of change and magnitude of causal relationships. The patterns of change may be the change in direction (positive to negative, or from any given X to Y, or vice versa.) As Menard (2002) puts it, once the data is collected over two or more distinct periods with the same sets of variables on the same topic, the research may be regarded as longitudinal. By repeating the data analysis of year 2003, the year before ‘don’t call be foreign bride’ movement starts, and 2012 with the same coding frame, this research is then able to show the pattern of change of the media representation of new immigrant women in Taiwan.


3.3 Data Collection

Data Base

The database this research employs is ‘iNews system ’ provided by the Parliamentary Library, which covers all published print media articles in Taiwan. To focus more on the main stream media, only four national-leveled daily newspapers were chosen for this research; Liberty Times, China Times, United Daily, and Apple Daily, known as the ‘big four’, which in total covers 94% of the market share among 32.3% (Rainmaker XKM International, 2012) of the Taiwanese population who reads newspapers on a daily basis. China Times and United Daily are two of the oldest newspapers in Taiwan and the mass readers for both are middle class, pro-China Taiwanese. Liberty Times was first published in 1980, in the post- martial law era of Taiwan, with a more pro-independence political stance and it is the country’s second largest circulation. Apple Daily, first published in May 2003, is now the best-selling newspaper in Taiwan.

Four Taiwanese-Mandarin terms most commonly used to stand for new immigrant women in a daily linguistic context are selected as keywords for this research:

  • Foreign Bride (外籍新娘);
  • Foreign Spouse (外籍配偶);
  • ‘Wai Pei’ (外配) (abbreviation of foreign spouse, commonly used in media)
  • New Immigrant (新移民).


Sample Size

3,291 articles with the keywords ‘Foreign bride OR Foreign Spouse OR Wai Pei OR New immigrants’ were found and all articles were examined and unrelated articles deleted. Considering the reasonable limitation for one researcher to conduct the coding, this research reduces the sample size by the researcher selecting every 4th article manually. This brought the total sample size to 707 (344 for year 2003 and 363 for year 2012). As the search result of iNews is arranged chronologically, we may view it as a random selection. Numbers and percentages of articles from each paper in the sample are as seen in the listed results:


Year Liberty Times China Times Apple Daily United Daily Total




































Total Number of Articles Analysed 707




3.4 Categories and coding

A beautiful coding frame is one that is internally coherent and simple in the sense that all codes flow from a single principle, rather than being grounded in the ‘dust-bowl’ empiricism of coding whatever comes to mind. (Bauer, 2000:141)

Categorizing, as suggested by Franks (1999), reduces the complexity of data sets, and allows the interpretation of data to be clearer. Constructing a transparent, reliable, replicable and exhaustive coding frame is crucial to a successful content analysis. To do so, all categories must be clear and mutually exclusive so that each unit can only fit one category. Most importantly, the coding frames must allow researchers to answer the research question of concern. (Marying 2000, Holtsi 1969)

The coding units in this research are by articles instead of paragraphs or sentences, to better grasp the general tone of each news report. The coding frame of this research is adapted from the system widely employed by most Taiwanese governmental departments, which the PR officers use as a daily media-monitoring tool. The coding frame is first set up according to previous studies on ‘foreign brides’ in the early 2000s (Hsia 2001, Lin 2005), and has been adjusted according to a random piloting of 100 articles. The final coding frame is as listed below:


Variable Definition
1 Date In the form of YEAR/Month/Day
2 Newspaper Liberty Times
China Times
United Daily
Apple Daily
3 Size Small < 400 words
Medium 400 words< X < 800 words
Large > 800 words
4 News Format News
5 Primary Source Government State and local government agents/individuals
NGOs All nongovernmental organizations
Academic Academic institutions/Scholars
Immigrant Immigrant women and their relatives
Other Other sources
6 Valuation Very Positive The strong positive image is directly linked to the immigrants and is indicated in the title.
Fairly Positive The fairly positive image is directly linked to the immigrants in the content
Slightly Positive Immigrants are somewhat linked with a positive image, but might be indirect.
Neutral The article does not show any valuation.
Slightly Negative Immigrants are somewhat linked with a negative image, but might be indirect.
Fairly Negative The fairly negative image is directly linked to the immigrants in the content
Very Negative The strong negative image is directly linked to the immigrants and is indicated in the title.
7 Positive Image Smart and talented
Willing to learn
Caring and helpful
Politically Influential
8 Negative Image Less educated
Low autonomy
Wasting social resources
Low Loyalty/Easily divorced
Incapable of taking care of children
9 Theme Policy



Coding was done with Microsoft Excel. Apart from the categories listed above, titles and the full articles were attached in the file to facilitate double checking.




This research evaluates the negative and positive value of newspaper articles with magnitudes and scales. From very negative to very positive, the valuations are presented in seven levels. When magnitudes and scales are used, Krippendorf (2004:136) suggests that ‘coders are expected to conceptualize the meaning of texts as continua, as having more or less of something, as possessing a metric.’ He further points out that semantic differential scales could be unreliable when the information about the characteristics to be recorded is not clearly defined. The more the coders have to guess, the less likely the result could turn out reliable.


3.5 Reliability

A process that is reliable, replicable, and yields valid is pivotal to any content analysis research. Krippendorf (2004) said the research procedure is reliable when it responds to the same phenomena in the same way regardless of the circumstances of its implementation. He identifies three types:

1)          Stability –The process should be unchanging over time, and coding results by the same coder must be consistent. It is also known as ‘intra-observer reliability’

2)          Reproducibility The process should be replicable under different conditions, at different locations, using different coders. It is often termed ‘inter-coder reliability’, or ‘inter-subjective agreement’.

3)          Accuracy – The process should functionally conform to an existing standard. That is to say, the result from one coder should be compared with the result that is known to be correct.

A second independent coder was involved in the piloting to test the inter-coder reliability of this coding frame. The second coder was chosen for he is a public communication professional and has three years of experience in conducting media monitoring and content analysis for the Government Information Office of Taiwan. Therefore the second coder did not require much previous training. The second coder coded 100 randomly selected articles for piloting, 50 articles from year 2003 and 50 articles from year 2012. All disagreements were discussed thoroughly between the author and second coder in first piloting, and then a revised version of the coding frame was employed. Inter-coder reliability is measured on a scale between 0, which stands for no agreement and 1, which indicates perfect agreement, and reliability is generally acceptable in the range 0.66r<0.79, high at r>0.8 and very high at r>0.9 (Bauer, 2000). The result of inter-coder reliability is presented as seen below:

Category Inter-coder Reliability Category Inter-coder Reliability
V1 1 V6 0.81
V2 1 V7 0.90
V3 1 V8 0.87
V4 1 V9 0.99
V5 0.94

The calculated inter-coder reliability for each variable is either very high or high, therefore the inter-coder reliability for this research shall be considered methodologically acceptable.



Data for coding is first collected on Microsoft Excel and further exported to Statistical Package for the Social Science (SPSS).


Don’t call me foreign bride?

Before analyzing the result of coding, the change in frequency of the term ’foreign bride’ being written in newspapers provides a broad view on the achievement of ‘don’t call me foreign bride’ movement. In 2003, 876 articles using the term ‘foreign bride’ were found in the four major newspapers – Liberty Times, China Times, United Daily and Apple Daily. The number increased a little in 2004, presumably because Apple Daily was not published until May 2003. The number of articles using ‘foreign bride’ to refer to new immigrant women started to drop significantly from 2007, which is two years after the book ‘Don’t Call Me Foreign Bride’ was published. Since 2009, less than 100 articles used the phrase ‘foreign bride’ per year. It could be concluded that in terms of reducing discriminative terminology, the movement was quite successful.


4.2 Data Description

The total number of articles analyzed is 707, with 344 articles from 2003 and 363 articles from 2012. Among the four newspapers, 269 articles are from United Daily; 235 articles are from China Times; 165 articles are from Liberty Times, and only 38 articles are from Apple Daily. Apple Daily is the print media with the highest circulation rate in Taiwan, but with its tabloid-style, it focuses less on immigrant social issues.

Year * Newspaper Cross tabulation
Newspaper Total
China Times Apple Daily United Daily Liberty Times
Year Year 2003 165 18 123 38 344
Year 2012 70 20 146 127 363
Total 235 38 269 165 707

In 2003, most articles are found in the last two quarters of the year. Issues on new immigrant women received large media attention in August and May, when the discussion of whether new immigrant mothers are more likely to give birth to mentally retarded babies occurred. In 2012, the policy debate took place in May on whether or not new immigrant women could donate breast milk to local hospitals to support babies in need of better nutrition attracted massive attention.


In terms of size of articles, more than 80% of the articles found in both 2003 and 2012 were between 400 and 800 words. Only 16% in 2003 and 13.8% in 2012 of the articles were larger than 800 words, the average words per article for major topics in Taiwanese newspapers, from which we could conclude that though new immigrant women issues are quite controversial in the society, it did not attract much attention from the mainstream media.

Year Frequency Percent Valid Percent Cumulative Percent
Year 2003 Valid small 24 7.0 7.0 7.0
medium 265 77.0 77.0 84.0
large 55 16.0 16.0 100.0
Total 344 100.0 100.0
Year 2012 Valid small 58 16.0 16.0 16.0
medium 255 70.2 70.2 86.2
large 50 13.8 13.8 100.0
Total 363 100.0 100.0



4.3 Positive and Negative Valuations

In 2003, the year before ‘Don’t Call Me Foreign Bride’ movement began, 54.7% of the articles found were slightly negative, negative or very negative, whereas in 2012, nearly ten years after the movement, the percentage of negative valued articles dropped to 27.8%. 14 articles had very negative valuations in 2003 and 4 very negative reports were in 2012. In terms of positive valuations, no article in 2003 had a very positive valuation, and only 19.5% were slightly positive or positive, while in 2012, 39.1% of the articles received positive valuations, and 3 were very positive.

Year Frequency Percent Valid Percent Cumulative Percent
Year 2003 Valid Very Negative 14 4.1 4.1 4.1
Negative 51 14.8 14.8 18.9
Slightly Negative 123 35.8 35.8 54.7
Neutral 89 25.9 25.9 80.5
Slightly Positive 53 15.4 15.4 95.9
Positive 14 4.1 4.1 100.0
Total 344 100.0 100.0
Year 2012 Valid Very Negative 4 1.1 1.1 1.1
Negative 25 6.9 6.9 8.0
Slightly Negative 72 19.8 19.8 27.8
Neutral 120 33.1 33.1 60.9
Slightly Positive 115 31.7 31.7 92.6
Positive 24 6.6 6.6 99.2
Very Positive 3 .8 .8 100.0
Total 363 100.0 100.0



The mean valuation with a scale of 1 to 7 in 2003 was computed by SPSS as 3.46, with a standard deviation of 1.160, whereas the mean in 2013 was commuted as 4.10, with a standard deviation of 1.105. One can conclude here that the overall valuation of immigrants’ image in newspapers has shifted in a more positive direction in the past ten years.

To create a clearer view, this research redefines the valuation variables by combining very negative, negative and slightly negative as ‘Negative’, and very positive, positive, slightly positive as ‘Positive’. The output of SPSS shows that in 2003, 54.7% of the articles are negative, whereas in 2012, the percentage drops to 27.8%. Meanwhile, the percentage of overall positive-valuated articles observed in 2003 is 19.5%, and in 2012, the percentage rose to 39.1%. The result of the chi-square test, p<0.001 shows that the differences in valuation tone between 2003 and 2012 is significant at all levels of significance.


Year * Valuation2 Cross tabulation
Valuation 2 Total
Negative Neutral Positive
Year Year 2003 Count 188 89 67 344
% within Year 54.7% 25.9% 19.5% 100.0%
% within Valuation 2 65.1% 42.6% 32.1% 48.7%
Year 2012 Count 101 120 142 363
% within Year 27.8% 33.1% 39.1% 100.0%
% within Valuation 2 34.9% 57.4% 67.9% 51.3%
Total Count 289 209 209 707
% within Year 40.9% 29.6% 29.6% 100.0%
% within Valuation 2 100.0% 100.0% 100.0% 100.0%




Chi-Square Tests
Value df Asymp. Sig. (2-sided)
Pearson Chi-Square 57.233a 2 .000
Likelihood Ratio 58.229 2 .000
Linear-by-Linear Association 55.070 1 .000
N of Valid Cases 707
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 101.69.


By comparing the sources of news and its valuation tone, one can identify the actors who contributed most to the formation of representations of minority groups in society. In 2003, 81.3% of the news released by or mainly with quotations from politicians had a negative valuation tone; 58.4% of the news articles with the government (central and local government units) being a primary source were negative. Whereas the percentages of negative news of other sources are: media -57.9%, NGO -55.8%, academic -38.5%, other -50.0%, immigrants -7.7%. In 2012, the percentage negative tone found in news released by politicians’ dropped to 64.5%, and the figure for the government significantly dropped to 25.7%. On the other hand, not surprisingly, immigrants were rarely the source of negative news, with 46.2% of the articles found in 2003 and 62.9% of the articles found in 2012 with immigrants or their relatives being the primary source had a positive valuation tone.


Primary Source * Valuation 2 Cross tabulation
Negative Neutral Positive
Year 2003 Primary Source Government Count 97 49 20 166
% within Primary Source 58.4% 29.5% 12.0% 100.0%
NGO Count 24 8 11 43
% within Primary Source 55.8% 18.6% 25.6% 100.0%
Academic Count 5 2 6 13
% withiSourcen Primary 38.5% 15.4% 46.2% 100.0%
Immigrant Count 1 6 6 13
% within Primary Source 7.7% 46.2% 46.2% 100.0%
Media Count 11 4 4 19
% within Primary Source 57.9% 21.1% 21.1% 100.0%
Politician Count 13 2 1 16
% within Primary Source 81.3% 12.5% 6.3% 100.0%
Other Count 37 18 19 74
% within Primary Source 50.0% 24.3% 25.7% 100.0%
Total Count 188 89 67 344
% within Primary Source 54.7% 25.9% 19.5% 100.0%
Year 2012 Primary Source Government Count 36 52 52 140
% within Primary Source 25.7% 37.1% 37.1% 100.0%
NGO Count 11 22 31 64
% within Primary Source 17.2% 34.4% 48.4% 100.0%
Academic Count 2 6 8 16
% within Primary Source 12.5% 37.5% 50.0% 100.0%
Immigrant Count 4 9 22 35
% within Primary Source 11.4% 25.7% 62.9% 100.0%
Media Count 3 10 8 21
% within Primary Source 14.3% 47.6% 38.1% 100.0%
Politician Count 16 4 5 25
% within Primary Source 64.0% 16.0% 20.0% 100.0%
Other Count 29 17 16 62
% within Primary Source 46.8% 27.4% 25.8% 100.0%
Total Count 101 120 142 363
% within Primary Source 27.8% 33.1% 39.1% 100.0%



Chi-Square Tests
Year Value df Asymp. Sig. (2-sided)
Year 2003 Pearson Chi-Square 32.358a 12 .001
Likelihood Ratio 33.628 12 .001
Linear-by-Linear Association 2.474 1 .116
N of Valid Cases 344
Year 2012 Pearson Chi-Square 46.490b 12 .000
Likelihood Ratio 44.364 12 .000
Linear-by-Linear Association 10.351 1 .001
N of Valid Cases 363
a. 8 cells (38.1%) have expected count less than 5. The minimum expected count is 2.53.
b. 1 cell (4.8%) has expected count less than 5. The minimum expected count is 4.45.

The p-values of chi-square test in 2003 is 0.001 and in 2012 is lower than 0.001 show that the differences in valuation tone between the seven sources both in 2003 and 2012 are significant at all levels of significance. It is clear that politicians and government officers are the main actors constructing negative discourses of new immigrant women.



4.4 Images of Immigrants

Positive Images

In 2003, among the 67 articles with positive images of immigrant women, 29.9% portrayed immigrant women as ‘willing to learn’; the second most frequently shown image was ‘smart and talented’ (20.9%), and the third was ‘caring and helpful’ (11.9). In 2012, among the 144 articles, 34 of them indicated the rich culture immigrant women contributed to Taiwanese society; 33 articles portrayed immigrant women as ‘caring and helpful’; and the third most frequent image was ‘smart and talented.’

It is asserted by Downing &Husband (2005:28) that one of the methodological drawbacks associated with content analysis is its ‘frequent failure to acknowledge the multiple ways in which language is used, including irony, parody, sarcasm, rhetorical over and understatement, phrases carrying a particular symbolic charge at a given moment in time […], and so on.’ In other words, they suggest that the positive valuation observed by this coding mechanism may in fact still contain a certain degree of discrimination or negative traits.



Positive Image
Year Frequency Percent Valid Percent Cumulative Percent
Year 2003 Valid Smart and talented 14 4.1 20.9 20.9
Willing to learn 20 5.8 29.9 50.7
Caring and helpful 8 2.3 11.9 62.7
Rich Culture 7 2.0 10.4 73.1
Politically influential 4 1.2 6.0 79.1
Other 14 4.1 20.9 100.0
Total 67 19.5 100.0
Missing 4 1 .3
None 276 80.2
Total 277 80.5
Total 344 100.0
Year 2012 Valid Smart and talented 29 8.0 20.1 20.1
Willing to learn 12 3.3 8.3 28.5
Caring and helpful 33 9.1 22.9 51.4
Rich Culture 34 9.4 23.6 75.0
Politically influential 6 1.7 4.2 79.2
Other 30 8.3 20.8 100.0
Total 144 39.7 100.0
Missing 4 1 .3
None 218 60.1
Total 219 60.3
Total 363 100.0



Looking deeper into the positive images of immigrant women presented in newspaper articles, the most frequently seen in 2003 was ‘willing to learn’. By praising the immigrants being ‘willing to learn’, the article was in fact implying that they e lack  certain abilities to fit into  society and that they need to ‘learn’ more. One example is as seen below:

Fan, Vietnamese married to a Taiwanese man 15 years ago completed the culture course in community college today […] Fan said that she once chopped mustard leaves into small pieces when she was cooking New Year’s Eve dinner for the family. Her mother-in-law felt offended and completely mad because mustard leaf symbolizes ’long life’ in Chinese culture. ‘Now I will never make this mistake again.’ She said to the press. (United Daily, 01/06/2012)

Similarly, while new immigrant women are valued as being ‘caring and helpful’, this reflects the fact that the ‘value’ of foreign brides is primarily seen as two-fold: a guarantee of offspring and to serve as a productive laborer. (Chen &Lin, 2006) They are expected to be both a good mother and a housekeeper, instead of an independent individual. Furthermore, while new immigrant women or their children perform well, the media tend to report the story as an ‘exception’, as if they should ‘normally be inferior.’ One example is as seen below:

Children of a foreign spouse rarely perform well  at school, but Chu is an exception. Her mother is Pilipino who speaks English fluently and helped with Chu’s school work. (China Times, 05/11/2003)


Negative Images

Among the 194 articles evaluated as negative in 2003, immigrant women were most frequently portrayed as ‘less educated’ (25.3%), ‘Incapable of taking care of children’ (16.5%) and ‘vulnerable’ (10.3%); where as in 2012, among the 104 articles with negative valuations, immigrant women were most frequently described as ‘vulnerable’ (28.8%), ‘committing crime’ (21.2%) and ‘less educated’ (10.6%).


Negative Image
Year Frequency Percent Valid Percent Cumulative Percent
Year 2003 Valid Less educated 49 14.2 25.3 25.3
Incapable of taking care 32 9.3 16.5 41.8
Vulnerable 20 5.8 10.3 52.1
Low Autonomy 15 4.4 7.7 59.8
Wasting Social Resource 3 .9 1.5 61.3
Lower Loyalty 19 5.5 9.8 71.1
Crime 15 4.4 7.7 78.9
Other 41 11.9 21.1 100.0
Total 194 56.4 100.0
Missing None 150 43.6
Total 344 100.0
Year 2012 Valid Less educated 11 3.0 10.6 10.6
Incapable of taking care 6 1.7 5.8 16.3
Vulnerable 30 8.3 28.8 45.2
Low Autonomy 7 1.9 6.7 51.9
Wasting Social Resource 4 1.1 3.8 55.8
Lower Loyalty 5 1.4 4.8 60.6
Crime 22 6.1 21.2 81.7
Other 19 5.2 18.3 100.0
Total 104 28.7 100.0
Missing None 259 71.3
Total 363 100.0



It is worth noting that in 2003; a piece of news titled ‘the incidence of developmental retardation is higher among foreign brides’ children’ was released by a public hospital:

More and more Taiwanese citizens are marrying foreign brides nowadays. It has been discovered that the newborns of Southeast Asian mothers fall into a high-risk bracket for many genetic disorders, including congenital growth retardation. (United Daily, 04/05/2003)

In the news release, new immigrant women were directly accused of having genetic disorders, but the causation for retardation from medical science’s point of view was not explained. This topic provoked an intense discussion on whether ‘foreign brides’ have ‘contaminated’ Taiwanese’s blood. Several opinions on whether the government should provide education on birth control to the new immigrant women and their spouses were discussed. Whereas statistics from the Ministry of Interior in 2003 showed that the incidence of retardation of newborn babies among all ethnic groups was around 0.1%, no significance was shown in differences between ethnic groups. Lan (2005) contends that such discourse showed the spirit of ‘racial hygiene’ still existed among political elites and reflects the anxiety of the ‘pure’ Taiwanese being mixed by ‘others’. Such anxiety also reflects in the data sample that 29 out of 35 health-related articles were slightly negative, negative or very negative in 2003.

One of the key statements of ‘Don’t Call Me Foreign Bride’ movement by new immigrant women is that they have the right to teach their children in their native language, and the fact that they do not read and write Chinese well does not mean that they are illiterate. In the articles found in 2003, new immigrant women were most frequently illustrated as ‘less educated’ (25.3%), mainly because Mandarin Chinese is not their mother tongue, and when problems occurred due to language or cultural barriers, the media tended to report them as ‘less educated.’



4.5 Who talks about new immigrant women?


To figure out what the primary sources of negative or positive news are is pivotal in identifying the power structure of discourse within a society. From the data selected, we can see that the government is always the primary source of immigrant-related news, where it released a total of 48.3% of the immigrant-related news in 2003, and 38.6% in 2012. NGOs were the second most frequently observed source in the data, while in 2003, 12.5% of the news was from NGOs, and the percentage in 2012 rose slightly to 17.6%.


Primary Source
Year Frequency Percent Valid Percent Cumulative Percent
Year 2003 Valid Government 166 48.3 48.3 48.3
NGO 43 12.5 12.5 60.8
Academic 13 3.8 3.8 64.5
Immigrant 13 3.8 3.8 68.3
Media 19 5.5 5.5 73.8
Politician 16 4.7 4.7 78.5
Other 74 21.5 21.5 100.0
Total 344 100.0 100.0
Year 2012 Valid Government 140 38.6 38.6 38.6
NGO 64 17.6 17.6 56.2
Academic 16 4.4 4.4 60.6
Immigrant 35 9.6 9.6 70.2
Media 21 5.8 5.8 76.0
Politician 25 6.9 6.9 82.9
Other 62 17.1 17.1 100.0
Total 363 100.0 100.0


It is worth noting that in 2003, the year before ‘Don’t Call Me Foreign Bride’ movement began; only 3.8% of the news observed was

released by or mainly quoted from new immigrants or their relatives, whereas in 2012, the percentage of immigrants being the primary source of a news report rose to 9.6%. One of the practical reasons for immigrants’ voice being absent from mainstream media, as suggested by Lee (2007) is that journalists and editors rarely have personal connections with new immigrants, therefore, in the fast-paced media industry; they tend to interview the easy-access sources who they consider more ‘trustworthy’—NGOs and government officials. Nevertheless, this short-cut for journalists has made new immigrant women the ‘symbolic annihilation’ (Tuchman, 1978), where their voices are absent even with the issue directly relating to themselves.





The methods which have been applied to media textual analysis vary, and content analysis is widely adopted as the heavily quantitative method. However, as Downing & Husband (2005) conclude, its major drawback is its ‘frequent failure to acknowledge the multiple ways in which language is used, including irony, parody, sarcasm, rhetorical over and understatement, phrases carrying a particular symbolic charge at a given moment in time […], and so on.’(p.28). This research focuses mainly on interpreting the quantitative results and puts only little effort on examining the content qualitatively. Since the full contents of newspaper articles are included in the data collected by this research, it would be easy to be employed as the database for future researches.

The second limitation of this research is that it targets only on Southeastern ‘foreign brides’, however Southeastern immigrants only account for around 30% of the population of foreign spouses, whereas Chinese immigrants account for 65%. Chinese new immigrant women are culturally and ethnically closer to local Taiwanese. (See Wang, 2005; Tsai, 2004). Therefore, with the comparison of the differences and similarity of how media portray the two groups, we can better analyze the representations of immigrant women in the Taiwanese media. Furthermore, since both Chinese and Southeastern Asian immigrant women are called ‘foreign brides’ in earlier times, it was unavoidable that ‘Chinese foreign brides’ were included in some of the articles selected, which shall be seen as a bias for data collection.

Thirdly, the iNews system provides easy access to all published newspaper articles in Taiwan, and its searching system allows researchers to quickly screen out the necessary information. However, its drawback is that all articles are text-only, where photos, tables and any other supporting visual materials are absent. In addition, the placement of articles on the pages and the size and typesetting of headlines are not shown in the system as well, which makes it difficult to analyze the attention structuring of any particular story.


Proofreading and editing service