Interest Groups in National Reading Policy: Perceived Influence and Beliefs on Teaching Reading
Mengli Song, University of Michigan
Cecil Miskel, University of Michigan
D uring the past decade, a sense of anxiety grew up around the reading performance of America's schoolchildren. Business leaders expressed concern that insufficient reading skills among the workforce were reducing America's international competitiveness; the public and policymakers lamented California's plunging reading scores on the NAEP; and scholars pointed to the dire personal, social, and economic consequences of reading failure. Sensing the nation's changing mood, political leaders placed reading high on the policy agenda. Former President Clinton and then-Secretary of Education Riley set as a top national priority that every child should read well and independently by the third grade. Representative Bill Goodling (R-PA), former chair of the House Committee on Education and the Workforce, held hearings and mobilized his colleagues in the House of Representatives to address the issue of reading achievement. With reading moving to the top of the political agenda, reading specialists promoted alternative solutions, and policymakers started creating program initiatives to address the reading problem. The two most prominent initiatives--the America Reads program announced by former President Clinton in 1996, and the Reading Excellence Act of 1998 initiated by then-Representative Goodling (McDaniel, Sims, & Miskel, 2001)--represented the advent of a new era in national reading policy. President Bush has made the improvement of reading a key component of his national education plan, thereby apparently signalling that he will keep the window of change open.
With the policy agenda for reading in flux, a growing number of interest groups and policymakers are trying to sway reading policy with their favored answers. Indeed, reading policy is a hotbed of controversy: those involved in this arena agree that the acquisition of reading skills is of fundamental importance to the well-being of individuals, society, and our economic system, but concur on little else. Debates about reading policy are increasingly pervasive, intense, and polarized. Driven by a range of ideologies, personal experiences, and conflicting research findings, parent advocacy groups, politicians from across the political spectrum, professional and business associations, and academics are joining debates around a number of issues, (e.g., the magnitude of the reading problem, pedagogy, standards and assessments).
The most polarized conflict is about the best way to teach reading. While a number of approaches are available, they are typically described in contentious terms, with war metaphors pitting the whole language camp against the phonics supporters. Calfee and Norman (1998, p. 244), for example, observe that "A battle is raging." Phonics proponents argue that reading achievement is in perilous decline because schools rely on whole language methods in their reading programs. Whole language advocates counter that phonics uses boring drills and practice, and delays children's access to real books. Firefights have been instigated or fueled by a plethora of writers--for example, Adams (1990), Allington (1997), Chall (1996; 2000), Cheney (1997), Coles (1998), Finn (1995), Goodman (1986), Moats (2000), and Shannon (1998).
In the conflict-ridden debate over this essential area of schooling, our knowledge about how reading policies develop is insufficient (Valencia & Wixson, 2000) and our understanding of the critical roles of interest groups and their beliefs in shaping educational policies is inadequate (Malen, 2001). To address these important needs, we imbued our investigation with a twofold purpose--first, to gauge the relative support of various interest groups for different approaches to teaching reading; and second, to examine whether interest groups' beliefs about teaching reading vary with their perceived level of influence.
The category of "interest groups" is defined for our study as including membership organizations, advocacy organizations not accepting members, businesses, other organizations or institutions, or any association of individuals, policy actors or groups, whether formally organized or not, that tries to influence public policy (Baumgartner & Leech, 1998; Hrebenar, 1997; Kollman, 1998; Sipple et al., 1997; Thomas & Hrebenar, 1991). Hence, a wide variety of organizations qualify as interest groups in the reading policy arena, including K-12 and higher education associations and unions, parent and citizen groups, think tanks or policy institutes, government agencies, media, business, and foundations. Specific examples include the American Federation of Teachers (AFT), the National Education Association (NEA), the Brookings Institution, the International Reading Association (IRA), and Education Week. Reading achievement, phonics, and whole language are not in themselves interests, but become so when those who care about reading demand that the government promote their values by means of public policy (e.g., Baumgartner & Leech, 1998; Salisbury, 1991). As the focus on early reading intensifies, many values become interests, and many private organizations become interest groups as they attempt to insure that their values are represented in the reading policy arena.
Sabatier and Jenkins-Smith (1993, 1999) theorize that the belief system of an advocacy group is not only essential for the group's formation, but also guides the group's perceptions and actions in the political arena. Group actions in turn influence policymakers' responses and the subsequent policy outcomes. One key set of commonly held beliefs about reading focuses on how reading should be taught, especially in the early grades. Phonics advocates call for explicit, direct, and systematic skills instruction that emphasizes the alphabetic principle, phonemic awareness, word recognition, decoding, and the relationship between sounds and spelling (Adams, 1990; Chall, 1996). In contrast, proponents of whole language underscore the importance of literature-based reading, meaning construction for purposeful functions, student-centeredness, teacher empowerment, and the naturalness of reading acquisition in reading instruction (Bergeron, 1990; Edelsky, 1993; Goodman, 1986; Weaver, 1994).
Chall (1996) chronicles the changing battle lines in this reading war. Meaning and communication-based approaches dominated reading instruction for much of the 20th century. Whole word and sight methods reigned from the 1920s through the 1960s, as whole language approaches did during the 1980s and 1990s. From the late 1960s onward, phonological methods based on learning the alphabetic principle--that is, the relationship between letters and their sounds--came into wide usage and dominated the teaching of reading. The list of interest groups touting phonics includes the Republican Party, the Christian Coalition, the American Family Association, and the Eagle Forum (Burnett, 1998). According to Ravitch (2000), whole language built a large and committed following in schools of education and professional associations such as IRA and the National Council of Teachers of English (NCTE) during the 1980s. Whole language proponents also won a major victory in 1987, when California adopted guidelines for teaching English that were driven by the tenets of whole language. During the past decade, a full-scale war for dominance raged between the two groups in a manner that Chall (2000) described as less reasoned than exchanges in earlier periods. Phonics advocates won major skirmishes during the 1990s, when the 1992, 1994, and 1998 NAEP scores showed drops in California's reading achievement. The state eventually rewrote its guidelines to emphasize phonemic awareness. Even some ardent whole language proponents such as Wolfe and Poynor (2001) acknowledge the decline of whole language, but attribute this decline to a campaign led by researchers, the popular media, and the religious right.
During different periods of the 20th century, phonics or whole language alternated in their domination of the way teachers taught reading. Whether in vogue or out, each side adamantly maintained its beliefs and conceded little to the other. Over the years, many efforts have been made to integrate the two approaches. Hirsch (1996) observes that middle-of-the-road approaches which use ideas from both phonics and whole language are needed. Indeed, calls for a truce have become increasingly strong during the late 1990s. For example, a report by the National Research Council (Snow, Burns, & Griffin, 1998) concluded that effective reading instruction integrates the best practices of different methods. The Learning First Alliance (1998), a collaborative of educational associations, calls for complex approaches that include phonics instruction. Moreover, both the IRA and AFT reversed their official policies and now specifically promote early phonics as part of a comprehensive reading program (Palmaffy, 1997). A number of scholars (e.g., Chall, 1996; Flippo, 1997; Lyon, 1998; Pressley, 1998; Stahl, 1992; Stanovich, 1990) argue that research evidence supports the balancing or integration of decoding and comprehension skills with elements of whole language, to create an effective and attractive early literacy curriculum. The various thoughtful and eclectic approaches to reading instruction go by a variety of names, including "comprehensive," "integrated," and "individualized" instruction, but the most common term is "balanced." While each side seems unwilling to entertain the other side's perspectives, powerful interests are talking about compromise and consensus on the question of how to teach reading. Hence, Hypothesis 1 is that interest groups in the national reading policy arena voice significantly stronger support for balanced approaches to reading instruction than for phonics or whole-language approaches.
Interest groups share the fundamental goal of swaying policymakers to incorporate the group's beliefs into public policy. The extent of their success depends largely on their relative influence. By definition then, the beliefs of groups who are perceived by policymakers as more influential carry more weight in the policymaking process than the beliefs of those groups perceived as less influential. The question of whether interest groups possessing different levels of influence hold different beliefs about how to teach reading--and if so, of how their beliefs differ--then becomes important. Current literature suggests that the influence of policy actors may be linked to their beliefs through the social network locations of the actors.
Influence, perceived influence, and network location. Influence is a fundamental dimension of power. As Knoke (1990) observes, influence occurs when one actor intentionally transmits information to another, thus altering the recipient's actions from what would have occurred without that information. Phrased somewhat differently, Bacharach and Mundell (1993) define influence as the informal dimension of power, consisting of attempts to persuade those with authority to make a decision along certain lines. While acknowledging that influence and power are inseparable, Pfeffer (1992) views influence and politics as the processes, actions, and behaviors through which potential power is utilized and realized.
Playing a crucial role in organizational behavior, power and influence originate from a multitude of sources, including money; political, organizational, and managerial skill; membership size; geographical distribution; group image or standing; membership attributes; affinity to coalitions; and member cohesion and dedication (Etzioni, 1964; Greenwald, 1977; Rosenthal, 1998). It would be virtually impossible to come up with an exhaustive list of the factors that contribute to group influence. Even if such a list could be constructed, it would still be impossible to precisely specify the equation for calculating group influence based on this set of factors, especially when we take into account the dynamic, transitory, and contextual nature of power and influence (Greenwald, 1977; Knoke, 1990; Mitchell, Agle, & Wood, 1997).
Indeed, the difficulty of measuring the slippery, albeit crucial, concept of influence has been troubling social scientists for decades, and has been regarded as a methodological challenge of the highest order. Baumgartner and Leech (1998) observe that studies designed around the false premise that we can observe the actions of influence and power are doomed to fail because they are intended to measure the unmeasurable
Examining perceptions of influence, instead of influence itself, not only makes the task methodologically more feasible, but also has logical appeal. After all, any influence attributed to interest groups must be mediated through policymakers before it can have an impact on policy decisions. Arguably, it is policymakers' perception of influence--rather than influence per se--that really counts. The perception of a group's influence by other groups may also have implications for coalition-building. Hence, for the purpose of this study, perceived influence is defined as the relative power that a group possesses in the perceptions of policymakers or other interest groups. We contend that this concept can be captured by standard interview schedules in a relatively simple and straightforward manner, without losing the measure's substantive relevance.
Beyond a group's own resources and membership characteristics, its influence also depends on its location within the social network in which it is embedded. Knoke (1990), for instance, asserts that network centrality is synonymous with influence, for the greater number of network ties which is a perquisite of centrality empowers central actors by giving them access to valuable information on conditions, opportunities, and constraints. Actors on the periphery of the network, on the other hand, are relegated to less influential positions because they do not have access to the same quantities and quality of information.
Social network analysts have repeatedly demonstrated the association between network location and influence, as well as perceived influence. Laumann and Pappi (1976) and Galaskiewicz (1979), for example, both found that the more central the organization, the greater its "influence reputation" (i.e., perceived influence) in community affairs or a functional area. Many other researchers have also detected a positive association between centrality and influence or perceived influence (e.g., Boje & Whetten, 1981; Knoke, 1983; Laumann & Knoke with Kim, 1987; Marsden & Laumann, 1977; Miller, 1980; Perrucci & Lewis, 1989). Although some researchers suggest that the association between centrality and influence may depend on the nature of the network and the system under study and the way in which centrality is measured, they note that most studies have revealed at least some substantively meaningful association between the two (Mizruchi, 1994; Mizruchi & Galaskiewicz, 1994).
Network location and beliefs. Network location is associated not only with policy actors' influence, but also with their beliefs. Such cognition is related to social context in at least two ways (Pattison, 1994). First, relations in social networks serve as channels for the transmission of information. Therefore, an actor's location in a social network determines, at least in part, the specific information to which the actor is exposed, and hence the cognitions that use such information. Second, social locale is related to certain regular patterns of social interaction, which are a "driving force" behind knowledge acquisition (Carley, 1986).
In short, the above arguments suggest that, because network ties serve as channels of communication, the information which individuals receive and the knowledge they construct will depend on their location in the social network (Pattison, 1994). In support of this argument, Knoke (1990; 1994) observes that centrally-located actors differ from peripherally-located actors in both the quantity and quality of information to which they have access: the former can tap into larger stores of information through their dense web of connections to other actors. It follows that differences in network locale may lead to the formation of different perceptions and beliefs, due to variations in both availability of information and influence from their regular social contacts.
Another interpretation of the cognitive difference between network locations stems from research on innovation diffusion. While some diffusion studies show that early innovators are central actors, others indicate that early innovators are in fact marginal and "underconform to norms to such a degree that they are perceived as highly deviant" (Rogers, 1962, p. 197). To explain this phenomenon, Becker (1970) contends that central actors have a greater desire to protect their professional reputation, and therefore are less likely to adopt an innovation early on if the innovation is perceived as being risky or controversial. If the innovation is thought to be safe and uncontroversial, then central actors are more likely to lead its adoption.
Mizruchi and Fein (1999) provide a concrete example of how network centrality affects people's cognition. By studying the uses of coercive, mimetic, and normative isomorphism, Mizruchi and Fein found that organizational researchers gave disproportionate attention to one component of DiMaggio and Powell's (1983) notion of institutional isomorphism--mimetic isomorphism, which conforms to the prevalent discourse in the field. More importantly, Mizruchi and Fein found that centrally located organizational researchers are more likely than other scholars to invoke this dominant interpretation of DiMaggio and Powell's work. Given that a dominant discourse is likely to gain its legitimacy only when it is adopted by socially central actors (Strang & Soule, 1998), Mizruchi and Fein argue that researchers occupying central positions in disciplinary social networks are more likely to adopt conceptions that accord with the dominant discourse in the field, since they have more access, and hence more exposure, to the dominant discourse. This discourse is disseminated informally through social networks as much as it is through direct contact with the original work. Conversely, peripherally located researchers are less conversant with the dominant discourse, and are therefore more likely to hold non-dominant or more literal views of a work.
Given the relationship between network location and influence or perceived influence, and the relationship between network location and beliefs, we postulate that the beliefs of policy actors are also related to their perceived influence. In an earlier study using the same data set, we found that policymakers distinguished three significantly different levels of influence among the 67 interest groups in the national reading policy arena, while the interest groups themselves distinguished only two levels for those same 67 groups (Song, Miskel, Young, & McDaniel, 2000). Based on this finding and the foregoing conceptual rationale, Hypothesis 2 is that interest groups with different levels of perceived influence in the national reading policy arena have significantly different beliefs about phonics, whole language, and balanced methods of teaching reading.
We first compiled a preliminary list of interest groups and policy organizations with substantial interest in shaping national reading policy through systematic searches of the research literature, relevant archival documents, the Internet, and the Congressional Information Service, as well as through consultations with knowledgeable informants. At least one individual from each group or organization was included in our initial sample, with the most likely candidates being the organizations' chairs, or their directors of government relations.
To increase our coverage of the reading policy domain and ensure that important members of the reading policy community were included, the initial sample was supplemented by a snowball sampling technique (Heinz et al., 1993; Kingdon, 1995). That is, during the interviews, individuals were asked to identify other key individuals and groups who were involved in setting national reading policy. Additional policy actors were identified in this way and the initial sample was expanded. A total of 118 individuals were identified as being of sufficient importance in national reading policy to merit inclusion in the study. Eleven of these people declined for various reasons, leaving 107 to participate, with a response rate of 90.7%.
Of these 107 participants, 4 were individual policy actors not affiliated with any particular interest group or policymaking body at the national level, and were therefore excluded from the study. The final sample thus consisted of 103 out of 118 (87.3%) policy actors, with 74 of the participants representing 67 interest groups, another 31 representing 20 government agencies, and 2 representing both interest groups and government agencies. The 67 interest groups make up a broad spectrum, ranging from teacher unions and professional associations to think tanks and media organizations. Examples include the AFT, the NEA, the American Educational Research Association, the Heritage Foundation, the Christian Coalition, and Education Week. The 20 governmental agencies are mainly Senate and House committees and offices of the U.S. Department of Education.
We used structured interviews as our primary method of data collection. The interviews, averaging 30-45 minutes, were guided by two standard open-ended schedules: one for interest group members and the other for policymakers. Of the 107 participants, 55 were interviewed in person, 50 via telephone, and 2 responded to the interview questions via e-mail. All of the interviews (except the e-mail responses) were recorded on audio tape and transcribed. In order to guarantee confidentiality, the names of the interviewees and the organizations they represent were removed from the transcripts and replaced with randomly assigned numerical codes before the transcripts were entered into our database.
We employed thematic analysis techniques (Boyatzis, 1998) to capture the qualitative richness of the interview data. A scheme of thematic codes was developed which mapped onto the major concepts examined in this study. Following is a detailed explanation of how these categories were developed and the data coded.
Perceived influence. For the purpose of this study, we drew upon the findings from our previous study on the perceived influence of 67 interest groups in national reading policy arena (Song et al., 2000). By measuring the perceived influence of each interest group as the number of times the group was mentioned as being most influential/powerful/effective, minus the number of times the group was mentioned as being least influential/powerful/effective, Song and her colleagues found that the 67 groups could be categorized into three distinct influence levels based on policymakers' perceptions, and into two distinct influence levels based on interest groups' perceptions.
Beliefs about teaching reading. We created three themes for interest groups' beliefs about the most appropriate approach to teaching reading: Phonics, Whole Language, and Balanced Approach. These themes represent the three currently dominant schools of thought about how to teach reading. To further delineate people's attitudes towards phonics and whole language, four subthemes were created within each of these categories: "Not Included," "Neutral," "Included," and "Primary Basis." These subthemes reflected a group's relative degree of emphasis on phonics or whole language, and were assigned ascending values from 1 to 4. An average score was calculated for each respondent's focus on phonics and whole language themes. For example, if, in a transcript, no statements were coded as "Phonics--Not Included" or "Phonics--Neutral," two statements were coded as "Phonics--Included," and three statements were coded as "Phonics-Primary Basis," then this respondent's score for the phonics approach would be the average score of the five statements that he or she made regarding the phonics theme: [(2x3) + (3x4)]/5 = 3.6. This value then served as an index of the respondent's beliefs about phonics, and was also assumed to represent his or her group's beliefs on the subject. For groups with more than one interviewee, the individual scores on each theme were averaged to calculate the value for the group.
Based on codings of the "Phonics--Primary Basis" and "Whole Language--Primary Basis" subthemes, we created two dummy variables which indicated whether an organization supported the phonics-only or the whole language-only approach to reading instruction. A third dummy variable, labeled "Balanced," was also created. The data suggest that the respondents' views on balanced approaches of reading instruction were straightforward and clear-cut. Hence, no subthemes were created under the balanced approaches rubric.
Reliability of coding. To ensure an acceptable degree of coding reliability, we checked both intracoder and intercoder reliability using the formula recommended by Miles and Huberman (1994): Coder reliability = number of agreements/(total number of agreements + disagreements). The intercoder reliabilities for the themes Perceived Influence, Phonics, Whole Language, and Balanced Approaches turned out to be 96.6%, 80.0%, 85.0%, and 87.0%, respectively. The intracoder reliabilities were substantially higher, ranging from 95% to 100%.
To test Hypothesis 1, we performed paired-samples t-tests to compare support for the balanced, phonics, and whole language approaches among the 67 interest groups. To test Hypothesis 2, we conducted an analysis of variance (ANOVA) to examine whether interest groups in the three distinct influence categories developed on the basis of policymakers' perceptions differed significantly in their beliefs about the three methods of teaching reading. We also performed independent-samples t-test to examine whether groups in the two distinct influence categories that were based on interest groups' perceptions differed significantly in their beliefs about the three methods of teaching reading.
Table 1 provides the mean scores for policy actors' beliefs about the different methods of teaching reading. As shown in the table, the mean "Phonics--average" index for interest groups is 2.83, on a scale from 1 to 4, with 1 representing "Phonics--Not Included" and 4 "Phonics--Primary Basis." This index indicates that, as a whole, interest groups can be described as believing that phonics should be included in reading instruction but should not be the primary or exclusive basis of reading instruction. These groups believe that whole language should also be included in reading instruction--but, again, not in an exclusive manner.
Consistent with the implications of the mean "Phonics--average" index and the mean "Whole Language--average" index, the means of the "Phonics--only dummy" and "Whole Language--only dummy" indices indicate that very few groups believe that either phonics or whole language should be the exclusive basis for reading instruction. Specifically, only 7% of the groups stated that phonics should be the primary basis of reading instruction, and only 1% stated that whole language should be the primary basis, with the majority of groups (81%) supporting balanced approaches to teaching reading. The results were similar for the beliefs of policymakers.
To test Hypothesis 1, we performed paired-samples t-tests which examined whether support for balanced approaches to reading instruction was significantly stronger among the 67 interest groups than support for extreme views (i.e., phonics-only or whole language-only approaches). The results are listed in Table 2.
Apparently, interest groups' support for balanced approaches is significantly stronger than their support for either the phonics-only or whole language-only approaches (p < .001). Specifically, of the 67 groups, 73% more groups supported balanced approaches than phonics-only approaches, and 79% more groups supported balanced approaches than whole-language-only approaches. Similar results were also obtained for policymakers' beliefs. Hypothesis 1 thus holds.
Song et al. (2000) found that interest groups in the national reading policy arena fell into three distinct influence categories, based on policymakers' perceptions. Of our 67 groups, four were perceived as having a low level of influence, 50 had a moderate level of influence, and 13 had a high level of influence. Interest groups, however, only distinguished two distinct influence levels in the Song et al. study. In this two-level scheme, 45 of 67 groups were of low to moderate influence, while 22 groups were of high influence.
To test Hypothesis 2, we first conducted an ANOVA to examine whether interest groups in the low, moderate, and high influence categories based on policymakers' perceptions differed significantly in their beliefs about the three methods of teaching reading. The results are shown in Table 3.
These results suggest that interest groups with different levels of influence as perceived by policymakers did not seem to differ significantly in their beliefs about the three approaches to teaching reading (p > .10). Independent-sample t-test results indicated that interest groups in the low to moderate influence category (based on interest groups' perceptions) did not differ significantly from groups in the high influence category in terms of their beliefs on phonics, whole language, and balanced methods of reading instruction (p > .10) (see Table 4). Thus, Hypothesis 2 is rejected. Contrary to our prediction, our analyses suggest that interest groups with different levels of perceived influence hold similar beliefs about how to teach reading.
The findings indicate that, overall, interest groups believe that both phonics and whole language should be included in reading instruction, but do not think that either approach should be the exclusive basis. Our analyses also confirmed the hypothesis that support among interest groups for balanced approaches was significantly stronger than support for phonics-only or whole-language-only approaches; similar findings were also obtained for policymakers. The findings indicate that pedagogical extremists may no longer dominate the beliefs of the reading policy community. Although a minority of interest groups still possessed an either/or mentality, over 80% of the policy actors in reading stayed away from the two extreme camps and instead positioned themselves somewhere in the middle. The following response favoring a middle-of-the-road approach came from an interest group member, and is typical:
Good reading instruction teaches kids to read, drawing on some of the necessary word deciphering skills that you get from phonics and phonemes and drawing heavily on whole language in storybook reading. Neither works alone.
One explanation of these findings is that the reading wars are over, and the powerful forces of moderation and integration as exemplified by the National Research Council have forged widespread agreement. Given the long duration and intense polarization of the reading wars and the continued publication of aggressive tracts, however, the agreement on the need for a balanced approach may seem somewhat hollow. If there is a perception that the dominant discourse is favoring balanced approaches--which is likely, given that many prominent voices have been advocating balance--then fewer grants or consultant positions would go to individuals or groups whose beliefs on reading instructional approaches are seen as being outside the mainstream. It might be that in order to be perceived as socially and politically acceptable rather than deviant, some policy actors simply act like political chameleons, changing their colors to conform to the dominant discourse. Taking on the mantle of a balanced approach as a political necessity, policy actors, like chameleons, can secure the benefits associated with mainstream status, while still entertaining extreme views in private. From this perspective, as more political chameleons have assumed the color of balanced approaches, balanced approaches have been reinforced as the dominant discourse in the field, which has in turn encouraged still more chameleons to change their colors and join the mainstream. Another explanation is that nebulous or individual definitions of balanced approaches simply paper over the differences, while the basic conflicts remain as salient as ever. For example, Moats (2000) contends that balanced reading instruction is just whole language in disguise.
In short, our findings suggest that the turbulence created by the clash between phonics and whole language proponents in the policy community has subsided appreciably--at least on the surface--and that balanced approaches to the teaching of reading have emerged as the most acceptable solutions to the reading problem. Such developments actually fit very well with Kingdon's (1995) argument about the generation of policy ideas. Kingdon contends that the development of policy alternatives is guided more by the recombination of familiar elements into new proposals than by rational problem-solving. This perpetual recombination of existing ideas produces alternative solutions whose origins might be difficult to trace, but which may have broader appeal than the original ideas. In our case, the balanced approaches of reading instruction can be seen as growing out of the phonics and whole language perspectives, but the relatively new balanced approaches apparently enjoy much greater popularity among policy actors and possibly among the warring armies than did the phonics and whole language ideas from which they originated.
With regard to the relationship between perceived influence and beliefs about teaching reading, we predicted that interest groups with higher levels of perceived influence would occupy more central network locations, and therefore have different beliefs than groups with lower levels of perceived influence. This difference would be explained by differences in access to information. However, our findings suggest that this was not necessarily the case. We found that despite the differential levels of perceived influence, interest groups held similar beliefs on phonics, whole language, and balanced methods of teaching reading.
To interpret the homogeneity of beliefs across groups with different levels of perceived influence, we draw on Heclo's (1978) notion of issue networks. Heclo asserts that, given the growth of government bureaucracy and the interest group system, policymaking is no longer dominated by small circles or closed "iron triangles" of participants, but instead takes place within relatively open "issue networks." Issue networks are communication webs of people who are knowledgeable about some policy area; they frequently include government officials, legislators, business people, lobbyists, academics, and journalists (McFarland, 1992). One distinct feature of issue networks is that they are "shared-knowledge" groups: the members of a network are likely to share a common language, information base, and understanding of a particular public policy problem (Heclo, 1978, p. 103). Although Heclo cautions that shared knowledge does not necessarily produce agreement, it is reasonable to assume that shared knowledge and information are conducive to consensus-building.
In the reading policy domain, the 67 interest groups we identified are likely to be important members of the issue networks in reading. Conceivably, with the development of modern information technologies as e-mail, the Internet, and fax machines, such communications have become even more efficient, and the circulation of information and the spread of ideas within the issue networks have also become faster and more far-reaching than ever before. Hence, it is understandable that, despite the century-long reading war, the discourse which dominates the networks of reading policy actors is now about balanced approaches to teaching reading.
Actually, the notion of issue networks is not necessarily in conflict with arguments about the relationship between network location and beliefs (Knoke, 1990, 1994; Pattison, 1994). It might be the case that once the density of communication ties exceeds a certain threshold in a network, then even those members with relatively sparse ties (i.e., less central and less influential actors) would have adequate and unencumbered access to the flow of information. If the network is rather loosely connected, on the other hand, then location within the network might make a crucial difference in the availability of information, and consequently in policy actors' beliefs.
As reading has become a mainstay in the political agenda, this study--dealing with the perceived influence and beliefs of interest groups in the national reading policy arena--is both timely and significant. It represents an initial empirical effort to better understand interest groups--a central force in educational policymaking. Building upon the findings of the current study, future research in reading and other education policy arenas should further explore issues such as interest groups' beliefs on other contemporary reading issues (e.g., levels of reading achievement, standards, and assessment) and determine how certain beliefs or policy ideas diffuse through education policy networks. In short, the findings of this study have substantial potential to enrich our limited knowledge of the educational policymaking process in general, and as that process applies to reading in particular. Ultimately we hope that our findings will contribute to successful policy actions and effective policies designed to enhance the early reading achievement of the nation's children in the new millennium.
Boje, D. M., & Whetten, D. A. (1981). Effects of organizational strategies and constraints on centrality and attributions of influence in interorganizational networks. Administrative Science Quarterly, 26, 378-395.
Marsden, P. V., & Laumann, E. O. (1977). Collective action in a community elite: Exchange, influence processes, and issue resolution. In R. J. Liebert & A. Imershein (Eds.), Power, paradigms, and community research (pp. 199-250). London: Sage.
McFarland, A. S. (1992). Interest groups and the policymaking process: Sources of countervailing power in America. In M. P. Petracca (Ed.), The politics of interests (pp. 58-79). Boulder, CO: Westview Press.
Mitchell, R. K., Agle, B. R., & Wood, D. J. (1997). Toward a theory of stakeholder identification and salience: Defining the principle of who and what really counts. Academy of Management Review, 22(4), 853-896.
Mizruchi, M. S., & Fein, L. C. (1999). The social construction of organizational knowledge: A study of the uses of coercive, mimetic, and normative isomorphism. Administrative Science Quarterly, 44, 653-683.
Pattison, P. (1994). Social cognition in context: Some applications of social network analysis. In S. Wasserman & J. Galaskiewicz (Eds.), Advances in social network analysis (pp. 79-109). Thousand Oaks, CA: Sage.
Salisbury, R. H. (1991). Putting interests back into interest groups. In A. J. Cigler & B. A. Loomis (Eds.), Interest group politics (3rd ed., pp. 371-384). Washington, DC: Congressional Quarterly Press.
Sipple, J. W., Miskel, C. G., Matheney, T. M., & Kearney, C. P. (1997). The creation and development of an interest group: Life at the intersection of big business and education reform. Educational Administration Quarterly, 33(4), 440-473.
Song, M., Miskel, C. G., Young, T. V., & McDaniel, J. E. (2000, November). Perceived influence of interest groups in national reading policy. Paper presented at the annual meeting of the University Council for Educational Administration, Albuquerque, NM.
Thomas, C. S., & Hrebenar, R. J. (1991). Nationalization of interest groups and lobbying in the states. In A. J. Cigler & B. A. Loomis (Eds.), Interest group politics (3rd ed.). Washington, DC: Congressional Quarterly Press.
Valencia, S. W., & Wixson, K. K. (2000). Policy-oriented research on literacy standards and assessment. In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research (vol. III, pp. 909-936). Mahwah, NJ: Erlbaum.