PA Assg#7–Please see attached instructions. Thank you.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms and Oxford University Press are collaborating with JSTOR to digitize, preserve and extend access to Journal of Public Administration Research and Theory: J-PART This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms J PART 23:103 Institutional Logic and Street-Level Discretion: The Case of HIV Test Counseling Eve E. Carrow*, Oscar Cruskyf *University of Michigan; ' University of California-Los Angeles ABSTRACT Research on variation in policy implementation has examined how individual- and organizational-level variables shape discretionary practices at the frontlines of implementa tion but has given less attention to the influence of the organizational field. From an institutional logics perspective, the practices of street-level workers are likely to be more congruent with policy intent when policy is consistent with the core institutional logic that structures the organizational field. Using data from a probability sample of 90 Los Angeles County HIV/AIDS health organizations and 216 frontline practitioners, our findings suggest that the extent to which implementation of HIV test counseling follows guidelines of the Centers for Disease Control and Prevention (CDCyf L V U H O D W H G W R F R Q V L V W H Q F \ E H W Z H H Q & ' C policy and core institutional logic, when accounting for individual-level factors such as client need; practitioners' professional orientation, knowledge, experience, and training; and organizational constraints such as workload pressures and formalization. It is well established that street-level workers in health and human service organizations, in the provision of services and other activities, often diverge from policy objectives and organizational expectations (e.g., Brodkin 2008; Fixsen et al. 2005; Handler and Hollingsworth 1971; Lipsky 1980; McDonald and Piliavin 1984; Mullen and Bacon 2004; Panzano and Herman 2005yf $ N H \ L Q V L J K W R I W K H U H V H D U F K R Q S R O L F \ L P S O H P H Q W D W L R n is that workers may diverge from policy intent because of the discretion afforded to them in street-level bureaucracies (Lipsky 1980yf 7 K H \ E H F R P H L Q H I I H F W O R Z H U O H Y H O S R O L F \ P D N H U s since their behaviors constitute policy in practice. Most of the research on street-level bureaucrats has examined how diverse occupa tions such as police officers or case workers use their discretion to respond to shared con ditions of work, such as resource limitations and client demands, in predictable ways (see, e.g., Brodkin 1997; Lin 2000; Lipsky 1980; Maynard-Moody and Musheno 2003yf . Although this microcentered theoretical approach has furthered our understanding, it has neglected to consider how and why street-level workers may differ systematically across organizational and institutional contexts in their discretionary behavior. This article contributes to the research on street-level behavior by demonstrating that frontline workers differ from one another in administering a standard protocol and that the We thank Yeheskel Hasenfeld for his insightful comments. We are also grateful to the anonymous reviewers for their helpful comments and suggestions. Address correspondence to the author at [email protected]. doi: 10.1093/jopart/mus009 Advance Access publication April 18, 2012 The Author 2012. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: [email protected] This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 104 Journal of Public Administration Research and Theory differences are related to the underlying institutional logic, or set of organizing principles (Friedland and Alford 1991, 248yf R I W K H R U J D Q L ] D W L R Q D O I L H O G L Q Z K L F K W K H Z R U N L V H P E H d ded. It adds to recent comparative work that has examined the influence of broad contextual factors such as county-level political culture (Fording, Soss, and Schram 2007yf D Q G R U J a nizational ideology and goals (Hasenfeld and Weaver 1996yf R Q V W U H H W O H Y H O G L V F U H W L R Q 7 o our knowledge, ours is the first study on street-level discretion to examine the influence of the organizational field. In a second contribution, our model provides a more nuanced and conditional under standing of worker agency than is found in most research on street-level bureaucracy. We contend that because street-level research has not sufficiently accounted for broader envi ronmental influences on discretion, it has overstated the agency of workers, who are viewed as exploiting the discretion afforded to them to maximize their self-interest. Discretion has been identified with workers' strategic reinterpretation of policy to accommodate the demands of work, limited resources, and their own beliefs and needs (e.g., Jewell and Glaser 2006; Prottas 1979; Smith and Donovan 2003yf $ Q R W K H U V W U D Q G R I U H V H D U F K D s sesses the influence of potential administrative and managerial controls over worker agency, including program design (Hill 2006yf W K H H [ W H Q W W R Z K L F K D G P L Q L V W U D W L R Q H P S K a sizes policy goals (Ewalt and Jennings 2004; Hill 2006; Riccucci et al. 2004yf D Q G P D n agerial supervision (Brehm and Gates 1997; Brewer 2005; Riccucci 2005yf 2 Y H U D O O W K e research indicates that these potential organizational controls have only modest influences on discretionary practices (Meyers and Vorsanger 2003yf . Our model delimits the agency of workers by applying one of the key insights from institutional theory—that agency is institutionally embedded—to the idea of street-level discretion (Friedland and Alford 1991; Thornton and Ocasio 2008yf 7 K H D Q D O \ V L V G H S L F W s street-level workers not as agents with objective interests and preferences, but rather as institutionally constructed actors whose values, interests, and practices are partially deter mined by the institutional logics that structure the organizational fields in which they op erate. Thus, although workers may use their discretion to advance their interests in response to their shared conditions of work, their actions, intentions, and interests are themselves institutionally conditioned in systematic ways. To explore this idea, we ask: Is the degree of compliance with policy mandates associated with the core institutional logic that structures the focal organizational field? To examine variation in policy implementation across organizational fields, we inves tigate the behaviors of practitioners who implement publicly funded human immunodefi ciency virus (HIVyf S U H W H V W F R X Q V H O L Q J D F U L W L F D O W R R O L Q + , 9 S U H Y H Q W L R Q S R O L F \ D Q G S U D F W L F e that is widely promoted by the Centers for Disease Control and Prevention (CDCyf $ V R X t lined in the CDC's revised guidelines (CDC 2001yf + , 9 S U H W H V W F R X Q V H O L Q J F D Q U H G X F H W K e spread of HIV by conveying information regarding HIV transmission and prevention and the meaning of HIV test results and by providing "client-centered" prevention counseling that ( 1 yf K H O S V F O L H Q W V L G H Q W L I \ W K H E H K D Y L R U V S X W W L Q J W K H P D W U L V N I R U D F T X L U L Q J R U W U D Q V P L W W L Q g HIV and (2yf K H O S V W K H P F R P P L W W R V W H S V W R U H G X F H W K L V U L V N 3 U H W H V W F R X Q V H O L Q J L Q F O X G H s a personalized risk assessment that: 1 These recommendations were revised again in 2006. We focus on the CDC guidelines for HIV testing and counseling that were released in 2001, which were in place during the data collection period of the study. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Carrow and Crusky Institutional Logic and Street-Level Discretion 105 ... encourages clients to identify, understand, and acknowledge the behaviors and circumstances that put them at increased risk for acquiring HIV. The session explores previous attempts to reduce risk and identifies successes and challenges in these efforts. This in-depth exploration of risk allows the counselor to help the client consider ways to reduce personal risk and commit to a single, explicit step to do so (CDC 2001yf . HIV pretest counseling provides an ideal opportunity for studying the relationship between field-level factors and street-level discretion. First, HIV pretest counseling is con ducted in a variety of organizational fields that vary substantially in terms of the compre hensiveness of counseling as defined by CDC guidelines. For example, in medicine, studies indicate that frontline practitioners in settings such as hospitals and emergency departments are often unlikely to provide HIV testing and counseling in a comprehensive manner (Cohan et al. 2009; Fincher-Mergi et al. 2002yf G H V S L W H W K H L Q F U H D V L Q J U R O H R I H P H U J H Q F y departments as the de facto health care safety net for poor and socially marginalized groups at high risk for HIV (Glauser 2001 ; Gordon et al. 2001 ; Richardson and Hwang 2001 ; Stern, Weissman, and Epstein 1991; Taylor 2001yf 6 H F R Q G D O W K R X J K W K H & ' & S U R Y L G H V I L Q D Q F L D l and technical support to organizations that conduct HIV testing and counseling, its enforce ment of CDC guidelines is weak. Thus, by studying the implementation of HIV pretest counseling, we can observe discretionary behavior that is relatively unadulterated by policy-level controls. In the case of HIV testing and counseling, implementation is a crucial public health issue, as between 252,000 and 312,000 individuals are currently unaware of their HIV infection and it is estimated that they account for 54yb R I D O O Q H Z L Q I H F W L R Q V L Q W K e United States (Marks, Crepaz, and Janssen 2006yf . The study is organized as follows. First, we bring together ideas from organiza tional theory to develop expectations on how field-level institutional logics shape pol icy implementation by street-level workers. Our central hypothesis is that street-level workers are more likely to implement policy mandates in a manner that is congruent with policy intent when the aims and assumptions of the policy are consistent with the core institutional logic of the organizational field. Next, using a probability sample of organizations that conduct HIV counseling and testing in Los Angeles County (LACyf , we identify four distinct core institutional logics (medical, public health, social move ment, and multiserviceyf D Q G D V V H V V W K H F R Q V L V W H Q F \ R I H D F K Z L W K W K H D L P V D Q G D V V X P p tions of HIV test counseling. In support of our expectations, we find that when counseling is consistent with the core field-level institutional logic, it is more likely that implementation will reflect policy intent, even when accounting for selected in dividual and organizational-level factors thought to influence street-level discretion. We conclude by discussing the implications of the findings. INSTITUTIONAL LOGICS AND EMBEDDED AGENCY From an institutional logics perspective, discretionary practices at the frontlines of policy implementation are embedded in a broad meaning system, reflected by the dominant field level logic, that defines the "interests, identities, values, and assumptions of individuals and organizations ..." (Thornton and Ocasio 2008, 103yf : R U N H U V D U H H [ S R V H G W R I L H O G O H Y H l institutional logics through their participation in and knowledge of the organizational field and most directly through their membership in organizations in which the logics are em bedded. Hence, decisions, courses of action, and frontline outcomes of practice result from This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 106 Journal of Public Administration Research and Theory embedded agency or agency that is constrained and enabled by the institutional logics structuring the organizational field and the organizations that constitute it (Jackall 1988; Friedland and Alford 1991; Thornton and Ocasio 1999yf . Embedded agency implies the interplay of agency and institutional logics and calls into question the assumption of individualistic interests that underlies much of the research on street-level discretion. On the one hand, workers use their discretion to maneuver and optimize their interests. In response to limited resources, conditions of work, organi zational context, and their interests and judgments, workers modify policy though their practice (Lipsky 1980yf ) R U H [ D P S O H W K H \ P D \ X V H W K H L U G L V F U H W L R Q W R G L V H Q W L W O H F O L H Q W s (e.g., Maynard-Moody and Musheno 2003; Smith and Donovan 2003yf D F W D V F L W L ] H Q D J H Q W s rather than state agents (Keiser 1999yf R U H Q I R U F H U X O H V D Q G K L H U D U F K \ 0 D \ Q D U G 0 R R G \ D Q d Musheno 2003yf 2 Q W K H R W K H U K D Q G E H F D X V H W K H L U L Q W H U H V W V D Q G D Y D L O D E O H U H S H U W R L U H R f practices are constituted by dominant field-level logics, street-level workers' behaviors are shaped by institutional structure in predictable ways. How do institutional logics delimit the discretionary practices of street-level workers? As noted by Thornton and Ocasio (2008, 114yf L Q V W L W X W L R Q D O O R J L F V D I I H F W W K H D O O R F D W L R Q R f attention to alternative schémas for perceiving, interpreting, evaluating, and responding to environmental situations ... [by providing] a set of values that order the legitimacy, im portance, and relevance of issues and solutions ... [and provide] decision makers with an understanding of their interests and identities." This insight is consistent with research showing that the implementation of human service technology is shaped by values and understandings of local users who share social context (McLaughlin et al. 1999yf 9 D O X e orientations, policy understanding, knowledge, and attitudes shape implementation across a range of policy areas including substance abuse treatment (Knudsen, Ducharme, and Roman 2007yf Z H O I D U H U H I R U P 0 H \ H U V * O D V H U D Q G ' R Q D O G \f, jobs programs (e.g., Sandfort 2000, 2010yf 6 R F L D O 6 H F X U L W \ ' L V D E L O L W \ S U R J U D P V . H L V H U \f, and early child hood education (Sandfort 2010yf 6 D Q G I R U W \f shows how collective schema held by frontline practitioners inhibited change. In a study of a welfare department, Watkins-Hayes (2009yf V K R Z V K R Z U D F L D O D Q G J H Q G H U L G H Q W L W L H V L Q W H U V H F W Z L W K R U J D Q L ] D W L R Q D O F X O W X U H D Q d rules to shape worker practices. Research by Meyers, Glaser, and Mac Donald (1998yf , Lin (2000yf D Q G . Q X G V H Q ' X F K D U P H D Q G 5 R P D Q \f demonstrate that workers reject practices that conflict with their understanding of the organization's core purpose and ac tivities. For example, Meyers, Glaser, and Mac Donald (1998yf V K R Z H G W K D W G H V S L W H W K H L r endorsement of welfare reform, welfare workers failed to implement new changes in wel fare policy that did not correspond with their understanding of the mission of their program. Because policy implementation is rarely examined across organizational fields, however, the role of field-level institutional logics as a determinant of organizational-level meanings and workers' understandings, values, interests, and practices is understudied. We posit that street-level practitioners are predisposed to accept core institutional log ics that structure the organizational field in which their organization is located and are likely to implement policy goals that accord with these logics. When policy features clash with the core logic, street-level workers are apt to use their discretion to disregard the policy or resist implementation. Our central hypothesis is as follows: the greater the consistency between the aims and assumptions of HIV test counseling and the core logic of the organizational field, the greater the comprehensiveness of counseling. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Crusky Institutional Logic and Street-Level Discretion 107 Components of Field-Level Institutional Logics Viewing discretion as institutionally delimited shifts the focus of inquiry from individual or organizational contexts to the organization's environment. Institutional logics have been described in a variety of ways and can occur at societal, industry, organizational field, or organizational levels. Field-level institutional logics are important empirical and theoret ical constructs because they provide members of the organizational field—defined as "those organizations that, in the aggregate, constitute a recognized area of institutional life" (DiMaggio and Powell 1991, 65yf Z L W K D V H Q V H R I F R P P R Q S X U S R V H D Q G X Q L W \ W K D t helps explain their connections and guide their interactions. Organizational fields are thought to coalesce around dominant institutional logics, although multiple logics may co exist (e.g., Scott 2008; Thornton and Ocasio 1999yf . Following Hasenfeld (1983yf D Q G 0 K U D Q G * X H U U D 3 H D U V R Q \f, we propose that the field-level institutional logics of health and human service organizations most likely to guide the practices of street-level workers are those that broadly define the purpose of the organizational field. They are delineated along the following dimensions: (ayf W D U J H W S R p ulation served, (byf V R F L D O S U R E O H P V D G G U H V V H G F \f diagnoses of problems, (dyf G H V L U H G R X t come of the organizational intervention, and (eyf D S S U R S U L D W H P H D Q V R I D G G U H V V L Q J W K e problems. Organizations share sets of practical understandings derived from assemblages of these elements and claim membership in the field by embedding them in their practices and structures (Baum and Oliver 1992; DiMaggio and Powell 1983; Meyer and Rowan 1977; Scott 2008yf 7 K H F R U H O R J L F F R Q V W L W X W H G E \ W K H F R P E L Q D W L R Q R I W K H V H H O H P H Q W V U H I O H F W s a theory of social provision (Hasenfeld 2000; Schneider and Ingram 1993yf W K D W K H O S V S U o vide a rationale for the organization's service delivery. The core field-level logic is ex pressed in the organization's mission, which is intentionally crafted to signal membership within a particular organizational field. That is, the mission both reflects and defines the field in which the organization is located. INDIVIDUAL- AND ORGANIZATIONAL-LEVEL FACTORS The research on street-level bureaucracy points to several individual-level factors that may explain frontline variation in policy implementation. These factors need to be considered when exploring the influence of field-level institutional logics. Abbot (1992yf S R L Q W V W R S U o fessional orientation as a potential factor in policy implementation. Professions make ju risdictional claims over particular tasks by ( 1 yf G L D J Q R V L Q J F O D V V L I \ L Q J F O L H Q W V E D V H G R Q W K H L r perceived issuesyf \f inferring (reasoning about client issuesyf D Q G \f treatment (taking action on these issuesyf 3 U D F W L W L R Q H U V W K X V U H I O H F W G L V W L Q F W L Y H F D S D E L O L W L H V W K D W D G G U H V V F O L H Q t issues, needs, and attributes (Hasenfeld 2010yf D Q G F R O O H F W L Y H L G H Q W L W L H V E D V H G R Q W K H V H F R m petencies (occupations and professionsyf J R Y H U Q W K H L U S H U F H S W L R Q V D Q G D F W L R Q V 2 F D V L o 1995yf ) R U H [ D P S O H P H P E H U V R I W K H P H G L F D O S U R I H V V L R Q V D U H V R P H W L P H V Y L H Z H G D V D F W L Q g in a manner incompatible with the logic of HIV counseling by relating to patients in narrow biomedical terms and emphasizing purely medical, rather than biopsychosocial factors (Friedson 1970yf & R Q V H T X H Q W O \ Z H H [ S H F W H G W K D W Q R Q P H G L F D O F R X Q V H O R U V Z R X O G E H P R U e likely to provide comprehensive HIV counseling than physicians and nurses. Work experience promotes adherence to established practices because workers be come more entrenched in ways of doing things and more resistant to change (Riccucci et al. 2004yf Z K H U H D V W U D L Q L Q J D Q G N Q R Z O H G J H I D F L O L W D W H D G D S W D W L R Q W R Q H Z S R O L F L H V - H Z H O l This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 108 Journal of Public Administration Research and Theory and Glaser 2006; Meyers, Riccucci, and Lurie 2001; Riccucci et al. 2004yf 7 K H U H I R U H Z e expected that greater experience, training, and knowledge in HIV counseling would result in greater comprehensiveness of counseling. Consistent with a citizen agent perspective on street-level discretion, workers some times use their discretion to tailor benefits to the level of client need. Keiser (1999yf G H m onstrated that states with greater shares of disabled people also had the highest disability benefit approval rates. We include two measures of client need. We expected that persons who test HIV positive would receive more comprehensive counseling than those who test HIV negative because the former are more likely to present more risk factors at pretest counseling. In addition, we expected that clients identified by the practitioner as belonging to well-established behavioral risk categories (men who have sex with men, intravenous drug users, and women at sexual risk—that is, women who have recently engaged in risky sexual behavior such as unprotected sex with multiple partners or trading sex for moneyyf would receive counseling that is more comprehensive than those without these risk factors. We also considered organizational-level factors that shape street-level discretion. First, to cope with workload pressures and time constraints, practitioners may prioritize among tasks and scale back services (e.g., Jewell and Glaser 2006; Prottas 1979; Smith and Donovan 2003yf ) R U H [ D P S O H - H Z H O O D Q G * O D V H U \f showed how caseworkers in welfare offices managed time constraints and high caseloads by falling back on case processing at the expense of a new mandate to identify and refer appropriate clients to social services. Consequently, we expected that practitioners in organizations with high workloads would be more likely than others to score low on comprehensiveness of pretest counseling. Second, formalization (automation or reliance on formal rulesyf F R Q V W U D L Q V G L V F U H W L R n by reducing the options available to frontline workers (Sandfort 2000yf : H Q J H U D Q d Wilkins (2008yf I R X Q G W K D W Z K H Q F R P S D U H G W R L Q S H U V R Q X Q H P S O R \ P H Q W F O D L P V W K H L Q W U o duction of automated telephone claims decreased the ability of workers to disentitle clients they deemed undeserving. We posit that automation and reliance on formal rules reduces the opportunity for discretionary practice, thus increasing the likelihood that HIV coun seling is implemented in a comprehensive manner. METHODS Sample and Data Data are from a representative sample of human service organizations in LAC that provide HTV counseling and testing. The sample frame was stratified by organizational type: for-profit (n = 33yf Q R Q S U R I L W Q \f, and public (n = 5yf K R V S L W D O V I R U S U R I L W S U L Y D W H S U D F W L F H s (n = 37yf Q R Q S U R I L W Q R Q K R V S L W D O V L Q F O X G L Q J F R P P X Q L W \ E D V H G V R F L D O V H U Y L F H D J H Q F L H s and primary health clinics (n = 125yf Q R Q K R V S L W D O & R X Q W \ * R Y H U Q P H Q W Z K L F K L Q F O X G H d public health STD clinics and personal health centers (n = 32yf D Q G S U L Y D W H I R U S U R I L t and County Government organizations with mobile testing sites (« = 14yf I R U D W R W D O R f 294 organizations. Hospital settings were derived from the list of State-licensed hospitals in California (as of 12/31/2002yf D V S U R Y L G H G E \ W K H & D O L I R U Q L D 2 I I L F H R I 6 W D W H Z L G H + H D O W K 3 O D Q Q L Q J D Q d Development. Hospitals that did not provide HIV testing were excluded from the sample frame. Lists of nonhospital settings were constructed from four publicly available lists: the This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion 109 Aids Project Los Angeles HIV Testing list, the Los Angeles County Department of Health Services HIV Testing Sites list, the National HIV Testing Resource's testing sites list for LAC, and the HIV L.A. testing sites list. A list of licensed primary care clinics in LAC that provided HIV testing was assembled using the American Academy of HIV Medicine's online roster of credentialed HIV specialists and other health care providers committed to HIV/AIDS care. Three key expert informants in the HIV care community reviewed the listing so that omitted practitioners could be added to the sampling frame. Public health testing sites were identified by contacting the three public health departments in LAC (Los Angeles Department of Health Services, Pasadena Public Health Department, and the Long Beach Department of Health and Human Servicesyf . In light of the small number of organizations in some strata, all organizations within each stratum were included, with the exception of the nonprofit nonhospital stratum, which was sampled using the probability-to-estimated size design, where the size for each orga nization was the estimated number of HIV tests conducted in a given year based on most recent data available. Using this method, large HIV testing sites were more likely than small ones to be sampled. Organizations were stratified into three groups: large, medium, and small volume of HIV testing, for a total of 47 nonprofit nonhospital organizations and a sam ple of 216. Weights were used to correct for stratification of the sample. Sixty-five organ izations declined to participate in the survey for a sample of 151 organizations and an acceptable response rate of 70yb V H H W D E O H $ O \f. In 2002—2006, 621 practitioners and administrators were randomly selected from these organizations for one-hour computer-assisted face-to-face interviews. Up to six front line practitioners of HIV counseling and testing services, including physicians, nurses, and counselors, along with up to three managers, including executives, clinic managers, and testing supervisors, were randomly selected for recruitment into the study. Organizational level data were collected through interviews with managers. Interviews with 428 frontline practitioners included questions on respondents' background characteristics, education, training, HIV-testing expertise, and work roles. Frontline service practitioners were asked to describe in detail their most recent HIV-negative and most recent HIV-positive coun seling session conducted during the previous 6 months (other than occupationally related incidents or "needlesticks"yf : H H [ F O X G H G I U R P W K H D Q D O \ V L V I R U S U R I L W S U L Y D W H S U D F W L F H s and the 40 practitioners among them who had conducted at least one HIV test in the pre vious 6 months.2 We also excluded 133 practitioners who reported they had not conducted an HIV test counseling session during the previous 6 months. The resulting sample con sisted of 255 frontline practitioners in 108 organizations. Listwise deletion of observations with incomplete data on all variables in the analysis further reduced the sample by ~ 15yb W o 216 frontline staff in 90 organizations.3 2 Physicians at for-profit private practices represented a unique case and their inclusion would have created misleading results. Private practices were located in a gay community, and these frontline practitioners generally offered HIV treatment as part of a larger array of medical services. They did not report comprehensive HIV counseling in their previous session with an HIV-negative and/or positive patient, not because the institutional logic of the physician practices was inconsistent with the logic of HIV counseling but because they provided counseling on an ongoing basis with their patients, with whom they generally shared long-term relationships. 3 The probability of missing data on each predictor variable did not depend on the value of the outcome, indicating that listwise deletion did not result in biased regression estimates. Following Allison (2002yf Z H F R Q F O X G H G W K D W W K H U e was no need to model the missing data mechanism as part of the estimation process and chose listwise deletion as the preferred method for handling missing data. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 110 Journal of Public Administration Research and Theory Variables Dependent Variable: Comprehensiveness of HIV Pretest Counseling The 2001 CDC guidelines recommended that HIV pretest counseling impart information on HIV transmission and prevention, the testing process, and the meaning of HIV test results and that practitioners engage clients in risk reduction counseling. Covering each domain during the counseling session is viewed as crucial because together the domains may have a greater likelihood of reducing HIV transmission than separately. For example, knowing how HIV is transmitted is designed to help clients make sense of safe sex methods and other risk reduction behaviors; understanding the benefits of knowing their serostatus may en courage clients to get regular HIV tests; and understanding the testing process and need for a window period can help ensure that clients test properly and understand the results of the test. To ascertain the extent to which respondents covered these domains in a comprehen sive manner, interviewers inquired about the following items and respondents indicated whether or not they had discussed each topic with the client in their most recent HIV pretest counseling session: (1yf P R G H V R I + , 9 W U D Q V P L V V L R Q \f safe sex practices, (3yf D Q R Q \ P R X s versus confidential testing, (4yf Z K D W W K H + , 9 D Q W L E R G \ P H D V X U H V \f explanation of the testing process, (6yf P H D Q L Q J R I W K H W H V W U H V X O W \f accuracy of the test result, (8yf F R Q I L G H n tiality of the test result, (9yf W K H Z L Q G R Z S H U L R G R U G D W H R I D Q \ Q H H G H G I R O O R Z X S W H V W \f the benefits of knowing one's serostatus, and (11yf U L V N U H G X F W L R Q D Q G J R D O V H W W L Q J 5 H V S R Q G H Q W s were permitted to consult their patients' charts to refresh their memories on the counseling session. The item was scored 1 if the respondent indicated that the topic was covered during the session and 0 otherwise. The items were summed for an aggregate measure ranging from 0 to 11. The more items covered, the greater the comprehensiveness of counseling. Four Core Institutional Logics According to an institutional logics approach, HIV test counseling is likely to be imple mented in accordance with CDC guidelines when the CDC's policy is consistent with the core institutional logic that structures the organizational field. To determine the degree of congruence, we extract four distinct institutional logics from the data and compare the aims and assumptions of HIV pretest counseling policy with each. Our model is consistent with Möhr and Guerra-Pearson's (2010, 14yf L Q V L J K W W K D W L W L s the essential duality of categories [of clients] and practices that both defines the institu tional logic ... and creates the tension through which the meaningful character of orga nizational activity becomes constituted." Thus, the application of a service technology to a specific target client group represents a "solution" for a category of people that presup poses specific client attributes and a particular problem, diagnosis, and desired outcome. Differences in target groups (and their assumed attributes, problems, and desired outcomesyf are reflected in the variety of services legitimately applied to them. At the same time, rep ertoires of practice are constituted as appropriate and legitimate when applied to categories of people with particular attributes, problems and desired outcomes. As noted by DiMaggio and Powell (1991, 148yf I L H O G V R Q O \ H [ L V W W R W K H H [ W H Q W W o which they are institutionally defined," a condition that depends on "a mutual awareness among participants in a set of organizations that they are involved in a common enterprise. " Institutional logics derived from the combination of service technology and target client group thus structure organizational fields and differentiate between them. At the same time, organizations claim membership in fields by virtue of the similarity of their combinatory This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion 111 logics as reflected in their missions. In keeping with this reasoning, we extracted field-level institutional logics from meaningful combinations of target client groups served and serv ices provided and assigned organizations to the same field when they exhibited similarity in their combination of these dimensions. Here, we follow the methodological advancements of scholars who have used organizational-level characteristics to measure the meaning structures of organizational fields or niches to which they belong and, sometimes, to de termine the boundaries of these higher level structures (e.g., DiMaggio and Mullen 2000; Möhr and Duquenne 1997; Möhr and Guerra-Pearson 1998; Tilly 1997yf 4 First, we coded the mission statements of the sample organizations by the services provided and the target client groups served. Mission statements, which are intentionally crafted to make jurisdictional claims upon particular regions of institutional space, often contain the dimensions of the field's core institutional logic. When mission statements did not provide information on a dimension, this information was extracted from the organ ization's website. Variables were coded 1 if the organization provided a service or served a particular target client group and 0 otherwise. Second, we used K-means cluster analysis to partition the organizations into nonover lapping groups based on similarity of their service and target client group profiles. Due to the small sample size, we were unable to use an inferential statistical strategy such as latent class analysis. The K-means algorithm creates the best solution for K clusters in the fol lowing way. It assigns observations to clusters using a random method and then calculates the means of each of the clusters. If the mean of an observation in a particular cluster is closer to the mean of another cluster, the observation is reassigned to that cluster. The process is repeated until no observation can be reassigned. Third, we ascertained the latent variable (core institutional logicyf W K D W G H I L Q H G H D F h cluster by examining its service profile and target client groups (see tables A2 and A3yf . We then assessed the congruence of each logic with the aims and assumptions of HIV pretest counseling. As shown in table 1, the results indicated that organizations providing HIV testing and counseling in LAC operate in four organizational fields typified by distinct institutional logics that vary in their congruence with the aims and assumptions of HIV pretest coun seling. First, a medical logic structures the field of medical organizations such as hospitals and emergency departments. This logic involves the provision of medical interventions for the general public. It tends to attribute illness or disease to pathology or biochemistry, con struct health as the desired outcome, and identify medical interventions as the most appro priate solution. The logic constructs service recipients in biomedical terms, puts greater emphasis on medical than social and psychological indices of pathology, and values medical treatment and the restoration to health more than behavioral change leading to the prevention of disease transmission. Frontline practitioners operating in the context of a biomedical logic are therefore expected to be fundamentally oriented, institutionally and professionally, toward goals other than the implementation of preventive and behavioral interventions such as HIV counseling. Second, a public health logic typifies a field consisting mostly of public health clinics. This logic links education, prevention (e.g., screening and immunizationsyf D Q G P H G L F D l For a review of methodological advancements in measuring higher level meaning structures from lower level data, Möhr (1998yf . This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Journal of Public Administration Research and Theory Table 1 Consistency of HIV Test Counseling with Institutional Logics I T3 § T3 3 'S O - -a I § « "2 lS9 >2 Cd u .2 to GO '§|«H .£ J ° W £ M o o •H M «J X> Oh —■« : *q w c o s 0yf > o s 'Sc^ cd Q CO *"2 Qh < G s £ o x 00 Q < > CO G O CO "O CO a o +-» CO G $ s cd J: o '5b o 'o 43 o fc> 5? »rt W Q g U g J £ D "T2 flj > .2 « ^ O _ » U o, O « &, o *-*< co o •2 « £ e "S> a .a, 8 .S ■§ _ _7 ° ts 2 .3 .2 S 5 o — 3 3 .g 3 ^ M O O > T3 05 a CO CO o .s Wyf A . C 2 2 T3 43 .2 J2 *-' o 8 -S <2 si« •8 2 5 o. § £ »-H cd CO CO O O cd O cd 43 44 (D 43 s D CQ jd O £ o > a> c o _ "-M a S -a 5 "3 •3 2 w S •Si T3 s 'S Oh G o h-J CO o oo \T3 C n.— B « C vi yb ® < u •o i & ac Cfl ^ S O < W3 ■c (n Q .. 5 M ^ 153 2 b O <2 00 Q < > s a> & .. u m Q g.< s Mh \ fll ts > c ti H U £ EC S3 CO *1 ^ 2^ s -a ° ■g 'c -3 ffl ts J2 c o (U 1 ^ Oh H O J2 < 'co O CO « ^ 'G & E c2 a 1 « g O £3 eo c 81 § 1 11 C/5 S> CO K* £ s ^ S -a *c « 2 3 2 J £ Oh £ Cd C o G Oh O Oh Xyf O I cd O CO o cd S -o c E o c U 'a o This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion 113 services to at-risk low-income groups and the general public. It emphasizes public health through the prevention of HIV and other infectious diseases; links transmission to behavior, awareness, knowledge, and poverty; and identifies educational, preventive, and medical interventions as the appropriate solution. In contrast to the medical logic, the preven tion-oriented logic directs workers to pursue goals such as HIV test counseling because it specifically targets sexually transmitted infectious diseases such as HIV, constructs ser vice recipients in bio-psycho-social terms, centers on social and psychological roots of infectious disease transmission, and puts a premium on prevention rather than treatment— meanings that are more consistent with the aims and assumptions of HIV counseling. Third, a social movement logic typifies a field consisting mostly of small nonprofit HIV and AIDS advocacy organizations. It links advocacy, prevention, education, and psy chosocial services to groups at risk for HIV, including lesbian, gay, bisexual, and trans gendered individuals, people who are homeless, and ethnic minorities. This logic emphasizes prevention of HIV, and programs are designed to further this purpose. Many organizations in this field were founded by HIV-positive individuals who died of AIDS and who served as an inspiration for organizational action aimed at client empowerment and social change. The logic links transmission to individual risk behaviors, level of awareness, and macrosocial conditions such as poverty and unjust policies; identifies advocacy and preventive, educational, and psycho-social strategies as the appropriate solution; and con structs behavior change, lowered HIV transmission rates, social and psychological health, and social justice as desired outcomes. The social movement logic is highly congruent with the goals of HIV pretest counseling. Finally, a multi-service logic structures a field of nonprofit health organizations that provide both social and medical services to low-income and uninsured clients. This logic attributes medical, social, and psychological problems to pathology, biochemistry, and so cial or psychological conditions, identifies medical, psychological, and social interventions as appropriate solutions, and identifies physical, social, and psychological well-being as desired outcomes. Unlike CDC's policy, the logic does not specifically aim to reduce HIV acquisition and transmission; however, consistent with the aims of HIV pretest coun seling, it attempts to effect behavior change and heightened awareness through counseling that addresses the psychosocial aspects of clients. As shown in table 1, the medical logic is least consistent, the public health and social movement logics are most consistent, and the multi-service logic is moderately consistent with the aims and assumptions of HIV test counseling. Accordingly, we expected that all else equal practitioners operating within the context of the medical logic will provide the least comprehensive counseling, those practicing within the context of public health and social movement logics the most comprehensive counseling, and those operating within the context of the multiservice logic will provide counseling somewhere in between. Table 2 provides key descriptive statistics for organizations in each institutional logic category. Most organizations structured by the medical logic are hospitals. On average, they are larger and older than other organizations. All the public health clinics, representing 61yb R I W K H R U J D Q L ] D W L R Q V L Q W K H F D W H J R U \ D U H V W U X F W X U H G E \ W K H S X E O L F K H D O W K O R J L F 7 K e public health logic category also includes a few nonprofit organizations and nonprofit or public hospitals. One hundred percent of organizations structured by the social move ment logic are nonprofit organizations. They are on average the smallest and youngest or ganizations. Ninety-one percent of organizations structured by the multiservice logic are This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Journal of Public Administration Research and Theory Table 2 Institutional Logic Variable (N = 108 organizationsyf Institutional Logic Medical Public Health Social Movement Multiservice Organization type For-profit hospital (yb \f 26 0 0 9 Nonprofit hospital (yb \f 61 8 0 0 Public hospital (yb \f 2 13 0 0 Public health clinic (yb \f 0 61 0 0 Nonprofit/nonhospital (yb \f 11 18 100 91 Total (yb \f 100 100 100 100 Mean size 1628 673 112 323 (number of full time employeesyf Mean year of founding 1941 1961 1986 1970 N (organizationsyf 46 38 13 11 N (individualsyf 80 96 46 33 nonprofit social service and primary care clinics. Multiservice organizations are on average smaller and younger than medical or public health organizations. Individual- and Organizational-Level Variables To determine professional orientation, respondents were asked, "Are you (ayf D S K \ V L F L D Q ; (byf O L F H Q V H G W R S U D F W L F H D V D Q X U V H R U F \f a nonmedical counselor?" A dummy variable was scored 0 if the practitioner was a medical professional (physician or nurseyf D Q G L I W K e practitioner was a nonmedical counselor. A variable indicating the practitioner's tenure was used as a measure of experience. Measured in years in the current position, the variable was constructed from responses to the following question: "How many years have you been working in your current position?" To measure training, a dummy variable was scored 1 for practitioners who answered "yes" to the question, "Have you received any training in HIV counseling?" and 0 for practitioners who answered "no" to this question. A variable measuring knowledge of HIV testing and counseling was constructed from the response to: "How knowledgeable do you consider yourself to be with regard to HIV antibody testing?" Responses were coded on a scale with 0 indicating "not at all," 1 "somewhat," 2 "very," and 3 "extremely." A dichotomous variable was constructed and scored 0 for those who answered "not at all" or "somewhat" and 1 for those who answered "very" or extremely." Two measures of client need were constructed. To measure client serostatus, a dummy variable was scored 1 if the HIV test result was seropositive and 0 if seronegative. Respond ents were allowed to consult their charts for this information. To determine whether the client was identified as belonging to a behavioral risk group, respondents were asked, "What behavioral risk group best identifies this client: men who have sex with men, men who have sex with men/women, men who have sex with men and are intravenous drug users, intravenous drug users, women at sexual risk, or other?" The item was dummy coded 1 for clients who did not fall into one of the behavioral risk groups and 0 otherwise. At the organizational level, a workload variable was constructed in response to the following questions: (1yf 2 Q D Y H U D J H K R Z P D Q \ K R X U V D Z H H N G R \ R X V S H Q G G R L Q J S D W L H Q t or client care (working with clients?yf D Q G \f on average, how many patients or clients do you see in a typical week? We divided the number of clients per week by the number of This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Crusky Institutional Logic and Street-Level Discretion 115 hours spent on client or patient care per week for each respondent to create a measure of average number of clients per hour and then created an organizational-level score by av eraging all responses in the organization. Formalization was measured in two ways. As a measure of automation we asked an administrator, "Does the organization use the LAC Department of Health and Human Serv ices, Office of AIDS Program and Policy's HIV Information Resources System (HIRSyf " " HIRS is a widely used standardized form in LAC that partially automates the counseling session by guiding the practitioner to address particular issues during the encounter with clients such as the identification of risk factors. The item was scored 1 if the organization used the form and 0 otherwise. As a measure of the organization's reliance on formal rules, we used the response to: "This is a formalized, structured organization" (Quinn and Kimberly 1984yf 5 H V S R Q G H Q W V U D W H G W K H R U J D Q L ] D W L R Q R Q D V F D O H I U R P W R V F R U H G 1 when the description did not fit the organization at all to 100 when the description completely fit the organization. Control Variables We included three organizational-level controls in the model. Larger organizations have more slack resources and may provide more opportunities for HIV training (Blau and Schoenherr 1971yf 7 K H \ P D \ E H P R U H E X U H D X F U D W L ] H G D Q G V S H F L D O L ] H G D Q G P D \ U H W D L Q P R U e workers who specialize in HIV counseling. Therefore, organizational size may be posi tively associated with comprehensive counseling. Organizational size was measured as the number of full-time paid employees. Increased age is associated with inertia and re sistance to change (Selznick 1957; Stinchcombe 1965yf 2 U J D Q L ] D W L R Q V I R X Q G H G S U L R U W R W K e 1980s, when AIDS emerged as a significant social problem, may be less likely than newer organizations to implement service technologies designed to stem the spread of AIDS in a comprehensive manner. We included year of founding to control for this factor. Finally, to control for the possible influence of ownership type (D'Aunno, Vaughn, and McElroy 1999yf Z H L Q F O X G H G G X P P \ Y D U L D E O H V L Q G L F D W L Q J Z K H W K H U W K H R U J D Q L ] D W L R Q Z D V Q R Q S U R I L W , for-profit, or public. Model Specification Hierarchical linear modeling (HLMyf Z D V X V H G W R W H V W W K H K \ S R W K H V H V Z K L F K L V D S S U R S U L D W e when data are nested (Raudenbush and Bryk 2002yf + / 0 F R U U H F W V I R U L Q W U D F O D V V F R U U H O a tion that arises when observations are not independent due to nesting and it provides a method for modeling variations in level 1 outcomes as a function of level 2 factors. Using the HLM 6.2 statistical package, we employed HLM random intercept models, which model the average within-organization outcomes in the comprehensiveness of counseling as a function of individual, organizational, and field-level factors. 5 Two hundred and sixteen individual respondents are nested in 90 organizations in the sample available for analysis. As discussed by Hox (2002yf W K H S R Z H U R I W K H W H V W I R U W K H V L J Q L I L F D Q F H R I W K H L Q G L Y L G X D O O H Y H O U H J U H V V L R Q F R H I I L F L H Q W s depends on the total sample size of individual respondents. The power of tests of higher level coefficients depends more strongly on the number of groups (here, organizationsyf W K D Q R Q W K H Q X P E H U R I L Q G L Y L G X D O V S H U J U R X S $ Q D O \ W L F Z R U k suggests a trade-off between sample size at different levels, with a large number of groups (in the literature 90-100 is considered "large"yf D O O R Z L Q J D V P D O O Q X P E H U R I R E V H U Y D W L R Q V L Q H D F K J U R X S V H H H J 5 D X G H Q E X V K D Q G / L X \f. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Journal of Public Administration Research and Theory Table 3 Individual-Level Descriptive Statistics and Pairwise Correlations (A/ = 255 Practitionersyf Mean (SDyf 3 U R S R U W L R n 1 2 3 4 5 6 7 1. No. topics covered 7.58 (3.6yf 1.00 2. Proportion of .44 .36 1.00 nonmedical counselors 3. Tenure (in yearsyf 5.5 (6.21yf -.16 -.11 1.00 4. Proportion with any .79 .41 .30 -.01 1.00 training 5. Proportion with .64 .39 .27 .02 .14 1.00 knowledge 6. Proportion with .08 .11 .05 -.17 .02 .10 1.00 seropositive test result 7. Proportion not in .44 -.14 -.17 .21 -.07 -.10 -.20 1.00 behavioral risk group Since the outcome variable is a count of the number of topics covered, it signals the appropriateness of a Poisson distribution. However, because the variance of the outcome variable was greater than its mean (the data were overdispersedyf Z H X V H G D Q H J D W L Y H E i nomial distribution, which is more appropriate in case of overdispersion. At level one, the negative binomial model estimated the number of topics covered in the counseling session as a function of individual-level predictors. All level 1 predictors were group mean cen tered, allowing us to interpret the level 1 regression intercept as an estimate of the average number of topics covered in the counseling session for a particular organization. Separate intercepts were estimated for each organization and the coefficients for the level 1 inter cepts were modeled in the level 2 equation as a function of institutional logics and orga nizational-level covariates. For ease of interpretation, we calculated the incident rate ratios, which are the exponentiated coefficients. The expected increase or decrease in the count for a unit change in the independent variable, holding other variables constant, was then com puted as 100[exp(ßt*8yf @ O . For example, an incident rate ratio of 1.30 indicates that a unit change in the inde pendent variable results in a 30yb L Q F U H D V H L Q W K H H [ S H F W H G P H D Q Q X P E H U R I W R S L F V F R Y H U H G . We used the penalized quasi-likelihood estimation procedure. RESULTS Descriptive statistics and pairwise correlations for all variables are shown in tables 3 and 4. Of note, the pairwise correlation in table 4 shows that, as expected, workload is moderately and negatively related to the number of topics covered in the counseling session. Figures 1 and 2 show the mean number of topics covered in the counseling session at the individual (pooledyf D Q G R U J D Q L ] D W L R Q D O O H Y H O V I R U D O O F D W H J R U L F D O Y D U L D E O H V . Figure 1 shows that professional orientation makes a difference. Consistent with ex pectations, nonmedical counselors cover on average more topics in the counseling session (9.06yf W K D Q S K \ V L F L D Q V \f or nurses (6.75yf & R P S U H K H Q V L Y H Q H V V R I F R X Q V H O L Q J D O V R V H H P s This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion Table 4 Organizational-Level Descriptive Statistics and Pairwise Correlations (N = 108 Organizationsyf o o o •o o o vo «o m I I © oo r O ^ o (N -H ^ q i VO vo O O O rn rH ^ ^ H I f. I I ^ © t 8 E w G o o f'l . o< x I 00 H s i »G W V £ a ^3 c<3 v I oo -J2 £3 ■§ Oh G G O O '€ '€ o o D- G, O O Uh i-i 0h Cl, cd 00 o G 1 "3 is £ £ c £ .2 £ N G •2 .a o ts *5 G g g o ft P u 8 § .a 0* Ph 1/3 to &o 8 .9 g "I § S a <2 § V. -p Z o S3 & U S >< <£ —' On 00 (N 00 a i CM CO ■f o o o M OO 1^1 N. o o sS w' sc w CO u O 4yf ►5 ^ c • B -ë 2 £ jg w 00 O « I § t . _ ^ u te c p a. -a 3 -9 g ot O c g > Ë ,°s O 3 3 U c ai b 1 ° C 4? (N rn r rn & Sc 3 (U ö t-yf mThis content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 122 Journal of Public Administration Research and Theory Individual- and Organizational-Level Factors None of the individual-level variables are significantly associated with comprehensiveness of counseling. As predicted, however, the organizational-level factors are significantly re lated to comprehensiveness of counseling. A one-unit increase in workload (average num ber of clients served per houryf U H V X O W V L Q D \b decrease in the number of topics covered during the pretest counseling session, net the effect of all other variables in the model. As predicted, use of the HIRS form apparently constrains counseling practices, resulting in a 17yb L Q F U H D V H L Q W K H Q X P E H U R I W R S L F V F R Y H U H G 8 Q H [ S H F W H G O \ J U H D W H U I R U P D O L ] D W L R Q V H H P s to suppress counseling. A one-unit increase in formalization (on a scale from 1 to 100yf results in an expected 1yb G H F U H D V H L Q W K H Q X P E H U R I F R X Q V H O L Q J W R S L F V F R Y H U H G 3 X W D Q R W K H r way, a 10-point increase in the formalization score decreases the expected number of topics by a little more than one. Control Variables Organizational size is significantly and negatively associated with comprehensiveness of counseling. This unexpected finding may reflect the correlation between size and core institutional logic as the medical logic, which is significantly and negatively associated with comprehensiveness, is positively associated with size. The social movement logic, which is significantly and positively associated with comprehensiveness, is moderately and negatively correlated with size (table 4yf 7 K H S D U D P H W H U H V W L P D W H I R U R Z Q H U V K L S W \ S H L s not significant. DISCUSSION A basic insight of the literature on street-level bureaucracy is that workers who implement policy on the frontlines become, in effect, lower level policy makers since their behaviors constitute policy in practice. Most research on street-level bureaucracy has examined how street-level workers use their discretion to respond to shared conditions of work in predict able ways but has largely neglected to consider how and why workers' discretionary be havior may differ systematically across wider contexts. Drawing on insights from the institutional logics perspective (Friedland and Alford 1991; Thornton and Ocasio 2008yf W K L V D U W L F O H W H V W V W K H L G H D W K D W Z R U N H U V G L I I H U L Q W K H L U G L V F U H W L R Q D U \ E H K D Y L R U D Q G W K D t the differences are related to the underlying logic of the organizational field in which the work is embedded. By extending the focus of explanation to include the broader institu tional context, our model implies the interplay of agency and institutional logics and calls into question the assumption of individualistic interests that underlies much of the research on street-level discretion. Institutional logics may influence policy implementation by providing cultural and material repertoires that shape workers' understandings of the means and ends of their interests. Logics allocate the attention of workers by defining the purpose of their organi zation, the nature of the problems they face, appropriate organizational responses to these problems, the relevant attributes of clients, the value of tasks, and their own roles and scope of work. We expected that when policies are consistent with core field-level institutional logics, it is more likely that implementation will be consistent with policy intent, even when This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Crusky Institutional Logic and Street-Level Discretion 123 accounting for variations in individual- and organizational-level factors shown to influence street-level discretion. Results provide support for this perspective. First, the organizational field seems to influence frontline practices. When compared to a model with individual factors and controls, a model that included institutional logics, individual-level factors, and controls explained about the same amount of between organization variance in HIV pretest counseling (around 60yb \f. Together, individual and organizational-level factors, institutional logics, and controls explained over 70yb of between-organization variance in counseling. Moreover, results demonstrate that institutional logics influence street-level practice by directing practitioners to implement service technologies congruent with core field-level logics. Among the four institutional logics that emerged from the analysis, two provide a stark contrast in fit with HIV pretest counseling policy. The medical logic, which is ori ented more toward medical care than prevention, provides the poorest fit for HIV test coun seling. By contrast, the social movement logic, which stems from HIV/AIDS social movements, provides the best fit with HIV pretest counseling policy. This logic emerged in opposition to prevailing logics as an alternative which identifies with and attempts to empower clients who are at risk for HIV, many of whom have been devalued by society (see, e.g., Epstein 1996yf 2 U J D Q L ] D W L R Q V R S H U D W L Q J X Q G H U W K H V R F L D O P R Y H P H Q W O R J L F P D y therefore make comprehensive HIV pretest counseling, which provides information and counseling to help the client gain insight and voluntarily make better choices, a top priority. As expected, workers in organizations structured by the medical logic implemented HIV test counseling to a significantly lesser degree than workers in organizations structured by social movement logic, even when controlling for individual- and organizational-level correlates. This finding provided strong support for the model. Unexpectedly, the public health logic was not significantly related to comprehensive ness of counseling. We reasoned that the public health logic provides a good fit with HIV pretest counseling because it emphasizes prevention rather than treatment and therefore expected that workers operating under this logic would perceive the value in HIV pretest counseling as a prevention tool and implement it in a comprehensive manner. In retrospect it may be that the public health logic's degree of fit with the policy intent of HIV pretest counseling was overstated and/or our measures were insufficiently robust. Specifically, the client-centered approach promoted by the CDC's policy on pretest counseling may conflict with the public health field's traditional disease control logic. His torically, the public health response to lethal infectious diseases has at times prioritized containment of disease control over private rights, justifying a coercive repertoire of prac tices such as "compulsory examination and screening, breaching the confidentiality of the clinical relationship by reporting to public health registries the names of those with diag noses of "dangerous diseases," imposing treatment, and in the most extreme cases, con fining persons through the power of quarantine" (Bayer 1991a, 1500yf , W Z D V W K H H I I R U W V R f gay rights advocacy groups during the early years of the AIDS epidemic that pushed na scent policies surrounding HIV testing and counseling toward voluntary, anonymous, and confidential practices that prioritized self-determination over public health concerns (Bayer 1991byf $ V D U H V X O W + , 9 W H V W L Q J D Q G F R X Q V H O L Q J S R O L F \ K D V E H H Q I U D X J K W Z L W K F R Q W U R Y H U V y and marked by sharp policy debates over issues such as reporting of the names of sero positives, partner notification, and informed voluntary consent to test. Indeed, in 2006, the CDC revised its recommendations for HIV testing to drop its recommendation for risk This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 124 Journal of Public Administration Research and Theory assessment and pretest counseling (Branson et al. 2006yf 7 K L V Z D V D W H V W D P H Q W W R W K H G L s puted legitimacy of pretest counseling as a tool of prevention in the public health arena. We expected that the multiservice logic, which provides a moderately good fit with HIV pretest counseling, would fall somewhere in-between the medical logic and the social movement logic in terms of comprehensiveness of counseling. This seemed to be the case, and practitioners operating in a field structured by this logic did not provide counseling significantly more comprehensive than counseling provided in organizations structured by the medical logic. It is often argued that professional orientation, experience, training, and knowledge determine frontline practice (Abbot 1992; Jewell and Glaser 2006; Meyers, Riccucci, and Lurie 2001; Riccucci et al. 2004yf + R Z H Y H U Z K H Q Z H D F F R X Q W H G I R U Y D U L D W L R Q D F U R V V R r ganizations, we found no significant differences between medical practitioners (physicians and nursesyf D Q G Q R Q P H G L F D O F R X Q V H O R U V L Q F R P S U H K H Q V L Y H Q H V V R I F R X Q V H O L Q J R U E H W Z H H n practitioners with varying levels of experience with, training in, or knowledge about HIV testing and counseling. We also failed to detect a relationship between comprehen siveness of counseling and behavioral risk group. Our inability to detect a significant relationship between individual-level factors and comprehensiveness of counseling may result from the mediating effect of institutional con text. For example, when comparing the mean number of topics covered in the counseling session shown in figure 1, it appears that physicians and nurses provide on average sub stantively less comprehensive HIV pretest counseling than nonmedical counselors. How ever, once variation across organizational and institutional settings is taken into account, these differences disappear, suggesting that the zero-order association between profes sional orientation and comprehensiveness may lie in physicians' and nurses' greater rep resentation (relative to nonmedical counselorsyf L Q L Q V W L W X W L R Q D O V H W W L Q J V W K D W G H S U L R U L W L ] e counseling. Future research should carefully test this proposition. In keeping with extant research, our findings suggest that increased workload pres sures (average number of clients served per houryf G H F U H D V H W K H F R P S U H K H Q V L Y H Q H V V R I F R X n seling and automation (use of the HIRS formyf L Q F U H D V H V F R P S U H K H Q V L Y H Q H V V 8 Q H [ S H F W H G O \ , the second measure of formalization, which captured the extent to which respondents per ceived that the organization was formalized and structured, was negatively related to the comprehensiveness of counseling. This finding may result in part from the particular in stitutional context in which HIV testing and counseling is carried out. The social movement logic, for example, which promotes comprehensive counseling, may be populated by small social movement organizations that eschew formal structure. This analysis has several implications for organizational research. First, by comparing policy implementation across organizational fields, the analysis illuminates how workers' discretionary practices varied systematically by field-level institutional logics, even when taking into account individual and organizational-level factors. Comprehensiveness of counseling was significantly associated with the institutional logics that structured the or ganizational context, suggesting that worker agency, as expressed through discretionary practice, is embedded in an institutional system that shapes how discretion is exercised by providing institutionally defined means, ends, and interests. Second, our analysis suggests how street-level discretion (as well as the variation in organizational practices it enablesyf L V E R X Q G H G E \ W K H L Q V W L W X W L R Q L Q Z K L F K L W R F F X U V 3 U D F W i tioners who carry out policies are subject to field-level institutional pressures that may This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion 125 conflict with or complement policy intent. Because the means and ends of their interests are institutionally shaped, workers may use their discretion to implement policies congruent with the core logics that structure the organizational field, while giving less attention to policies that are incongruent with field-level logics. Our approach enriches the literature on street-level discretion by tempering notions of agency with the insight that agency is institutionally embedded. It also diverges from the efforts of some institutional theorists to accommodate agency and interests by arguing that conformity to institutional pressures is a strategic choice driven by organizational interests (Oliver 1991; Scott 1991; Goodstein 1994yf 5 D W K H U R X U I L Q G L Q J V L Q G L F D W H W K D W R U J D Q L ] D W L R Q s and street-level workers engage in discretionary practices but within the context of con straints and opportunities imposed by the institutional orders under which they operate, rather than outside of them (e.g., Friedland and Alford 1991; Goodrick and Salancik 1996; Seo and Creed 2002yf . The results also have implications for policy implementation. In calling attention to the ways in which policies are filtered through the structuring logics of organizational fields, the analysis suggests that policy implementation should, at least in part, focus on the organizational field. The results demonstrate that when organizations operate in the context of field logics that are consistent with the policy intent they are more likely to implement the policy. To encourage effective policy implementation, policy makers could encourage and support fields that uphold the intent of policy or create incentives for fields to adopt the policy's logic. Although the results provide support for our model, several methodological issues limit the ability to generalize from this study. First, the cross-sectional study design does not allow ruling out the possibility that practitioners self-selected into organizations com patible with their world views, although a number of individual-level factors, such as pro fessional orientation, training, and knowledge, which help shape practitioners' perspectives were controlled. For example, practitioners who are HIV + or members of high-risk groups targeted by HIV test counseling may be more likely to fully implement the counseling protocol than others and may also be more likely to work in the field structured by the social movement logic (see Watkins-Hayes 2009 for an example of how the interaction between institutional context and attributes of workers such as race and class influences street-level discretionyf : H G R Q R W K D Y H W K H G D W D W R F R Q W U R O I R U W K L V S R V V L E L O L W \ 6 H F R Q G 4 organizations were dropped from the analysis because none of the interviewed workers had conducted an HIV test in the last 6 months, creating the potential for bias in the observed results.8 We tentatively suggest that the findings can be generalized to organizations that regularly conduct HIV testing and counseling but not to organizations that offer, but have not recently conducted, HIV testing and counseling. Third, we lacked data to control for the effects of resource dependence. It is possible that, in an effort to ensure continued funding, organizations dependent on revenues from HIV testing implement HIV pretest counseling in order to comply with funding requirements (Pfeffer and Salancik 1978yf , Q I D F W U H V R X U F e dependence could explain not only conformity to CDC recommended protocol but also reporting that such guidelines are followed, regardless of actual practices. However, an 8 For example, it is possible that people seeking HIV testing and counseling are less likely to patronize organizations that provide less comprehensive counseling. Nonetheless, we felt that excluding from the analysis organizations in which none of the interviewed workers had conducted an HIV test in the last 6 months was justified because of potential bias caused by faulty recall of the counseling session. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms 126 Journal of Public Administration Research and Theory explanation based on resource dependence may not be applicable in this case because most costs of HIV counseling and testing are reimbursed by the CDC, which disburses funds to county health departments who then distribute them to service organizations and their pro viders. Although the CDC does provide financial and technical support for HIV antibody counseling, testing, and referral services and risk-reduction counseling, it does not hold organizations accountable for implementation of best practices. Therefore, exposure to re source-based rewards or sanctions linked to compliance with CDC guidelines for HIV test counseling was probably quite minimal. Fourth, we are not able to control for the com plexity of organizational work. Respondents may be primary care providers, or do intake, or provide more complex services. Although we have controlled for professional orienta tion and workload, we cannot completely capture in our measures the nature of professional responsibilities. Finally, for the sake of simplicity, we identified one dominant institutional logic for each organization. However, the assumption that there is a single predominate institutional logic to which an organization adheres rather than multiple and potentially competing institutional logics is questionable (Kraatz and Block 2008yf , W L V U H D V R Q D E O e to assume that the more complex, highly differentiated, heterogeneous, and distributed the organization the greater the number of institutional logics to which it attempts to con form. CONCLUSION This study provides evidence that, in addition to organizational-level factors, the degree of compliance with policy mandates is associated with characteristics of the organizational field. Specifically, we showed that the core field-level logic can impact critical organiza tional outputs such as the well-being of service recipients and society by influencing how policies are or are not implemented by frontline practitioners. These findings support the perspective that institutional logics have important consequences for health and human services because they supply the moral categories and legitimated practices that play a key role in shaping the quality of services provided to vulnerable client groups. Work on or with people is justified on the basis of adherence to institutionalized moral rules that define target client groups, their relevant attributes, and what can be done to or with them. Future research should examine the influence of core logics on other routine frontline prac tices in the health and human services sector, such as mental health treatment or basic needs assistance, which, like HIV test counseling, are conducted across diverse institutional and organizational contexts. Appendix Table AI Organization Participation by Stratum (A/yf Organizations For-profit Nonprofit, Public, with Mobile Hospitals Private Practices Nonhospitals Nonhospitals Testing Units Total Sample frame 86 37 125 32 14 294 Sample 100yb \f 100yb \f 38yb \f 100yb \f 100yb \f 74yb \f Response rate 63yb \f 51yb \f 77yb \f 88yb \f 100yb \f 70yb \f This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion Table A2 Share of Organizations with Type of Service, by Institutional Logic Category Services Logic Medical Public health Social movement Multiservice All organizations Social Case Mental Advocacy Services Management Counseling Education Prevention Health Medical 0yb \b 2.63 2.63 76.92 76.92 18.18 54.55 12.04yb \b 2.17yb \b 2.63 2.63 76.92 76.92 72.73 72.73 18.52yb \b 4.35yb \b 15.79 42.11 46.15 30.77 72.73 9.09 20.37yb \b 0yb \b 5.26 97.37 53.85 15.38 36.36 100 12.04yb \b Table A3 Share of Organizations that Serve Target Group, by Institutional Logic Category Target groups Logic Medical Public health Social movement Multiservice All organizations Low Minority Vulnerable, General income Uninsured ethnic group HIV LGBT Homeless other public 0yb \b 0yb \b 0yb \b 0yb \b 100 10.53 5.26 7.89 0 2.63 0 65.79 30.77 0 30.77 84.62 23.08 15.38 23.08 0 100 63.64 18.18 0 0 0 18.18 49.07yb \b 7.41yb \b 2.78yb \b 4.63yb FUNDING This work was supported by the National Institutes of Health, National Institute of Mental Health (RO1 -MH-62709-PI-O.Gruskyyf . REFERENCES Abbot, Andrew. 1992. Professional work. In Human services as complex organizations, ed. Yeheskel Hasenfeld, 145-62. Newbury Park, CA: Sage. Allison, Paul. 2002. Missing data. Thousand Oaks, CA: Sage. Baum, Joel A. C., and Christine Oliver. 1992. Institutional embeddedness and the dynamics of organi zational populations. American Sociological Review 57:540-59. Bayer, Ronald. 1991a. Public health policy and the AIDS epidemic: An end to HIV exceptionalism? The New England Journal of Medicine 324:1500-4. . 1991 b. Private acts, social consequences: AIDS and the politics ofpublic health. New Brunswick, NJ: Rutgers Univ. Press. Blau, Peter M., and Richard A. Schoenherr. 1971. The structure of organizations. New York, NY: Basic Books. Branson, Bernard M., H. Hunter Handsfield, Margaret A. Lampe, Robert S. Janssen, Allan W. Taylor, Sheryl B. Lyss, and Jill E. Clark. 2006. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in healthcare settings. MMWR 55:1-18. Brehm, John, and Scott Gates. 1997. Working, shirking, and sabotage: Bureaucratic response to a democratic republic. Ann Arbor, MI: Univ. of Michigan Press. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Journal of Public Administration Research and Theory Brewer, Gene A. 2005. In the eye of the storm: Frontline supervisors and federal agency performance. Journal of Public Administration Research and Theory 15:505—27. Brodkin, Evelyn Z. 1997. Inside the welfare contract: Discretion and accountability in state welfare administration. Social Service Review 71:1-33. . 2008. Accountability in street-level organizations. International Journal of Public Administration 31:317-36. CDC. 2001. Revised guidelines for HIV counseling, testing, and referral. MMWR 50:1—58. http:// www.cdc.gov/mmwr/preview/mmwrhtml/rr5019al.htm (accessed February 9, 2012yf . Cohan, Deborah, Elvira Gomez, Teri Dowling, Nicola Zetola, Beth Kaplan, and Jeffrey Klausner. 2009. HIV testing attitudes and practices among clinicians in the era of updated Centers for Disease Control and Prevention recommendations. Journal of Acquired Immune Deficiency Syndromes 50:114-6. D'Aunno, Thomas, Thomas E. Vaughn, and Peter McElroy. 1999. An institutional analysis of HIV prevention efforts by the nation's outpatient drug abuse treatment units. Journal of Health and Social Behavior 40:175—92. DiMaggio, Paul, and Ann Mullen. 2000. Enacting community in progressive America: Civic rituals in National Music Week, 1924. Poetics 27:135-62. DiMaggio, Paul J., and Walter W. Powell. 1983. The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review 48:147-60. . 1991. The iron cage revisited: Institutional isomorphism and collective rationality in organi zational fields. In The new institutionalism in organizational analysis, ed. Walter Powell and Paul DiMaggio, 63-82. Chicago, IL: Univ. of Chicago Press. Epstein, Steven. 1996. Impure science: AIDS, activism, and the politics of knowledge. Berkeley, CA: Univ. of California Press. Ewalt, Jo Ann G., and Edward T. Jennings, Jr. 2004. Administration, governance, and policy tools in welfare administration. Public Administration Review 64:449-62. Fincher-Mergi, Melissa, Kathy Jo Cartone, Jean Mischler, Patricia Pasieka, E. Brooke Lerner, and Anthony J. Billittier. 2002. Assessment of emergency department health care professionals' be haviors regarding HIV testing and referral for patients with STDs. AIDS Patient Care and STDs 16:549-53. Fixsen, Dean L., Sandra F. Naoom, Karen A. Blase, Robert M. Friedman, and Frances Wallace. 2005. Implementation research: A synthesis of the literature. Tampa, FL: Univ. of South Florida, Louis de la Parte Florida Mental Health Institute, the National Implementation Research Network. Fording, Richard C., Joe Soss, and Sanford F. Schram. 2007. Devolution, discretion, and the effect of local political values on TANF sanctioning. Social Service Review 81:285—316. Friedland, Roger, and Roger R. Alford. 1991. Bringing society back in: Symbols, practices, and insti tutional contradictions. In The new institutionalism in organizational analysis, ed. Walter Powell and Paul DiMaggio, 232-63. Chicago, IL: Univ. of Chicago Press. Friedson, Eliot. 1970. Profession of medicine. New York, NY: Harper and Row. Glauser, Jonathan. 2001. Rationing and the role of the emergency department as society's safety net. Academic Emergency Medicine 8:1101-6. Goodrick, Elizabeth, and Gerald R. Salancik. 1996. Organizational discretion in responding to institutional practices: Hospitals and cesarean births. Administrative Science Quarterly 41:1—28. Goodstein, Jerry D. 1994. Institutional pressures and strategic responsiveness: Employer involvement in work-family issues. Academy of Management Journal 37:350-82. Gordon, James A., John Billings, Brent R. Asplin, and Karin V. Rhodes. 2001. Safety net research in emergency medicine proceedings of the Academic Emergency Medicine Consensus Conference on "The Unraveling Safety Net". Academic Emergency Medicine 8:1024—9. Handler, Joel F., and Ellen Jane Hollingsworth. 1971. The deserving poor: A study of welfare admin istration. New York, NY: Academic Press. Hasenfeld, Yeheskel. 1983. Human service organizations. Upper Saddle River, NJ: Prentice Hall. . 2000. Organizational forms as moral practices: The case of welfare departments. Social Service Review 74:329-51. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Garrow and Grusky Institutional Logic and Street-Level Discretion . 2010. Worker-client relations. Social policy in practice. In Human Services as Complex Organizations, ed. Yeheskel Hasenfeld, 405—26. Los Angeles, CA: Sage. Hasenfeld, Yeheskel, and Dale Weaver. 1996. Enforcement, compliance, and disputes in welfare-to-work requirements. Social Service Review 70:235—56. Hill, Carolyn J. 2006. Casework job design and client outcomes in welfare-to-work offices. Journal of Public Administration Research and Theory 16:263—88. Hox, Joop. 2002. Multilevel analysis: Techniques and applications. Mahwah, NJ: Lawrence Erlbaum. Jackall, Robert. 1988. Moral mazes: The world of corporate managers. New York, NY: Oxford Univ. Press. Jewell, Christopher J., and Bonnie E. Glaser. 2006. Toward a general analytic framework: Organizational settings, policy goals, and street-level behavior. Administration & Society 38:335-64. Keiser, Lael R. 1999. State bureaucratic discretion and the administration of social welfare programs: The case of social security disability. Journal of Public Administration Research and Theory 9:87—106. Knudsen, Hannah K., Lori J. Ducharme, and Paul M. Roman. 2007. The use of antidepressant medications in substance abuse treatment: The public-private distinction, organizational compatibility, and the environment. Journal of Health and Social Behavior 48:195-210. Kraatz, Matthew S., and Emily S. Block. 2008. Organizational implications of institutional pluralism. In Organizational institutionalism, ed. R. Greenwood, C. Oliver, R. Suddaby, and K. Sahlin, 243-75. Los Angeles, CA: Sage. Lin, Ann C. 2000. Reform in the making: The implementation of social policy in prison. Princeton, NJ: Princeton Univ. Press. Lipsky, Michael. 1980. Street-level bureaucracy. New York, NY: Russell Sage. Marks, Gary, Nicole Crepaz, and Robert S. Janssen. 2006. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 20:1447-50. Maynard-Moody, Steven, and Michael C. Musheno. 2003. Cops, teachers, and counselors: Stories from the frontlines of public service. Ann Arbor, MI: Univ. Michigan Press. McDonald, Thomas P., and Irving Piliavin. 1984. Failure to participate in AFDC: Some correlates and possible influences. Social Work Research and Abstracts 20:17-22. McLaughlin, Janice, Paul Rosen, David Skinner, and Andrew Webster. 1999. Valuing technology: Organizations, culture, and change. London, UK: Routledge. Meyer, John W., and Brian Rowan. 1977. Institutionalized organizations: Formal structure as myth and ceremony. The American Journal of Sociology 83:340-63. Meyers, Marcia K., Bonnie Glaser, and Karin Mac Donald. 1998. On the front lines of welfare delivery: Are workers implementing policy reforms? Journal of Policy Analysis and Management 17:1-22. Meyers, Marcia K., Norma M. Riccucci, and Irene Lurie. 2001. Achieving goal congruence in complex environments: The case of welfare reform. Journal of Public Administration Research and Theory 11:165-202. Meyers, Marcia K, and Susan Vorsanger. 2003. Street-level bureaucrats and the implementation of public policy. In Handbook of public administration, ed. B. Guy Peters and John Pierre, 245—55. Thousand Oaks, CA: Sage. Möhr, John W. 1998. Measuring meaning structures. Annual Review of Sociology 24:345—70. Möhr, John W., and Vincent Duquenne. 1997. The duality of culture and practice: Poverty relief in New York City. Theory and Society 26:305—56. Möhr, John, and Francesca Guerra-Pearson. 2010. The duality of niche and form: The differentiation of institutional space in New York City, 1888—1917. In Categories in markets: Origins and evolution (Research in the sociology of organizations, Volume 31yf H G * U H W D + V X * L D F R P R 1 H J U R D Q d Ozgecan Koçak, 321-68. Bingley, UK: Emerald Group. Mullen, Edward J., and William Bacon. 2004. A survey of practitioner adoption and implementation of practice guidelines and evidence-based treatments. In Evidence-based practice manual: Research and outcome measures in health and human services, ed. Albert R. Roberts and Kenneth R. Yeager, 210-8. New York, NY: Oxford Univ. Press. Ocasio, William. 1995. The enactment of economic adversity: A reconciliation of theories of failure induced change and threat-rigidity. Research in Organizational Behavior 17:287—331. This content downloaded from f:ffff:ffff:ffff:ffff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC All use subject to https://about.jstor.org/terms Journal of Public Administration Research and Theory Oliver, Christine. 1991. Strategie responses to institutional processes. Academy of Management Review 16:145-79. Panzano, Phyllis, and Lon Herman. 2005. Developing and sustaining evidence-based systems of mental health services. In Evidence-based mental health practice, eds. Robert E. Drake, Matthew R. Merrens, and David Lynde, 167—87. New York, NY: Norton. Pfeffer, Jeffrey, and Gerald R. Salancik. 1978. The external control of organizations. New York, NY: Harper and Row. Prottas, Jeffrey Manditch. 1979. People-processing: The street-level bureaucrat in public service bu reaucracies. Lexington, MA: Lexington Books. Quinn, Robert E., and John R. Kimberly. 1984. Managing organizational transitions. Homewood, IL: Dow Jones-Irwin. Raudenbush, Stephen W., and Anthony S. Bryk. 2002. Hierarchical linear models: Applications and data analysis methods, 2nd ed. Thousand Oaks, CA: Sage. Raudenbush, Stephen W., and Xiaofeng Liu. 2000. Statistical power and optimal design for multisite randomized trials. Psychological Methods 5:199-213. Riccucci, Norma M. 2005. How management matters: Street-level bureaucrats and welfare reform. Washington, DC: Georgetown Univ. Press. Riccucci, Norma M., Marcia K. Meyers, Irene Lurie, and Jun Seop Han. 2004. The implementation of welfare reform policy: The role of public managers in front-line practices. Public Administration Review 64:438—48. Richardson, Lynne D., and Ula Hwang. 2001. Access to care: A review of the emergency medicine literature. Academic Emergency Medicine 8:1030-6. Sandfort, Jodi R. 1999. The structural impediments to human service collaboration: Examining welfare reform at the front lines. Social Service Review 73:314—39. . 2000. Moving beyond discretion and outcomes: Examining public management from the front lines of the welfare system. Journal of Public Administration Research and Theory 10:729-56. . 2010. Human service organizational technology: Improving understanding and advancing research. In Human services as complex organizations, 2nd ed., ed. Yeheskel Hasenfeld, 269-90. Thousand Oaks, CA: Sage. Schneider, Anne, and Helen Ingram. 1993. Social construction of target populations: Implications for politics and policy. American Political Science Review 87:334—47. Scott, W. Richard. 1991. Unpacking institutional arguments. In The new institutionalism in organizational analysis, eds. Walter Powell and Paul DiMaggio, 164—82. Chicago, IL: Univ. of Chicago Press. . 2008. Institutions and organizations, 3rd ed. Thousand Oaks, CA: Sage. Selznick, Philip. 1957. Leadership in administration: A sociological interpretation. New York, NY: Harper and Row. Seo, Myeong-Gu, and W. E. Douglas Creed. 2002. Institutional contradictions, praxis, and institutional change: A dialectic perspective. The Academy of Management Review 27:222-47. Smith, Brenda D., and Stella Donovan. 2003. Child welfare practice in organizational and institutional context. Social Service Review 77:541-63. Stern, Robert S., Joel S. Weissman, and Arnold M. Epstein. 1991. The emergency department as a pathway to admission for poor and high-cost patients. JAMA 266:2238-43. Stinchcombe, Arthur L. 1965. Social structure and organizations. In Handbook of organizations, ed. James G. March, 142—93. Chicago, IL: Rand McNally. Taylor, Todd B. 2001. Threats to the health care safety net. Academic Emergency Medicine 8:1080-7. Tilly, Charles. 1997. Parlimentarization of popular contention in Great Britain, 1758-1834. Theory and Society 26:245—73. Thornton, Patricia H., and William Ocasio. 1999. Institutional logics and the historical contingency of power in organizations: Executive succession in the higher education publishing industry, 1958— 1990. The American Journal of Sociology 105:801—43. . 2008. Institutional logics. In The SAGE handbook of organizational institutionalism, ed. Royston Greenwood, Christine Oliver, Roy Suddaby, and Kerstin Sahlin-Andersson, 99-129. Thousand Oaks, CA: Sage. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms Carrow and Grusky Institutional Logic and Street-Level Discretion Watkins-Hayes, Celeste. 2009. The new welfare bureaucrats: Entanglements of race, class, and policy reform. Chicago, IL: Univ. of Chicago Press. Wenger, Jeffrey B., and Vicky M. Wilkins. 2009. At the discretion of rogue agents: How automation improves women's outcomes in unemployment insurance. Journal of Public Administration Research and Theory 19:313—33. This content downloaded from 128.6.45.205 on Thu, 26 Aug 2021 00:37:48 UTC All use subject to https://about.jstor.org/terms