Fig 12 - uploaded by Manuel S Gonzalez Canche
Content may be subject to copyright.
Context in source publication
Context 1
... manipulation #Transforming the graph to a matrix, following the same rationale as in equation (7), this renders Table 3. v2m<-as.matrix(get.adjacency(g2m)) #The following line solves the issue discussed in the content of the chapter with respect to over-representation of zeroes v2m<-v2m[1:i,(i+1):(i+j)] v2m#Trimmed version #Two-to one-mode keeping human actors as shown in Table 4 and equation (12) v1mactor<-v2m%*%t(v2m) v1mactor #Two-to one-mode keeping non-human actors as shown in Table 5, equation (13) v1mcommunity<-t(v2m)%*%v2m v1mcommunity #A resulting warning is ok #Plotting the map for Figure 3 x11()#This opens a plotting window. ee+ geom_point(aes(x=lon, y=lat, colour = Sector), data=coord_df[InstChar$UNITID==215062 ,], size=15) + scale_color_manual(values=c("red"))+ ylab("Latitude") + xlab("Longitude") + labs(caption = "Map created using official NCES, IPEDS data") + ggtitle("Big Dot Represents the University of Pennsylvania") #Plotting the map for Figure 4(a) x11()#This opens a plotting window. ...
Similar publications
En el presente documento se desarrollan un conjunto
de ideas respecto a la investigación en educación ambiental,
considerando los siguientes capítulos: el campo de la educación
ambiental, la investigación en el campo de la educación ambiental,
paradigmas de la investigación en educación ambiental, perspectivas
metodológicas de la investigación en e...
This report from the Metropolitan Educational Research Consortium (MERC) explores research related to Advanced Placement (AP) courses through an equity lens. It answers five questions: 1) What are AP classes? 2) Who enrolls and succeeds in AP classes? 3) Why do disparities in AP matter? 4) What factors contribute to disparities in AP participation...
Citations
... This indicates that quantitatively these measures may present significant correlations. Nonetheless, these correlations are not perfectly linear (González Canché, 2019, 2024, for the algorithms were designed to qualitative capture distinct roles in a given network. From this view, our integrative mixed methods analysis will build from their quantitative positions but will also highlight separate roles and levels of influence of participants and their codes (Borgatti, 2006). ...
We offer insights into the factors impacting faculty-led academic/research collaborations between Mexican scholars employed in the USA and their Mexican colleagues working in Mexico. Founded on the idea that diasporic relationships include people involved in cross-border migrations yet maintaining ties with their homeland, we are referring to these faculty-led collaborations as diasporic. To offer nuanced understandings, data analyzed were obtained from 25 semi-structured interviews exploring collaboration in different professional, institutional, disciplinary, and regional contexts. Relying on Network Analysis of Qualitative Data, we were able to identify the most relevant drivers (e.g., personal relationships, common research interests, and cross-cultural understandings) and deterrents (e.g., political and legal challenges and institutional contexts) of diasporic collaborations influenced by institutional, national, and sociopolitical power dynamics. Our use of diasporic academic collaborations is intended to transcend this study; that is, although our analytic sample is comprised by diasporic Mexican academics, we argue that similar barriers and drivers may apply to academics from other countries who may be interested in participating in diasporic academic collaborations. Accordingly, we invite other researchers to expand this understudied research topic by providing access to our interview protocols and the detailed list of codes used to apply Network Analysis of Qualitative Data.
... In multilevel modeling, ∆ is a block diagonal design matrix that is traditionally referred to as an ascription matrix in the network analysis literature (Breiger, 1974, 388 8 SSEM Regression Based Analyses 2018a, 2019aWolf et al., 2021), where units of different types are related to one another. Earlier in this book we have referred to these connections as two-mode networks. ...
Although spatial outcome dependence (SODA) is informative in its own right for it allows us to observe social and spatial influence, the findings are univariate, which constrain our capacity to model factors that may impact outcome variation. Specifically, in the social sciences we are more interested in multivariate analyses, wherein we may explain the expected variation of an outcome as a function of predictor and control indicators in a statistical model. From this view, in this chapter we build from the insights gained with SODA and illustrate that, when outcome dependence is corroborated in our data structures, this lack of independence among units makes it necessary for us to rely on regression-based approaches capable of handling SODA. Accordingly, in this chapter we present four spatial regression models: Simultaneous Autoregressive (SAR) models, Multilevel SAR models, Geographically Weighted Regression, and Spatio-temporal or Panel SAR models. The first two (SAR and Multilevel SAR) were designed to handle SODA relying on coefficient estimates that are applied “globally” (i.e., across all areas or zones regions included in the analytic sample) and “cross-sectionally” (i.e., one point in time). In this respect, to the extent that the relationships among splace-based indicators in these areas or zones are somehow spatially homogeneous these areas or zones are somehow spatially homogeneous, the assumption of global regression coefficients being sufficient to explain spatial mechanisms, may be feasible. However, when at least some relationships in these areas, zones, or regions are spatially heterogeneous in nature (i.e., the impact of poverty rates on the outcome vary across the splace studied), the relationships among outcomes and predictors of interest may vary across zones. If this is true then, global coefficients may not realistically capture this variation. From this view, Geographically Weighted regression tests whether spatial heterogeneity may better explain the relationships observed among our splaces. Similarly, in instances where, in addition to place-based indicators, we observe these splaces in multiple time points, cross-sectional estimates may fall short in modeling these temporal and spatial sources of variation. Accordingly, the fourth regression model we discuss in this chapter, offers an expansion of SAR by modeling longitudinal panel or repeated cross-sectional spatial data structures. Since the level of feature engineering or data management complexity increases with multi-level, geographically weighted, and spatio-temporal approaches, we have designed minimal code functions to ease access to the data handling and manipulation demanded by the use of multiple cross-sectional matrices of influence or longitudinal spatial matrices. In each of these spatial regression models we illustrate how these four analytic strategies may be applied and assess their performances relying on fully reproducible databases.
... (González Canché, 2019). This proximity for students reinforces how community colleges continue to offer their commitment to their local communities by serving students who live and attend college locally. ...
Public 2-year colleges were designed to serve students coming primarily from their local communities, hence explaining their “community college” title. This study focuses on an understudied factor capable of altering the local workforce impact of community colleges while also potentially strengthening their 4-year steppingstone function: Transfer articulation agreements with 4-year institutions located out of state. This article documents the magnitude and structure of these agreements focusing on institutions’, states’, and programs’ prevalence. The analyses may serve as a reference or guide to institution and system leaders who may be considering engaging in these agreements. The discussion identifies programs and states wherein these agreements may be more easily established and offers recommendations that aim to protect the financial and academic well-being of transfer students. Current demographic changes place institution and system leaders in an advantaged position to negotiate agreements and take actions to avoid jeopardizing the local impact of community colleges.
... Such analytic approaches are rarely applied in study abroad research generally (McAllister-Grande & Whatley, 2020) and have not been used in studies focused on the two-year sector. Second, we acknowledge and incorporate information about students' geographic origins in our analyses predicting study abroad participation, accounting for possibly the most prevalent source of omitted variable bias in higher education research (González Canché, 2019a. ...
... Our study offers a multimodal narrative that comprehensively accounts for both student-level indicators and indicators associated with students' geographic origins (heretofore we refer to these latter indicators as place-based indicators) in exploring study abroad participation and resulting academic outcomes at a two-year technical college located in the U.S. Midwest. Inclusion of both individual and place-based indicators is an important contribution given the primacy of place and space in students' higher education prospects (González Canché, 2019aHillman, 2016;Turley, 2009). Although one might argue that place-based indicators are less important at the community college, given the community college mission to serve local communities, ignoring such indicators in research assumes that all students attending a given community college originate from similar places, providing them with access to similar resources in their places of origin, such as economic prosperity or average levels of educational attainment in their home communities. ...
This study contributes to a line of research that examines the relationship between study abroad participation and academic success among students attending two-year institutions. We offer a multimodal narrative that comprehensively accounts for both student-level characteristics and indicators associated with students’ geographic origins in exploring study abroad participation and resulting academic outcomes, namely completion of any credential, completion of an Associate’s degree, enrollment in a four-year institution, percentage of credits passed, and cumulative GPA, at a two-year technical college in the U.S. Midwest. Our results suggest that study abroad is consistently and positively associated with students’ academic outcomes. This study has important implications for how we think about both who is able to access study abroad in the two-year sector and study abroad’s role in promoting student success at these institutions
... Despite its potential, the use of mixedmethods has been an exception rather than a rule in network research within human geography over the last decades. Combining qualitative and quantitative network analysis in mixed-method designs is still an emerging field (Crossley, 2010;Crossley and Edwards, 2016;Domínguez and Hollstein, 2014;Edwards, 2010;Glückler et al., 2020;González Canché, 2019;Nooraie et al., 2018). Recent approaches range from network ethnographies (Berthod et al., 2017), cultural network approaches (Edelmann, 2018), and the use of participant observation in formal network analysis (Conti and Doreian, 2010) to the 'quantitization' of qualitative content (Williams and Shepherd, 2017). ...
Social network analysis has become a popular methodology in human geography. This article develops three propositions – connectivity, contextuality, and reflexivity – for relational analysis to overcome the dualism between universalist network science on the one hand and idiosyncratic network stories on the other. Building on a detailed meta-analysis of over 300 network studies published between 1990 and 2018, we offer a sympathetic critique of the actual use of network methods in human geography. To unleash further potential of network research, we propose several research strategies for more inclusive relational analysis, including informed pre-specification, triangulation, and communicative validation.
... Scholars have extensively studied the concept of centrality within the social network literature. It is a measure that accounts for an actor's influence within the larger network based on their connections to other actors (González Canché, 2019). In general, more central actors are perceived as being in more powerful positions than those located at the periphery (Marsden & Laumann, 1984). ...
Despite pledges to increase student body diversity, low-income students and Black, Latinx, and Native American students remain grossly underrepresented at elite colleges. To date, most studies exploring stratification in college enrollment patterns have looked at this phenomenon from the student or organizational perspective—examining either students’ college choice process or the enrollment management practices of elite colleges. However, in this study, I suggest that one unexplored but potentially powerful mechanism shaping college access and student body diversity is the network of feeder schools that are connected to highly selective colleges in the United States. Guided broadly by relational sociology and social network theory, this dissertation uncovers the ways that feeder school relationships shape the field of highly selective admissions. Using an exploratory, mixed methods design that combines social network analysis with a qualitative case study approach, the first phase of the study examines the overall network architecture of feeder ties between 3,200 secondary schools and 76 highly selective colleges and universities. Then, in the second phase, I conducted interviews with 44 admissions officers and 38 college counselors who are a part of this network to explore the relational dynamics between feeder schools and highly selective colleges, attending to their consequences for access to and diversity in higher education. I find that private, predominantly white, and affluent high schools have more and stronger feeder ties to elite colleges compared to their public school, majority-minority high school, and less resourced counterparts. The network analysis in phase one revealed that four of the top 10 most central high schools in the network are private elite boarding schools. Through interviews in the second phase of the study, I also discovered that certain features of schools—namely, their curricula, resources, status, and capacity for relationship-building—contribute to the strength of ties they have to highly selective colleges. Qualitative findings suggest that feeder relationships are driven by the mechanisms of homophily, reciprocity, and trust, and these connections are facilitated by the frequent interactions and interpersonal relationships between admissions officers and college counselors. Based on these findings, I conclude that selective college admissions is, in fact, a relational process that unfolds within webs of social relations. Both feeder schools and elite colleges adopt habits and routines which serve to reproduce these relationships over time, with broad implications for access to elite higher education. Even if feeder relationships no longer guarantee entry into America’s most elite institutions, they continue to provide substantial advantages to already-advantaged high schools. If elite colleges are committed to increasing access and diversity, then they must critically evaluate existing interorganizational arrangements and routinized practices that drive decision-making in highly selective admissions.
... Geographical bias in testing is informed by the frameworks of concentrated advantages or disadvantages based on individuals' location (Elijah, 1990;Jargowsky & Tursi, 2015;Pacione, 1997). In this framework, individuals' prospects for social/upward mobility are affected by their places of residence, based on the divergent, structural, and systematic forms of access (or lack of access) to resources associated with experiencing life in particular geolocated (or localized) contexts (González Canché, 2019a, 2019b. By extrapolating social/upward mobility with test-performance prospects, one can argue that to the extent that spatially concentrated circumstances, such as crime, poverty, and unemployment, continue to systematically affect students' performance on these tests, their use in admission decisions and financial-aid eligibility will continue to fuel social inequality and yield estimates that are biased as a function of ignoring these geographical effects (González Canché, 2019a, 2019b. ...
... In this framework, individuals' prospects for social/upward mobility are affected by their places of residence, based on the divergent, structural, and systematic forms of access (or lack of access) to resources associated with experiencing life in particular geolocated (or localized) contexts (González Canché, 2019a, 2019b. By extrapolating social/upward mobility with test-performance prospects, one can argue that to the extent that spatially concentrated circumstances, such as crime, poverty, and unemployment, continue to systematically affect students' performance on these tests, their use in admission decisions and financial-aid eligibility will continue to fuel social inequality and yield estimates that are biased as a function of ignoring these geographical effects (González Canché, 2019a, 2019b. ...
... The findings corroborated that geographically/spatially contextualized indicators (e.g., poverty, unemployment, education attainment levels, and family composition in students' zip code of residence) affect students' performance in SAT/ACT tests, over and above individual-level factors such as test-takers' SES, ethnicity, and gender. Thus, even if these tests were robust to individuallevel cultural, racial, and socioeconomic biases, the fact that students' attainment prospects in these standardized scores can be explained by their localized contexts constitutes a bias-generating problem (Bivand, Pebesma, & Gomez-Rubio, 2013;Cliff & Ord, 1972;González Canché, 2014a, 2017a, 2018b, 2019b. ...
Recent Federal Bureau of Investigation (FBI) criminal charges involving standardized testing have renewed public scrutiny over the tests’ role in reproducing inequities and wealth-based privileges. Although some continue to call for the eradication of standardized testing, this article presents a different perspective with a threefold purpose. First it provides a critical review of the existing literature on standardized testing including the newest development: geographical bias in testing or the systemic and systematic ways in which place-based resources influence test-takers’ performance. Second, it discusses the goals and consequences of the College Board’s recently unveiled “adversity score,” designed as a tool to “boost admission prospects” of students facing additional challenges in their neighborhoods and schools. Finally, it offers a conceptual and methodological framework and policy insights to shift the role of these tests as reproducers of inequalities to enhancers of equity and opportunity into what is termed true college access: enrollment with realistic prospects of success.
... Considering the current study's interest in testing whether policy implementation resulted in different outcomes between GMU and a synthetic counterfactual, random permutation tests are more appropriate to test findings (Good, 2000). control units is recorded (González Canché, 2019). This procedure enables the recreation of a sampling distribution where mean differences of a given sample size is recorded x number of times. ...
A growing number of postsecondary institutions in the United States
have removed standardized testing as a requirement for admission. Researchers, however, have suggested that these “test-optional” policies may not benefit underrepresented populations as intended, but instead serve as an additional revenue source for the institution. In this study, we utilize a synthetic control method to extend this research by considering whether a more nuanced “testflexible” policy, which allows qualifying students to decide whether to submit test scores instead of an institution-wide policy, influences student enrollment demographics at a public university.
Although proponents of test-optional and test-flexible policies argue that these policies can increase student diversity, in line with previous research, findings from this study suggest GMU’s test-flexible policy has had a statistically insignificant influence in increasing the racial and socioeconomic diversity of its student body.
Even complete removal of standardized admissions tests will not be sufficient to close the persistent gaps alone without the provision of adequate information, resources, and support for students to ensure their ability to thrive.
... The third purpose is to offer an empirical example using actual institutional data obtained from two community colleges in different regions of the country. This example will highlight the use of spatial econometrics Cliff & Ord, 1969, 1972Cressie, 2015) using network analysis principles (González Canché, 2018, 2019aGonzález Canché & Rios-Aguilar, 2015) to assess and address issues of social dependence while using official institutional data typically gathered by institutional researchers. Particular emphasis will be placed on depicting the steps and rationale followed to conduct the analyses so that institutional and higher education researchers can use these analytic techniques that are yet to be broadly implemented in education research. ...
... The third purpose is to offer an empirical example using actual institutional data obtained from two community colleges in different regions of the country. This example will highlight the use of spatial econometrics Cliff & Ord, 1969, 1972Cressie, 2015) using network analysis principles , 2019aGonzález Canché & Rios-Aguilar, 2015) to assess and address issues of social dependence while using official institutional data typically gathered by institutional researchers. Particular emphasis will be placed on depicting the steps and rationale followed to conduct the analyses so that institutional and higher education researchers can use these analytic techniques that are yet to be broadly implemented in education research. ...
A common and truly relevant question that many institutional researchers commonly ask is why are so many students dropping out of their declared major? This chapter guides institutional research professionals through the rationale and steps to identify and address one of such factors: outcome dependence, a pragmatically and timely relevant endeavor when considering the current emphasis on outcome‐based or performance‐based funding policies.
