Microsoft Word - Volume37_final insna.org | Issues 1&2 | Volume 37 | 45 A Visual Data Collection Method: German Local Parties and Associations Dr. Isabelle Borucki (Corresponding author) University of Trier Trier, Germany Abstract This research captures local networks of German political parties and welfare agencies in re- gards to poverty. The article explores whether there are differences in regards to homophily and brokerage between the two studied groups using a dataset of 33 egonetworks in two Ger- man cities. The computer assisted drawn networks were collected in an interactive participa- tive way together with the interviewed egonetworks. To achieve the theoretical aim of analys- ing homophily and brokerage between politicians and welfare workers, two hypotheses are examined, resting upon social capital theory. The hypotheses were quantified and explicated with different variables. The first hypothesis states that heterophile networks imply more so- cial capital, which referred to different measurements (size, density, homophily). This could be partially validated since the analysed networks of association representatives (n=12) were denser and slightly more heterophile than those of party representatives (n=21). Second, it was assumed that politicians, because of their function as elected representatives, would be more likely to take on an interface function within the communities than representatives of civil society institutions. Results based on calculated EI-indices, subgraphs and brokerage show that party representatives do indeed have larger networks, but these networks split into fewer subgraphs than association representatives’ networks. Author Isabelle Borucki currently is leading “DIPART. Digital Party Research” – a junior research group on digital party research, located at the NRW School of Governance, University of Duisburg-Essen. Before that she worked at the Department of Political Science, University of Trier. Isabelle does research in Political Organizations and Parties, Comparative Politics, and Information Technology/Digitalization and Politics. Acknowledgements This research was funded by the German Research Foundation (DFG) within the project “Po- litical Representation of Poverty by Political Parties in Germany” of the Collaborative Re- search Centre (CRC) 600 at Trier University, Germany. The author wishes to thank Dimitris Christopoulos for his recommendations and Markus Gamper, Linda Reschke and Peter Starke for their remarks on earlier versions of the paper. Connections A Visual Data Collection Method 46 | Volume 37 | Issues 1&2 | insna.org 1. Introduction Poverty and social exclusion are pressing social issues in advanced and industrial- ized countries, such as Germany. In Ger- many, these issues are becoming more vis- ible through a public discussion of the German social security reform program ‘Hartz IV’ or the poverty of children. For German local communities, the fields of poverty prevention and fighting against al- ready existing poverty structures pose a va- riety of challenges. Officially, local institu- tions are responsible for a majority of so- cial benefits and their distribution. The in- terplay between different political actors and actors from civil society requires net- working, cooperation and coming to agreements at various levels. It is exactly this process, and the accompanying rela- tions and connections, which make munic- ipal efforts to fight poverty in Germany an interesting field for political science; a field which has hardly been explored in terms of networking structures. This is surprising, because municipalities are es- pecially restricted in their actions and ca- pabilities and depend on the cooperation and involvement from the civil sector, which is a challenge specifically visible in the area of fighting poverty. The thematic aim of this paper is to capture the local network of political par- ties and welfare agencies regarding pov- erty, and to explore whether there are dif- ferences concerning homophily and bro- kerage between politicians and welfare workers. The research question is: How exactly are local political actors connected with organisations from civil society? The methodological goal is to show that social network analysis is well suited to statistically represent the relations be- tween local politicians and welfare agen- cies. Additionally, the findings demon- strate that visual data collection allows for a simultaneous validation of data through the participation of the subject group. As a result, it can be illustrated how a visual in- quiry through digital networking maps en- ables a collection of quantitative data that can then be evaluated (Gamper et al., 2012). Lastly, it will be assessed how use- ful the method of visually surveying quan- titative data can be for the research field of local politics. Here, the question is: which advantages and disadvantages emerged for participants during the survey because of structured and standardized digital network maps? 2. State of Research and Theoretical Framework Some studies exist that take a closer look at the state of the German local level and its existing structures of decision-making. However, social network analysis has not been significantly featured in these works (Heinze & Voelzkow, 1991; Helbling, Egli, & Matter, 2005; Fowler, 2006; Ohm, 2009; Werner, 1998). Instead, the intent of this research is to offer a comparative look at the structures of fighting poverty at a lo- cal level by analysing the networks be- tween local politicians, welfare agencies and domestic associations. One exception to achieve this is the work of Sören Peter- mann (2008). He has studied the political influence of local politicians in the context of a social capital model in major cities, medium-sized cities and counties (Peter- mann, 2008, p. 156). He wanted to show how the social capital (Burt, 1995; Cole- man, 1988) of established politicians in the municipality (mayor, county commission- er, parliamentary party leaders) affects their political influence within the local power structures. Here, he measured the centrality of the actors and their broker po- sition in egonetworks (Petermann, 2008, p. 153, 155). Moreover, he calculated regres- sion models to estimate the social capital of politicians within interaction networks and found out that finding consensus and bargaining in networks depend on politi- cians’ prestige in the cities (ibid., p. 171). In methodological concerns, he showed that local top politicians are highly con- nected within their community, but admits that his studied networks focused on strong ties and interactions within cities. Howev- A Visual Data Collection Method Connections insna.org | Issues 1&2 | Volume 37 | 47 er, no studies exist that explicitly focus on the field of poverty at the local level in Germany. The study introduced here is meant to, at least partially, fill this academ- ic void. This study draws from Bourdieu’s social capital theory (Bourdieu, 1986; Coleman 1988; Lin, Cook, & Burt, 2001), because social capital enables inclusion in social networks, and entering social rela- tionships based on access to material and immaterial resources and support from other people (Crossley et al. 2015, p. 26– 37). At this point, it is assumed that social capital can be an individual, as well as a collective good. In institutionalized rela- tionships, resources can be tapped, which can prove beneficial for the individual, as well as for the collective. Here, it is helpful to understand networks as fields, in which social capital is distributed and exchanged (Bourdieu, 1985). Because one group dis- tinguishes itself from its outside, the pre- sent article is based on the assumption that, according to reciprocal recognition and re- inforcement, similar and homogenous ac- tors are to be found in local networks (Bourdieu, 1986). Consequently, the study is based on the following premises: The in- terrelationship between local party activ- ists, city administration and welfare agency officials serves as the exchange and foster- ing of one’s social capital. Depending on how much social capital an ego can accu- mulate, meaning how many potential communicative relations it has, it is more likely to be able to articulate its interests and act inclusively at the political level. Another, more special, type of capital is in- formation, because it is passed on through weak ties (Granovetter, 1973) within, and between, networks. This research is guided by asking about the cooperation of political parties, welfare agencies and the administration: How are these networks structured? What kind of different or similar actors can be found in these networks, how many of them are there and what is their function? Two hypotheses serve as guidelines: First: If actors from different areas are represented in a network, heterophile networks would contain more social capi- tal. This is based on the assumption that welfare agencies and domestic associations represent poverty-stricken people, because of their institutional foundation, and are, therefore, potential contacts for politicians. Consequently, their networks should be bigger and presumably denser and more heterophilic. Networks can be considered homophilic if alteri (network contacts and nodes), which resemble the ego and its fea- tures, are placed in the areas where the people questioned are supposed to locate their contacts (Lin, Cook, & Burt, 2001; Pfenning & Pfenning, 1987, p. 72). These three characteristics of networks (size, density, and homophily) will be calculated individually in this case. Second: Due to their function and position as representatives of the people, officials of political parties have higher brokerage values than representatives of welfare and domestic agencies. A person, who is located in a defining interface of the network, is called a broker. Brokerage is comparatively operationalized via the number of subgraphs in each egonetwork (Burt, 1995). The more subgraphs that are in an egonetwork, the higher the ego bro- kerage value. 3. Methodology For this study, egonetworks were collected and analysed (Fischer et al., 1977; Well- man & Leighton, 1978; Crossley et al. 2015). Because egonetworks explain rela- tions between the ego and its alteri, they reflect the individual “bounded rationality” of actors. The qualitative approach to net- works that gained attention recently (Bel- lotti 2008; Hollstein 2011) proved to be the best way to access the discussed field of research. Benefitting from qualitative SNA, a mixed methods design was real- ized based on visual network maps with targets (Crossley et al. 2015, p. 60). In this case, the program VennMaker was used to collect data visually through network maps Connections A Visual Data Collection Method 48 | Volume 37 | Issues 1&2 | insna.org (Kahn & Antonucci, 1980). This software allows an immediate visualization of the participant’s network during the conversa- tion. Additionally, network structures can be immediately validated by repeated and clarified questions during the interview process. The emphasis, therefore, lies on the qualitative aspects of respondents’ networks. Complementary, or, depending on the question, primarily, quantitative da- ta can be calculated by standardizing and structuring the network maps prior to the interview, so that results can be compared. Overall, the use of VennMaker is a stand- ardized collection, in which alteri attrib- utes, for example importance, type of rela- tionship, etc., are visualized through the program and are quantitatively recorded and standardized. This visual collection method combines qualitative and quantita- tive approaches and integrates both (Hollstein 2014, p. 10–1). The elicitation for this research was embedded in “tradi- tional”, in-depth expert interviews (Bogner, Littig, & Menz, 2009). Regarding the sampling parties, welfare agencies and domestic associations at the local level in an East German city (Jena) and a city in West Germany (Trier) were chosen. Both cities are home to a university, are of similar size, are of a similar socio- economic structure, and have to tackle similar challenges in areas of deprivation, which is what makes them participants of the federal “Social City Program”. Accordingly, a systematic prob- ability method was chosen to approach representative position holders in the field: the chairmen or executive directors of as- sociations and charities were interviewed and their political counterparts, the party leaders, faction leaders, or the politicians who had social policies as their main fo- cus. By then using the snowball-approach (Babbie, 2013, p. 191), other central actors were included in the course of the process. The sample includes 36 egonetworks in to- tal. For Jena, the sample includes six asso- ciation representatives, 12 party represent- atives and one city administrator. For Trier, five association representatives are included, as well as nine party representa- tives and two city administrators. Alto- gether, 19 women and 17 men between 31 and 72 years were represented. For four networks, not all data is available, and for three networks, interview effects can be assumed, which were conducted by anoth- er interviewer. The networks of admin- istration officials were not included, be- cause they represent a functional group of only three egos, and thus too small for an intended mean value comparison. The egonetworks were established together with the respondents using the “free net- work drawing” function of VennMaker based on targets. For the networks, alteri were collected through a name generator, in order to functionally separate the field (Burt, 1997; Campbell & Lee, 1991). For this analysis, the question asked focused explicitly on relationships of professional information exchange: “Now I am curious about the people with whom you work to- gether in the field of fighting poverty. If you could give me a specific example, for instance, who do you contact or ask for help when you administer benefits?” The generator was connected to qualitative guiding questions of the expert interview with the intention to reveal rele- vant cooperation: Most of the time, partic- ipants named organisations important to them in the course of the qualitative expert interview. These organisations were noted, and the participant would be asked, with which individuals from these organisations she worked together with in the field of welfare. The named alteri were then drawn in VennMaker, together with the interview partner, in a visually-participatory manner. Afterwards, the name interpreters were queried, which generated standardized in- formation regarding the alteri and their re- lations to ego. The nature of the relation- ship between the ego and its alteri is col- lected through the interpreters, as well as alter-alter relationships. Interpreters, com- prised of several common measures, such as the duration of contact, frequency of contact and the type of contact, the im- portance of alteri for the ego (visualized as A Visual Data Collection Method Connections insna.org | Issues 1&2 | Volume 37 | 49 the size of alteri), function, party member- ship, age and gender, were used. Interpret- ers were operationalized as follows: con- tact frequency: 1 = very often (daily, week- ly), 2 = often (up to three months), 3 = sel- dom (once or twice a year); duration of contact: 1 = 1–2 years, 2 = 2–5 years, 3 = 5–10 years, 4 = 10–20 years, 5 = more than 20 years; type of contact (personal, via phone, via mail, email, cell phone). The visual collection and participa- tory positioning together with the partici- pants proved to be very beneficial for both sides. At no point in time did the partici- pants feel as if they were producing the type of quantitative data, which is usually the case with questionnaires when collect- ing network data. Plus, this collection pro- cedure is not as time consuming as the classic approach. This method is, therefore, especially interesting for smaller and more sensitive research fields, and leads to equally valid results. In the course of the processing of the digital network map, it became appar- ent that categorical variables, such as fre- quency of contact, age or duration of con- tact, can be collected very well through the additionally configurable ‘actors chart’ in VennMaker, which is comparable to a questionnaire, meaning these were not added visually, but via a catalogue of ques- tions that can be called upon in the pro- gram. Instead, the form and formalization of cooperation was represented through re- lations, and afterwards the drawings were complemented with this information to- gether with participants. Here, varying forms of relations were offered to depict multiplexity, from which the participant could select the most fitting for the present relationship in their perception (see Figure 1). Figure 1: Visualization example of the operational- ization Source: Collection, calculation and figure by the author To distinguish those multiplex rela- tionships, several forms for collaboration were drawn with one alter. More specifi- cally, these were the exchange of infor- mation and experiences (turquoise), ex- change or procurement of means (finan- cially and otherwise; blue), planning and implementation of concrete measures (pur- ple) and timely loose or sporadic (orange), as well as institutionalized and formalized cooperation (green). The alteris’ colours represent party membership in line with parties’ colours or non-partisanship (white). Via three concentric circles, the accessibility of alteri was classified from “very good” to “less good” (see Figure 2). The sectors, meaning the circle segments in different grey shades, illustrate the areas of the municipality: political parties, city administration, agencies, charities and as- sociations, businesses, unions, media. For the quantitative analysis, the collected re- sults were controlled and processed in Ex- cel through the export function of the pro- gram. Afterwards, the network parameters and measures for homophily and brokerage were calculated by using UCINET; here, the calculation of density was balanced with the one from VennMaker. As a third factor, apart from size and density, ho- mophily was calculated using the EI-index. This index carries out values from −1 to 1; Connections A Visual Data Collection Method 50 | Volume 37 | Issues 1&2 | insna.org 1 meaning heterophily and −1 standing for homophily: the lower the value, the more homophilic the network. In order to com- pare the networks amongst each other, in terms of their homophily, this research takes the indices of the entire network into account. Finally, the comparison of mean values was calculated with a parametric test in SPSS. 4. Empirical Findings When looking at the network size, politi- cians have a slightly bigger network than the association representatives. The larger the network, the lower the density (Borgat- ti & Everett, 1997). For both groups, dif- ferences could be found in density rather than in size. A comparison of the networks of party representatives regarding the EI- index reveals that the networks show the entire spectrum of EI. For the mean value, all networks are located in the homophile field of the index, yet the networks of party representatives are more heterophilic at −0.122 (SD = 0.588) than those of the as- sociation representatives at −0.245 (SD = 0.696) (see Table 1). Thus, politicians have more heterophilic networks in terms of lo- cating alteri in sectors. The mean values show that the association networks display less varied values than the ones of the poli- ticians. Therefore, politicians act as infor- mation intermediaries according to the second hypothesis, which stated politicians enjoy a kind of monopoly regarding the passing of information, which is operation- alized through the number of subgraphs. Table 1: Comparison of the mean value for size, density, EI-index and subgraphs Collection, calculation and figure by the author.15 Significance level at 0.05. The difference in the amount of subgraphs from 21.38 (parties) to 23 on average (as- sociations) indicates that association repre- sentatives seem to have the higher broker- age power, and hypothesis two is rejected. The t-test was only relevant for the density. Regarding the number of subgraphs and function, the mean value comparison was not significant (see Table 1). Since the dif- ference of the mean values for density is accidental by up to 20 per cent, a very low effect between density and function may 15 The purely star-networks (no. 9, 10, 31, 34, 37) were excluded here. exist. As a result, one can state that the networks of the association representatives contribute more to the inclusion of affected interests than those of the party representa- tives. Contrary to this, the politicians’ net- works seem more secluded, due to their slight homogenous constitution. 5. Conclusion The thematic aim of this contribution was to illustrate the network structure between local politicians and welfare associations in poverty politics. The networks of stud- ied egos were quantified and explicated with different variables (size, density, ho- mophily, and subgraphs). The first hypoth- esis could be partially validated, because the networks of association representatives were denser and slightly more heterophilic Function Size Density EI-index via sector affiliation Number of subgraphs party (n=21) mean 24.950 0.12870 −0.122520 21.380 SD 9.897 0.12527 0.588792 13.261 association (n=12) mean 23.670 0.25010 −0.245500 23.000 SD 12.272 0.38182 0.696312 13.423 total (n=33) mean 24.480 0.17280 −0.167240 21.970 SD 10.648 0.25187 0.622052 13.133 significance (n=26) t 0.599 0.08000 0.775000 0.489 squared eta (n=26) ŋ2 0.012 0.00400 0.001000 0.210 A Visual Data Collection Method Connections insna.org | Issues 1&2 | Volume 37 | 51 than those of party representatives. Re- garding the second hypothesis, results show that party representatives have larger networks, yet these networks split into fewer subgraphs than the networks of as- sociation representatives. In terms of methodology, the aim was to show how digital network maps fa- cilitate collecting and analysing quantita- tive data. For this, data collection with VennMaker has proven to be very effec- tive. Structures and the manifestations of relations with the associations, as well as alteri-characteristics were standardized; here, the scaling and classification of cate- gories was very time-consuming. The posi- tioning of the alteri could be singled out from VennMaker for the quantitative cal- culations presented here. Insofar, it was as- sured that the shown network structure was also depicted in the quantitative data set. Sketching the networks together with the interview partners allowed a direct captur- ing of the inherent thought processes of the participants during the conversation. While this may have led, in some cases, to very confusing network maps, the fact that par- ticipants could see their networks structur- ally visualized proved to be interesting and surprising to them. This also prevented boredom or frustration from emerging dur- ing the collection of alter-alter-relations (McCarty, Killworth, & Rennell, 2007). The joint sketching caused participants to become more interested in the inquiry, and it was possible to illustrate the structure of their professional networks for them. Such a combined approach is currently without equal because a simultaneous positioning of the alteri and validation of network structures on a visual network map is a clear advantage of the program, whereas the analysis for statistical calculations had to be done with other programs. Conse- quently, VennMaker had an interface func- tion for the quantitative part of the study. The collection and processing of data was enabled and highly facilitated, in particu- lar, because of the program. Compared to this, the processing and calculating of pa- rameters, especially the EI-index and the ttest, was elaborate and tedious in UCI- NET. The data collected and exported with VennMaker was further analysed, or man- ually edited in other programs. Here, SPSS was fitting for the hypotheses tests. Initial- ly, VennMaker caused the most problems in the beginning of the analysis compared to the other programs. This was mainly be- cause the software was first used as a be- taversion in the research project and later on in the 1.0-version. These two versions already differed greatly in their usage, es- pecially the statistics functions, which were needed for a smooth export and fur- ther analysis: they were in need of further development. For the collection of network data, VennMaker has proven successful in an interview setting. As a result, the meth- od of using digital network cards is well suited for the visual-participatory collec- tion of egonetworks together with partici- pants. The possibility of sketching the network and the following revision with the participants should be emphasized as a positive feature. By proceeding this way, the interpretations of the interview partner could be directly included and analysed quantitatively afterwards. Because of that, a distortion by the researcher can be mini- mized, and, almost in passing, quantitative data can be collected. It could, therefore, be demonstrated that visual collection in- struments are well suited to gather and evaluate quantitative data. References Babbie, E. R. (2013). The practice of social research (13th ed.). Belmont, CA: Wadsworth Cengage Learning. Bellotti, E. (2008). What are friends for?: Elective communities of single people. Social Networks, 30(4), 318–329. doi:10.1016/j.socnet.2008.07.001 Bogner, A., Littig, B., & Menz, W. (Eds.). (2009). Re- search methods series. Interviewing experts. Ba- singstoke [England], New York: Palgrave Macmil- lan. Borgatti, S. P., & Everett, M. G. (1997). Network analysis of 2-mode data. Social Networks, 19(3), 243–269. doi:10.1016/S0378-8733(96)00301-2 Bourdieu, P. (1985). The genesis of the concepts of habitus and field. Sociocriticism, 2(2), 11–24. Bourdieu, P. (1986). The forms of capital. In J. G. Richardson (Ed.), Handbook of theory and re- Connections A Visual Data Collection Method 52 | Volume 37 | Issues 1&2 | insna.org search for the sociology of education (pp. 241– 258). New York: Greenwood Press. Burt, R. S. (1995). Structural holes: The social struc- ture of competition (2nd ed.). Cambridge, Mass.: Harvard University Press. Burt, R. S. (1997). A note on social capital and net- work content. Social Networks, 19(4), 355–373. doi:10.1016/S0378-8733(97)00003-8 Campbell, K. E., & Lee, B. A. (1991). Name genera- tors in surveys of personal networks. Social Net- works, 13(3), 203–221. doi:10.1016/0378- 8733(91)90006-F Coleman, J. S. (1988). Social capital in the creation of human capital. The American Journal of Sociolo- gy, 94, 95–120. Crossley, N., Bellotti, E., Edwards, G., Everett, M. G., Koskinen, J. & Tranmer, M. (2015). Social net- work analysis for ego-nets. Los Angeles: SAGE. Fischer, C. S., Stueve, C., Jones, L. M., Jackson, R. M., Gerson, K., & Baldassare, M. (1977). Net- works and places: Social relations in the urban setting. New York: Free Press. Gamper, M., Schönhuth, M., & Kronenwett, M. (2012). Bringing qualitative and quantitative data together – Collecting and analyzing network data with the help of the software tool VennMaker. In H. Safar, M. Safar, & K. A. Mahdi (Eds.), Social Networking and Community Behavior Modeling: Qualitative and Quantitative Measures. Qualita- tive and quantitative measures. Hershey: Infor- mation Science Reference. Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(4), 1360–1380. Heinze, R., & Voelzkow, H. (1991). Kommunalpolitik und Verbände: Inszenierter Korporatismus auf lokaler und regionaler Ebene? In H. Heinelt & H. Wollmann (Eds.), Stadtforschung aktuell: Vol. 31. Brennpunkt Stadt. Stadtpolitik und lokale Politikforschung in den 80er und 90er Jahren (pp. 187–206). Basel: Birkhäuser. Helbling, M., Egli, S., & Matter, S. (2005). Lokale Eliten und kommunale Politiknetzwerke - Einflussreiche Akteure in der Einbürgerungspolitik einer Schweizer Gemeinde. In U. Serdült (Ed.): Vol. Nr. 3. Zürcher Politik- & Evaluationsstudien, Anwendungen sozialer Netzwerkanalyse. Beiträge zur Tagung vom 14. und 15. Oktober 2004 (pp. 105–118). Zürich: Institut für Politikwissenschaft. Forschungsbereich Policy-Analyse und Evaluation. Hollstein, B. (2011). Qualitative Approaches. In P. J. Carrington & J. Scott (Eds.), The SAGE handbook of social network analysis (pp. 404–416). London: SAGE. Hollstein, B. (2014). Mixed Methods Social Networks Research: An Introduction. In S. Domínguez & B. Hollstein (Eds.), Structural analysis in the social sciences: Vol. 36. Mixed methods social networks research. Design and applications (pp. 8–34). New York: Cambridge Univ. Press. Fowler. James H. (2006). Connecting the Congress: A Study of Cosponsorship Networks. Political Anal- ysis, (14), 456–487. Retrieved from 10.1093/pan/mpl002 Kahn, R. L., & Antonucci, T. C. (1980). Convoys of life course: Attachment, roles, and social support. In P. B. Baltes & O. G. Brim (Eds.), Life-span de- velopment and behavior (pp. 253–286). New York: Academic Press. Lin, N., Cook, K. S., & Burt, R. S. (2001). Social cap- ital: Theory and research. New York: Aldine de Gruyter. Retrieved from http://www.worldcat.org/oclc/45320849 McCarty, C., Killworth, P. D., & Rennell, J. (2007). Impact of methods for reducing respondent burden on personal network structural measures. Social Networks, 29(2), 300–315. doi:10.1016/j.socnet.2006.12.005 Ohm, A. (2009). Die Machtstruktur kommunaler Entscheidungsträger - Eine Netzwerkanalyse. In V. Schneider & F. Janning (Eds.), Politiknetzwerke. Modelle, Anwendungen und Visualisierungen (pp. 285–303). Wiesbaden: VS Verlag für Sozialwissenschaften. Petermann, S. (2008). Soziale Netzwerke und politischer Einfluss von Kommunalpolitikern. In W. Matiaske & G. Grözinger (Eds.), Sozialkapital. Eine (un)bequeme Kategorie (pp. 139–177). Marburg: Metropolis-Verlag. Pfenning, U., & Pfenning, A. (1987). Egozentrierte Netzwerke: Verschiedene Instrumente - verschiedene Ergebnisse. ZUMA Nachrichten, (21), 64–77. Retrieved from http://www.ssoar.info/ssoar/files/2011/393/zuma- nachrichten_1987_21_64-77.pdf Wellman, B., & Leighton, B. (1978). Networks, neighborhoods and communities: Approaches to the study of the community question. Toronto: Centre for Urban and Community Studies and the Dept. of Sociology, University of Toronto. Werner, W. (1998). Armut und Obdachlosigkeit in der Kommune. In H. Wollmann & R. Roth (Eds.), Kommunalpolitik – Politisches Handeln in der Gemeinde (2nd ed., pp. 703–716). Opladen: Westdeutscher Verlag.