These "classic" data were originally collected by Theodore Newcomb (1961) from 1953 to 1956 and consist of 15 matrices that record weekly sociometric preference rankings from 17 men attending the University of Michigan in the fall of 1956; data from week 9 are missing. They are also available in UCINET (Borgatti, Freeman, & Everett, 2006), a popular general social network software package reviewed in Chapter 12. A 1 indicates first friendship preference, a 16 represents the last preference, and no ties were allowed. The men were recruited to live in off-campus (fraternity) housing, rented for them as part of the Michigan Group Study Project supervised by Newcomb. All were incoming transfer students with no prior acquaintance with one another. Specifically, the following files derived from these data are available from this companion website (in the form of adjacency matrices):

1. Fraternity Data Chapter 5. This adjacency matrix is from the beginning of Newcomb's study (week 0) and consists of relational data that are directed and unweighted. Friendship rankings originally coded as 1, 2, or 3 became 1s, 0 otherwise. These are the same data reported in Table 5.1 (in node-list format) and Table 5.2 (matrix format).
2. Fraternity Data Chapter 6. This adjacency matrix, from the beginning of Newcomb's study (week 0), consists of relational data that are non-directed (symmetrical) and unweighted. A friendship tie exists if one student nominated another as either a friend or best friend. These are the same data that are presented in Table 6.1 and used for many of the "top-down" and "bottom-up" approaches to group analyses addressed in Chapter 6.

Collected from students in four middle school science classrooms taught by the same teacher at two points in the school year, these data (reported in Pittinsky & Carolan, 2008) provide opportunities to examine two interesting features of students' social networks. First, because network data were collected in the fall and spring, it provides an opportunity to examine how within-classroom friendship patterns change over time. Second, not only were students asked to nominate their friends, the teacher was also asked to rate who was friends with whom. This provides an opportunity to examine how well the teacher perceived friendship patterns among students and the degree to which this accuracy increased or decreased over time. Specifically, the following files derived from these data are available from this companion website (in the form of adjacency matrices):

1. Peer Groups Data Chapter 3_a. This adjacency matrix consists of student-reported friendship relations among 27 students in one class in the fall semester. These data are directed and unweighted; a friendship tie is present if the student reported that another was either a best friend or friend. These are the same data used to generate Figure 3.1 and reported in Table 3.3.
2. Peer Groups Data Chapter 3_b. This adjacency matrix also consists of student-reported friendship relations among 27 students in one class in the fall semester. However, the adjacency matrix contains directed and valued (weighted) data, with "best friends" coded as 1s and "friends" coded as 2s. These are the same data used to generate Figure 3.4.
3. Peer Groups Data Chapter 3_c. This adjacency matrix consists of the teacher-reported friendship relations among 27 students in one class in the fall semester. These data are directed and unweighted; a friendship tie is present if the student reported that another was either a best friend or friend. Used with the data file, Peer Groups Chapter 3_a, these data can be used to investigate cognitive social structure (CSS). A subset of these same data is reported in Table 3.4.

The third data set that will be used throughout this text is Daly's Network of School District Leaders. Leadership network data were collected at two school districts over 3 consecutive years. For each consecutive year, school district leaders were invited to complete a survey that collected individual demographic information (e.g., gender, ethnicity, marital status, age, years of experiences), 11 different network relationships (e.g., collaboration, confidential, energy, expertise, support you approach, support approach you, work-related issues, input, recognition, best practices, and innovation), efficacy, and trusting relationships. The social network questions asked the participants to assess the frequency of interactions they have with those nominated individuals on a four-point frequency scale ranging from 1 (the least frequent) to 4 (1–2 times a week). The efficacy items were designed based on the Principal Efficacy Scale used in Daly et al. (2011) and Tschannen-Moran and Gareis's (2004) studies. The efficacy scale includes 18 items rated on a 9-point Likert scale ranging from 1 (None at all) to 9 (A great deal). The trust scale contains eight items rated on a 7-point Likert scale ranging from 1 (Strongly disagree) to 7 (Strongly agree) modified from Tschannen-Moran and Hoy (2003). Specifically, the following files derived from these data are available from this companion website in the form of 1) adjacency matrices, or 2) for attribute data, rectangular matrices:

1. School Leaders Data Chapter 7. This non-directed and unweighted adjacency matrix reports on whether two school leaders turn to each other for "information about work-related topics." Collected as part of a complete network study, these same data were used to describe and define the ego-level concepts and measures discussed in Chapter 7. The data have been dichotomized (unweighted) so that only those ties that were originally coded a 3 or 4 are now 1s, indicating that there is a relation between ego and alter. Any tie that was originally absent or coded 1 or 2 is now 0, indicating that a tie does not exist. Furthermore, the relations were symmetrized (non-directed) using a weak criterion: If one school leader reported a tie (1), then the tie was considered present.
2. School Leaders Data Chapter 9_a. This adjacency matrix reports on "collaboration" ties among 43 school leaders in year 1 of a three-year study. This is a directed valued (weighted) network measured on five-point scale ranging from 0 to 4, with higher values indicating more frequent collaborations (1–2 times/week). These data are used throughout Chapter 9.
3. School Leaders Data Chapter 9_b. This adjacency matrix reports on "collaboration" ties among 43 school leaders in year 3 of a three-year study. This is a directed valued (weighted) network measured on five-point scale ranging from 0 to 4, with higher values indicating more frequent collaborations (1–2 times/week). These data are used throughout Chapter 9
4. School Leaders Data Chapter 9_c. This adjacency matrix reports on "confidential help" ties among 43 school leaders in year 1 of a three-year study. This is a directed valued (weighted) network measured on five-point scale ranging from 0 to 4, with higher values indicating more frequent collaborations (1–2 times/week). These data are used throughout Chapter 9.
5. School Leaders Data Chapter 9_d. This adjacency matrix reports on "confidential help" ties among 43 school leaders in year 3 of a three-year study. This is a directed valued (weighted) network measured on five-point scale ranging from 0 to 4, with higher values indicating more frequent collaborations (1–2 times/week). These data are used throughout Chapter 9.
6. School Leaders Data Chapter 9_e. This rectangular matrix consists of four attribute vectors for 43 school leaders. Following the first ID column, the matrix includes an efficacy score, trust score, and indicators for whether one works at the district-level and is male (1 = yes, 0 = no). These attribute variables can be used as covariates in some of the statistical models covered in Chapter 9, including regression and p* models.