پذیرش مقالات و شبکه های اجتماعی در جستجوی کار

We examine how Dartmouth College seniors use social networks to obtain their first jobs. We do this by analyzing self reports of networking and by examining the correlation in employment outcomes among randomly assigned freshman roommates and hallmates. We find that the use of social networks differs for men and women and for white and nonwhite students. Networking also differs greatly across career types. Students networking with fraternity and sorority members and alumni are the most likely to obtain high paying jobs. There is a strong connection between own employment outcomes and outcomes for randomly assigned freshmen hallmates.

The simplest models of career choice and job search assume that people choose a career from a large set of options using full information. In reality, it seems likely that people use social networks of friends, peers, parents, and teachers to obtain career advice and information on jobs (see Holzer, 1987). This paper uses data from a survey of Dartmouth College seniors to examine how careers chosen, jobs obtained, and salaries obtained correlate with the various forms of networking that were used in the job search process. Our first set of results is descriptive and indicates that the students perceive networking to be quite important in finding a job. Furthermore, the type of networking used differs by job type, gender, and race. Students who solicit help from current and alumni members of their fraternity or sorority are also the most likely to obtain high paying jobs. This is an interesting finding given that many lawsuits against single sex organizations focus on the notion that social clubs provide economic benefits for their members.1 The second set of results shows that employment outcomes for a student are correlated with mean outcomes for that student's randomly assigned freshmen hallmates and dormmates. We discuss the conditions under which these results can be interpreted as evidence of peer effects among students in a rooming group. If networking is important to the job search process, and if selective private universities provide better networking opportunities than competing schools, then the private schools may be able to charge a large price premium. This hypothesis is consistent with Dale and Krueger (1998) who find that students’ post-graduation earnings are positively related to the tuition at their undergraduate school.2 The value of networking may also explain students’ willingness to incur the costs of attending highly selective MBA programs. Recent research has supported the existence of peer effects among college roommates as in Sacerdote (2001), Zimmerman (2002), Kremer and Levy (2000), and Stinebrickner and Stinebrickner (2001). More broadly, there is a rich and growing literature that finds evidence for peer effects among elementary school students as in Hoxby (2000) and Schunk et al. (1987), and junior high and high school students as in Gaviria and Raphael (2001), and Evans et al. (1992).3 This paper extends the literature by focusing on employment outcomes for graduating college seniors and using survey data to examine students’ own perceptions of the importance of peer influences. The peer effects measured here cannot be interpreted as causal due to the identification problems discussed in Manski (1993). Specifically, the selection issues may be quite severe. For example, we find that students who receive job help and advice from other fraternity (sorority) members are much more likely to enter the high paying fields of investment banking and financial sales and trading. It is likely that such correlation does not come purely from the treatment effect of fraternity help on career choice. Students who join fraternities (or sororities) may be inherently more outgoing and socially able people. The prevalence of high paying jobs among such students may be a return to a skill or the result of underlying preferences which drive certain types of students to join fraternities and to search for high paying jobs. To control for the selection issue, we examine the correlation in outcomes among randomly assigned freshmen roommates, hallmates, and dormmates. The data indicate that there is a statistically and economically significant correlation in working status and salary among students who were housed together as freshmen. Because of the random assignment, we know that this correlation in outcomes does not arise from students self-selecting into rooming groups. However, due to the reflection problem as described by Manski (1993), we cannot interpret our coefficients directly as the actual size of the peer effect.4 The rest of the paper is organized as follows: Section 2 presents a description of the data. Section 3 presents results and Section 4 concludes.

The survey data explored in this paper demonstrate that students perceive networking with peers, alumni, faculty and relatives to be an important part of the job search process. Students entering different careers use different forms of networking. The data also show that men and women and white and nonwhite students use different networking strategies. There is a very strong connection between obtaining help from fraternity and sorority members and obtaining prestigious, high paying jobs. Some of this connection might be causal in the sense that a given student using fraternity help would do worse on employment if he were denied the opportunity to network with fraternity brothers and alumni. Identifying this causal effect would require a good instrument for fraternity participation and we do not have such an instrument. There is also a strong correlation between own employment outcomes and average outcomes for a student's randomly assigned freshmen hallmates and dormmates. This suggests that peer networks from friendships made freshman year may be influencing career choices or helping students obtain jobs. Overall the data provide good evidence that the job search process involves collecting information and receiving assistance from a variety of social networks including peer networks. The availability of networks may partially explain why admission to certain colleges and universities is considered so desirable. Hopefully future work will be able to further document causal links between the availability of such networks and employment outcomes and whether a lack of networks can explain poor outcomes for certain groups of people.