* Abstract

This article analyses a series of emails thanking Nigel for his stewardship of JASSS and the characteristics of their authors. It identifies a correlation between two measures of author activity in social simulation research, but no pattern between these activity measures and the email timing. Instead, the sequence suggests a classic standing ovation effect.

Keywords:
Gilbert Number, Social Contagion

* Introduction

1.1
On 4 November 2014, in response to the retirement of Professor Nigel Gilbert as editor of JASSS, Dr David Sallach emailed the SimSoc mailing list (SimSoc 2014) thanking Nigel for the 17 years of JASSS work. Over the following week, there was a series of further thanking emails, both to the SimSoc list and privately. This article analyses key indicators of sender involvement in the social simulation community to assess whether these provide greater insight than a simple bandwagon effect (Macy & Willer 2002).

* Methods

2.1
We extracted those emails thanking Nigel Gilbert for editing JASSS or some related activity from both the SimSoc mailing list archive (SimSoc 2014) and Gilbert's private email account (using administrative measures within the University of Surrey). Two types of messages were excluded: those in a thanking chain but did not explicitly thank Gilbert, and second or later emails by the same sender. We retained 36 SimSoc and 6 personal emails.

2.2
Each email was coded in four ways. These were timing relative to the initial message, country, Gilbert number, and JASSS publications. Time and country were intended as descriptive variables. Gilbert number and JASSS publications are indicators of engagement with the social simulation community, which might influence a person's decision to write an email.

2.3
Length of time after the initiating email was calculated from the sending time recorded by the SimSoc archive or message header metadata. The country in which the author currently works was identified with manual internet searching.

2.4
Inspired by the Erdős number (ERDOS NUMBER PROJECT 2014), the 'Gilbert number' for each email author represents the distance to Gilbert along the collaboration network. Initial values were provided by the Co-author Path tool of Microsoft Academic Search (Microsoft 2014). However, data quality is somewhat poor, with different authors combined into a single profile, or a single author appearing as separate profiles. Thus, individual links proposed by Co-author Path were traced and the proposed collaboration papers checked, with Gilbert numbers revised as required. In addition, distances greater than 4 and authors unable to be linked were assigned a Gilbert number of 4 as longer distances were more difficult to correct. Given Gilbert's long history in the social simulation community (such as establishing JASSS in 1998, an early President of ESSA, and co-organiser of one of the first relevant workshops in 1992) and extensive cited output (Google Scholar h index of 51), it is reasonable to expect him to hold a central position in the collaboration network. Hence, the Gilbert number serves as an indicator of involvement in the social simulation community.

2.5
Finally, the number of articles published in JASSS was extracted from the Web of Science (Thomson Reuters 2014) with a journal specific enquiry in combination with the email sender as author. This returns all material indexed by the Web of Science, such as articles and book reviews, and the material was manually counted. Number of JASSS publications is an indicator of direct involvement with JASSS, and hence involvement with Gilbert as editor of JASSS.

* Analysis

3.1
The first observation is the broad geographic distribution of the senders of 'thank you' messages (see Figure 1). It is very clear that the social simulation community is indeed global, with centres in the USA, South America, Europe and Australia.

Figure 1
Figure 1. Map of countries in which message senders are based. The map was built through http://edit.freemap.jp/en/trial_version/edit/world. .

3.2
Figure 2 displays the time sequence of emails and the Gilbert number of their senders. Most people have replied on the day after the announcement of Nigel Gilbert stepping down. The time series has been plotted by message counter (or cumulative number of messages) to show the social contagion process (Macy & Willer 2002), which appears as a sigmoidal curve.

Figure 16
Figure 2. Gilbert numbers by time and type of message. Gilbert numbers of people who sent emails thanking Nigel Gilbert either to the SimSoc list (blue squares) or privately (red dots). Time is truncated at 84 hours (3.5 days), which excludes one data point at 306 hours. The initial message was sent at 21:35 GMT on 4 November 2014, and vertical dotted lines indicate 0:00 GMT for each day.

3.3
Also in Figure 2, darker colours (either blue or red) indicate a shorter distance from the person to Nigel Gilbert in the coauthorship network. The darkest markers mean direct coauthorship (Gilbert number of 1). There are no clear coauthorship patterns in the thank you emails sequence. The first person to write a thank you has a large Gilbert Number. While the few immediately following messages are close coauthors, the later senders have a mixture of high and low Gilbert Numbers. This observation is supported by the negligible correlation between message counter and Gilbert number (ρ = 0.07).

3.4
Figure 3 uses a similar format, but reports the number of publications that message senders have had in JASSS. Darker shades indicate more publications. As for coauthorship, senders with large and small JASSS publication counts appear throughout the 'thank you' email sequence. However, there is a weak relationship, with early messages more likely to be sent by those with a higher number of publications (ρ = −0.27).

Figure 3
Figure 3. JASSS articles by time and type of message. Number of JASSS publications of people who sent emails thanking Nigel Gilbert either to the SimSoc list (blue squares) or privately (red dots). Time is truncated at 84 hours (3.5 days), which excludes one data point at 306 hours. The initial message was sent at 21:35 GMT on 4 November 2014, and vertical dotted lines indicate 0:00 GMT for each day.

* Discussion

4.1
The timing of the messages demonstrates a typical social contagion curve. There is a weak relationship between message order and number of articles published in JASSS by the sender. A natural interpretation is that those researchers most engaged with JASSS are the quickest to thank Gilbert for creating and maintaining JASSS.

4.2
The messages thanked Gilbert for several different aspects of JASSS editorship, including broader benefits to the social simulation community. One sentiment thanked Nigel Gilbert for running JASSS, including performing various mundane administrative tasks such as article formatting. A second recognised the role of both JASSS and Gilbert in increasing the impact and reach of social simulation. The outcome is demonstrated by the map of sender countries (Figure 1) and the even broader readership of JASSS (Figure 4).

4.3
Finally, Gilbert was thanked for establishing social simulation as a field and building the community. While the Gilbert number is not correlated with message order, there is a clear relationship with the number of JASSS articles (ρ = −0.51); those message senders with more JASSS publications also have a shorter coauthor path to Gilbert. As JASSS is the relevant specialist journal for social simulation, this relationship emphasises the central role of Gilbert in the community, supporting the recognition expressed in the messages.

Figure 4
Figure 4. JASSS reader map. Map of cities in which JASSS was accessed on 24 November 2014 (JASSS reader map).

4.4
Overall we conclude that the key driver in the thank you messages was the standing ovation social contagion effect rather than relying on a close relationship with Gilbert. It is a spontaneous eruption of gratitude to someone who is always slightly ahead of the game (promoting social simulation in its infancy and starting one of the very first open-access internet-only academic journals) and is able to build a community not by overt networking but by attracting people through producing excellent research, tireless collaboration on research and research proposals and enriching every discussion through immense focus, creativity and intellectual curiosity.

* References

ERDOS NUMBER PROJECT (2014). Hosted by Oakland University at https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=SIMSOC.

JASSS reader map obtained from http://jasss.soc.surrey.ac.uk/admin/rdr_map.html on 25 November 2014.

MACY, M. W. & Willer, R. (2002). From factors to actors: Computational sociology and agent-based modeling. Annual Review of Sociology 28, 143–166. [doi:10.1146/annurev.soc.28.110601.141117]

MICROSOFT (2014). Microsoft Academic Search, Co-author Path, available at http://academic.research.microsoft.com/VisualExplorer.

SIMSOC (2014) SimSoc mailing list managed by JISCM@il, available at https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=SIMSOC.

THOMSON REUTERS (2014). Web of Science, available (by subscription) at http://wokinfo.com.