Coding together – coding alone: the role of trust in collaborative programming

ABSTRACT

In the digital economy, innovation processes increasingly rely on highly specialised know-how and open-source software shared on digital platforms on collaborative programming. The information that feeds into the content on these platforms is provided voluntarily by a vast crowd of knowledgeable users from all over the world. In contributing to the platforms, users invest their time and share knowledge with strangers to add to the rising body of digital knowledge. This requires an open mindset and trust. In this study, we argue that such a mindset is not just an individual asset but determined by the local communities the users are embedded in. We, therefore, hypothesise that places with higher levels of trust should contribute more to Stack Overflow, the world's largest question-and-answer platform for programming questions. In relating the city-level contributions of 266 OECD metropolitan areas to infrastructure, economic, and trust measures, we find support for this hypothesis. In contrast, click rates to the platform are solely driven by infrastructure and economic variables, but not by trust. These findings highlight the importance of societal values in the twenty-first-century knowledge economy: if policymakers want to develop a lively local digital economy, it is not enough to provide fast Internet access and business opportunities. Instead, it is equally important to establish a trust-building environment that fosters sharing of innovative ideas, collaborations, and knowledge spillovers.

Code and Data: www.github.com/Braesemann/CodingTogether

KEYWORDS:

• Digital knowledge economy
• Social data science
• platform innovation
• collaborative programming
• stack overflow
• trust

Acknowledgments

This work benefited from comments made at the Complexity Science Hub Vienna Winter School 2019 in Obergurgl, and remarks from participants of the Innovation and Entrepreneurship Group Meeting at the Alexander von Humboldt Institute for Internet and Society in March 2019 in Berlin.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In this study, we do not claim that platform contributions are innovations, in the sense that these activities were focused on marketable outcomes, but they are a form of digital knowledge creation. We relate platform contributions to the patent-based innovation literature, as it provides a guideline on how the geography of such activities can empirically be modelled.

2 At the same time, a society's level of generalised trust is strongly related to societal settings, such as inequality (Stephany, 2017) or educational attainment (Stephany, 2019).

3 We examine patent applications as one of the most commonly applied metrics in economic geography for measuring local innovation activities. While we highlight how the online platforms have fundamentally changed the innovation process, we still perceive ‘traditional’ patent applications as a valid measure of innovation in ‘20th century’ economic geography.

4 For example, Wikipedia editors can gain awards for editing; on GitHub, a user's contribution history is publicly displayed on the profile page; and on Stack Overflow, users 'earn' reputation points for contributions (Bosu et al., 2013).

5 The level of generalised trust is measured by the share of individuals, who agree to the question 'Generally speaking, do you think that most people can be trusted (or do you think that you can never be to careful in dealing with people)?' The answers to this question ranges on a 10-item scale from 0 (no trust) to 10 (complete trust). The WVS, used in our analyses, aggregated the data to the regional share of people who answered with five or higher, as described in Table 1.

6 All Stack Overflow data are publicly available: https://archive.org/details/stackexchange. Details are described in (Braesemann et al., 2019).

7 https://www.quantcast.com/stackoverflow.com.

8 World Bank: https://data.worldbank.org/

9 OECD metropolitan area dataset: https://stats.oecd.org/Index.aspx?Datasetcode=CITIES

10 http://www.worldvaluessurvey.org/wvs.jsp.

11 Affinity describes how likely an Internet user in a given city is to visit Stack Overflow, relative to the global average of all Internet users. For example, in a world with only two cities, A (1,000,000 Internet users and 1000 Stack Overflow visitors) and B (5,000,000 Internet users and 2000 Stack Overflow visitors), city A would have an affinity value of 2.0 and city B a value of 0.8. Thereby, affinity does not depend on the size of a city's Internet population: it captures differences in Internet infrastructure and general affinity of the population to use the Internet; in other words the variable is used in this study as a normalised measure of the topical affinity to or specialisation in programming content on the Internet. This variable should, on its own, capture a lot of the variation of Stack Overflow contributions and clicks and helps us to control for some of the unobserved heterogeneity that is due to differences in the use of the Internet.

Source: https://www.quantcast.com/measure/stackoverflow.com?country=US#/generalInterestsCard

12 $y=\log (x+\sqrt{x^2+1})$. This transformation is an alternative to the more commonly used logarithmic transformation to handle extreme values but has the advantage of being defined for zero-values, see Burbidge et al. (1988).

13 Furthermore, the interpretation of the other parameter estimates is more straightforward, as they are on a logarithmic scale. For example, GDP's effect on contribution would be as follows: on average, a 1% increase in city-GDP would be associated with a 1.89% increase in city-level Stack Overflow contributions, holding everything else constant.

14 The conceptualisation and cross-cultural comparability of generalised trust has been criticised in the past by Sobel (2002) and others (Stephany & Braesemann, 2017b). For the case of collaborative programming activities, we were interested in the willingness to trustworthy collaborate with strangers; a concept that we have approximated by generalised trust. In order to validate this presumption, we run a separate analysis in which we use the share of a city's population engaged in charitable organisations as a control instead of trust. Similarly to the results reported here, we observe a strong relationship of charitable engagement and contributions, while clicks and voluntary engagement are only weakly related.

15 Certainly, platforms, like Stack Overflow, have established their own reputation metrics, such as user ratings, in order to overcome information asymmetries. However, while these metrics indicate levels of trustworthiness of individual contributors, they do not facilitate the general collaborative mindset of a society that facilitates users to make contributions to the global community in the first place.

16 Sustainable Development Goals

17 https://www.itu.int/en/sustainable-world/Pages/default.aspx

18 It could be, for instance, that users from countries where English is not so commonly used, rather prefer to use programming platforms in other languages. However, an exploratory search on www.similarweb.com, a website to measure country-specific web traffic, conducted on in March 2019, shows that Stack Overflow is the number-one platform in the category 'Computer and Electronics' in 54 out of 57 countries listed on SimilarWeb. In Japan it is ranked fifth, in Russia and Colombia ranked second. In all OECD, but Japan, it is ranked first. Considering the high correlation ($\text{β}=0.81$, see Figure 1(C)) between Stack Overflow clicks and contributions, and the high popularity of the website globally and across OECD countries, we conclude that language barriers (even if we do not control explicitly for such barriers in the analysis) should, thus, be of minor importance in explaining the global distribution of Stack Overflow contributions.

Additional information

Funding

This research was in part supported by the H2020 European Research Council [grant number 335716].

Notes on contributors

Fabian Stephany

Fabian Stephany is a Computational Social Scientist at the Oxford Internet Institute, University of Oxford. His research focuses on the application of data science and social science statistics in education, migration, and public policy. [email: [email protected]]

Fabian Braesemann

Fabian Braesemann is Research Fellow and Data Scientist at the Saïd Business School and Research Associate at the Oxford Internet Institute, University of Oxford. His research focuses on data mining and the statistical analysis of large-scale online data to understand market and information dynamics in a digitally connected world. [email: [email protected]]

Mark Graham

Mark Graham is Professor of Internet Geography at the Oxford Internet Institute, a Faculty Fellow at the Alan Turing Institute, a Senior Research Fellow at Green Templeton College, and an Associate in the University of Oxford's School of Geography and the Environment. He leads a range of research projects spanning topics between digital labour, the gig economy, internet geographies, and ICTs and development; with Digital Geographies being his most long-standing research area. [email: [email protected]]