We have no idea — same analysis, different results

In a recent paper, Akira Igarashi and James Laurence look at anti-immigrant attitudes in the UK and Japan. Like my 2019 paper in JEMS, they and how they highlight the limited research on non-Western countries, but they analysis they do is much more similar to what Sjoerdje van Heerden and I did in Urban Studies. Them like us relied on panel data to get a better handle on changing attitudes to immigrants. Them like us looked at the share of foreigners in the area (this relates to theoretical expectations that individual attitudes to immigrants reflect changes in the share of foreigners in the area; we refer to the same theories). We both used fixed-effect panel models. They find that “increasing immigration harms attitudes towards immigrant”, while we report that “a larger change in the proportion of immigrant residents is associated with more positive views on immigrants among natives” — yes, the exact opposite!

Need another example? Several studies examine the impact of sudden exposure to refugees on attitudes to immigrants and votes for radical-right parties. Such sudden exposure happened for example in Austria and Germany in 2015. In separate analyses, Andreas Steinmayr 2020 finds a clear increase in support for the radical-right, as we find in the work by Lukas Rudolph and Markus Wagner. Max Schaub, Johanna Gereke and Delia Baldassarri, by contrast “record null effects for all outcomes”. Same situation, same strategy to obtain the results.

We could now start the detective work, examining the small differences in modelling, ponder about the impact of how we define neighbourhoods, invoke possible differences between the countries (are the Netherlands an expectation, when the UK and Japan yield the same results? — not likely). Or we could admit how little we know, how much uncertainty there is in what we do, how vague our theories are in the social sciences that we can come to quite different conclusions in quite similar papers. I guess what we can see here is simply a scientific search for answers (it’s not like our research output would otherwise disagree so clearly). It’s probably also a call for more meta-level research: systematic analyses that synthesize what we do and don’t know, because even though individual papers sometimes contradict, we know quite a lot!

Heerden, Sjoerdje van, and Didier Ruedin. 2019. ‘How Attitudes towards Immigrants Are Shaped by Residential Context: The Role of Neighbourhood Dynamics, Immigrant Visibility, and Areal Attachment’. Urban Studies 56 (2): 317–34. https://doi.org/10.1177/0042098017732692.

Igarashi, Akira, and James Laurence. 2021. ‘How Does Immigration Affect Anti-Immigrant Sentiment, and Who Is Affected Most? A Longitudinal Analysis of the UK and Japan Cases’. Comparative Migration Studies 9 (1): 24. https://doi.org/10.1186/s40878-021-00231-7.

Rudolph, Lukas, and Markus Wagner. 2021. ‘Europe’s Migration Crisis: Local Contact and Out‐group Hostility’. European Journal of Political Research, May, 1475-6765.12455. https://doi.org/10.1111/1475-6765.12455.

Ruedin, Didier. 2019. ‘Attitudes to Immigrants in South Africa: Personality and Vulnerability’. Journal of Ethnic and Migration Studies 45 (7): 1108–26. https://doi.org/10.1080/1369183X.2018.1428086.

Schaub, Max, Johanna Gereke, and Delia Baldassarri. 2020. ‘Strangers in Hostile Lands: Exposure to Refugees and Right-Wing Support in Germany’s Eastern Regions’. Comparative Political Studies, September, 001041402095767. https://doi.org/10.1177/0010414020957675.

Steinmayr, Andreas. 2020. ‘Contact versus Exposure: Refugee Presence and Voting for the Far-Right’. The Review of Economics and Statistics, May, 1–47. https://doi.org/10.1162/rest_a_00922.

Zschirnt, Eva, and Didier Ruedin. 2016. ‘Ethnic Discrimination in Hiring Decisions: A Meta-Analysis of Correspondence Tests 1990–2015’. Journal of Ethnic and Migration Studies 42 (7): 1115–34. https://doi.org/10.1080/1369183X.2015.1133279.

p-hacking: try it yourself!

It’s not new, but it’s still worth sharing:

The instructions go: “You’re a social scientist with a hunch: The U.S. economy is affected by whether Republicans or Democrats are in office. Try to show that a connection exists, using real data going back to 1948. For your results to be publishable in an academic journal, you’ll need to prove that they are “statistically significant” by achieving a low enough p-value.”

The tool is here: https://projects.fivethirtyeight.com/p-hacking/

And more on p-hacking here: Wikipedia — to understand why “success” in the above is not what it seems.

Excellence by Nonsense

I know it’s 5 years old, but I still think this description of academia deserves a wider audience.

In this chapter, Binswanger (a critic of the current scientific process) explains how artificially staged competitions affect science and how they result in nonsense. An economist himself, Binswanger provides examples from his field and shows how impact factors and publication pressure reduce the quality of scientific publications. Some might know his work and arguments from his book ‘Sinnlose Wettbewerbe’.

Binswanger, Mathias. 2014. ‘Excellence by Nonsense: The Competition for Publications in Modern Science’. In Opening Science: The Evolving Guide on How the Internet Is Changing Research, Collaboration and Scholarly Publishing, edited by Sönke Bartling and Sascha Friesike, 49–72. New York: Springer. https://doi.org/10.1007/978-3-319-00026-8_3. [open access]