How did cooperation research change over time?

By CoDa

by Daniel Balliet, Giuliana Spadaro, Benny Markovitch, & Wouter Beek

The Cooperation Databank (CoDa) contains studies about human cooperation using social dilemmas that were conducted between 1958 and 2017. Social dilemmas may be one of the longest used standardized methods to study human social behavior. This research tradition began with the use of the prisoner’s dilemma. Although researchers continue to use this method, several other paradigms have been developed to study cooperation, including public goods dilemmas and resource dilemmas. Also, the participants involved in these studies may have changed over time. And certainly the topics (e.g., independent variables) that researchers have focused on has changed over the past 60 years.

For example, earlier research was based mostly on student samples in psychology laboratories, but more recent years may have seen changes in terms of student samples and the country of participants. Early research focused on a few topics, such as communication and individual differences, but the topics have expanded across the years to include other variables that can explain cooperation. Here, we report how methods, samples, and topics have changed over the course of 60 years of research on human cooperation using social dilemmas.

Importantly, this review is a living document that will be automatically updated with the entry of new studies to CoDa.

This data story will address the following questions:

  1. Did methods change over time?
  2. Did samples change over time?
  3. How did research topics change over time?

In this data story we do not define the many different method characteristics, sample characteristics, and independent variables used in cooperation research. You can learn about the definitions and values of these variables at the CoDa website and using the ontology explorer in the CoDa platform.

Did methods change over time?

We begin by plotting studies over time (year of data collection) according to the different types of social dilemma paradigms. As displayed in Figure 1, studies began using the prisoner's dilemma, but since 2000, relatively more studies have been conducted using the public goods dilemma. Resource dilemmas have received relatively much less attention in the history of research on cooperation.

Figure 1 - Displays the number of studies using the different social dilemma paradigms, plotted according to the year the data was collected.

In the following three figures, we plot studies over time (year of data collection) according to several aspects of methods that can differ between studies. For each figure, you can select which variables to display in the figure, by first selecting True or False in the corresponding dropdown menu, and then selecting to run the query.

As you can see in Figure 2a (default), early research relied mostly on partner matching, while more recent studies used a balance between both stranger matching and partner matching experimental designs. In Figure 2b, you can see that early research mostly studied dyadic interactions, and this gradually declined over time, with modern research allocating more attention to cooperation in groups. In Figure 2b, you can also plot whether studies used one-shot versus repeated interactions, and here you can see that most early research used repeated interactions, and that one-shot interactions became much more common since the 1990s. Figure 2c plots studies according to the degree of conflict experienced in the social dilemma, operationalized as either MPCR or the K value. Research over the last 60 years has continued to study cooperation in situations that vary substantially in terms of the degree of conflicting interests people experience, and this hasn't changed much over time.

Make your own decisions about what to display in the figures below, and explore how the methods used in cooperation research have changed over the last 60 years.

Figure 2a ― Number of studies plotted by the year the study was conducted and according to (up to 3) different methods used in cooperation research (stranger versus partner matching, continuous versus step-level public good, and type of incentives).

Figure 2b ― Percentage of studies (per year) plotted by the year the study was conducted and according to (up to 3) different methods used in cooperation research (dyads versus groups; use of deception; and one-shot versus repeated interactions).

Figure 2c ― Number of studies plotted by the year the study was conducted and according to (up to 2) indices of the degree of conflicting interests in the social dilemma (i.e., values of MPCR and K index).

Did samples change over time?

The participants included in studies on human cooperation did change over time. In Figure 3a, you can select to display the average percentage of samples that were male participants (per year). You can see that earlier studies had relatively more male participants, while in more recent years there is a gender balance in the samples used in cooperation research. Figure 3a also displays that sample sizes have become larger over time. The average age of the sample has also increased over time. Figure 3b reports that the mean age of samples has slightly increased over time, while there is a sharp increase in the maximum age of participants in samples.

Figure 3a ― Plots the number of studies per year, including the averages of two sample characteristics (per year)(i.e., the average sample size and the average percent male).

Figure 3b ― Plots the average mean age of samples per year, including the average mean minimum age and mean maximum age of samples.

The following figures display information about student samples and the recruitment methods used to acquire samples in research on cooperation. As displayed in Figure 4a, much earlier research on cooperation relied on student samples that were psychology students. Figure 4b shows that although more recent studies continue to have student samples, there has been a decline in the relative proportion of studies using student samples. As shown in Figure 4c, this decline in student samples could be attributed, in part, because of the recent uptick of studies using Mechanical Turk to recruit samples.

Figure 4a ― Plots the number of studies per year that involved student samples and the academic discipline of students.

Figure 4b ― Plots the proportion of studies per year that involve student samples, and the proportion of studies that have student samples from different academic disciplines (i.e., psychology, economics, sociology, mixed, other)

Figure 4c ― Plots the number of studies per year according to the recruitment methods for participants. You can select to display the studies by (1) the number of studies per year or (2) the proportion of studies per year using a specific recruitment method.

CoDa includes studies that were conducted all around the world. The original search for studies was done in English, Japanese and Chinese. Figure 5a plots studies according to whether the documents that reported the studies were written in English, Japanese or Chinese. Most studies were published in English, but this figure also tracks the increase in research on cooperation published in Japanese and Chinese. For example, research published in Japanese began in the 1970's, while research published in Chinese stated in1999.

Figure 5b plots the papers, studies, and effect sizes reported in studies that were conducted in different countries/regions. This research tradition began in the United States. You can see this in Figure 5b by selecting the range of years you would like to display in the figure (1958-2017). However, when defining the range of years to be between 1958 to 1965, shows that research on cooperation had expanded beyond the United States by then to being conducted in Canada and the United Kingdom. In 2018, research on cooperation using social dilemmas has been conducted in a total of 61 countries. You can set the range of dates in this figure to observe how this research tradition spread around the globe.

Figure 5a ― The number of studies plotted over time that were documented in English, Japanese, and Chinese. You can select to display the studies by (1) the number of studies per year or (2) the proportion of studies per year reported in a specific language.

Figure 5b ― Number of studies and effect sizes reported per country around the globe. You can select a range of years to zoom in on a specific time period, or to observe how this research tradition spread around the world over time.

The figures below report the correlation between the year a study was conducted and several characteristics of the methods and samples used in those studies. Figure 6a reports the correlation between year a study was conducted and several continuous study and sample characteristics. Figure 6b reports the correlation between the year the study was conducted and dichotomous study and sample characteristics.

How did research topics change over time?

One way that cooperation research has changed over time are the variables that are measured and manipulated within studies and used to predict cooperation. Below we report the number of studies that have measured or manipulated the different variables that have been annotated in CoDa. It's a long list of variables, and the table captures the diversity of topics that have been studied in this area of research.

In Figure 7, you can select the range of years to display, to narrow your attention on a specific time period of cooperation research. You can also select whether the data are reported according to any number of years that fall within that range. Finally, you can restrict the Figure to only report a specific independent variable.

Figure 7 -- Number of studies per independent variable.

In conclusion, this data story provides an overview of how research methods, samples, and topics in research on human cooperation have changed over time. Importantly several of the research methodology has remain constant across the years. Social dilemma paradigms are continually used that have incentives for behavior, a well-defined conflict of interests, and younger, student samples. That said, some features of the methods have changed over time, such as an increase use of stranger matching experimental protocol, and one-shot interactions, perhaps in part because of the awareness that reputational concerns can shape how people behave. The samples in cooperation research have also improved over time. Samples have become larger, more gender balanced, having a broader range of age, and involved more than student samples. The topics of cooperation research have also expanded with the field.

This data story will be automatically updated when new studies are added to the databank - and so this review will continually monitor and document the many trends that exist in cooperation research over time.

How to cite this report:

Balliet, D., Spadaro, G., Markovitch, B., & Beek, W. (2020, date). How did cooperation research change over time? Cooperation Databank. Link.