Embedded Files
Public good induced values as determinant of voluntary contributions to research activities
The times they are a-changin’ – Bob Dylan
With a little help of my friends – The Beatles
Don’t let a good crisis go to waste – W. Churchill
The shoemaker’s children always go barefoot – Anonymous
Date: 2020-06-26; Authors: Marianne Lefebvre, Jens Rommel & Jesus Barreiro-Hurle; Image: Unkown origin
The times they are indeed a-changin’ and having unexpected impacts on our life. At this moment of time, we would have been expecting a response from our first attempt at obtaining competitive funding for consolidating our Research Network (COST action), we would have been planning our trip to Prague to participate at the organized sessions at the EAAE Congress; instead, we are still coming to terms with the new situation.
To keep the momentum up, the REECAP Secretariat decided to ask our friends for a little help, that is to kick-start the activities planned under the COST Action. We believed that some of the activities planned in the proposal could already be initiated based on the altruistic and voluntary contribution of members in the network. While researchers face the pressure of improving their individual publication records, their participation to international networks is also valued. It requires extra time and somehow different skills, but can be rewarding. Some REECAP members may therefore face a public good dilemma. Nudging may foster voluntary contributions to this public good.
At the same time, we thought it would be a pity to let this crisis go to waste and decided to set up an experiment that could be embedded in the call for volunteers. Following Etner et al. (2009) we considered that a good theoretical basis for the experiment would be whether the initial status of the public good to which individuals are asked to contribute has an impact on the total contribution. As they state in their paper (focusing on environmental quality as a public good) “if the individual considers the initial level of environmental quality as sufficiently high, they prefer […] not to contribute to the improvement of environmental quality. This situation appears independent from the environmental risk. Second, if the individual considers environmental quality level as sufficiently low, […] they contribute more to improve environmental quality” (page 313). We thought that a similar behaviour could be expected if we shifted from an environmental public good to a research public good. Our experiment hinges on the assumption that the level of research environment quality in the highly uncertain COVID context could be induced exogenously via a script included in the email sent to the REECAP network.
Our experiment involved a two-step approach. People were first contacted with a message that featured the experimental conditions. Each of the REECAP members was randomly assigned to one of three experimental conditions described in Table 1. [1] Then there was a follow-up survey to gather additional data.
[1] We tested for the effectiveness of the randomization process and found no differences across the 3 conditions for gender, years of postdoctoral experience, share of less research-intensive (vs and other EU countries), share of practitioners working for agricultural administration or consultancy (vs researchers) and membership in the Board of REECAP. Note the small sample size and see Briz et al. (2017) for an interesting discussion of balance tests in experimental economics.
The sample consisted of people that have shown interest in the REECAP COST action. They were contacted via email, informing them about the postponement of the COST action submission date and the possibility of starting the planned activities before learning about the outcome of the funding decision. The email had a link to an on-line survey in which people could declare their willingness to participate in one or more of the working groups of the REECAP network [2]. We also asked to provide some basic background on their REECAP-related experience [3]. We obtained 60 responses of which 41 declared a willingness to contribute to at least one working group. Once the deadline set in the first email expired, the original sample was contacted again disclosing the experiment and asked a series of follow-up questions to measure the level of optimism (pessimism) and risk aversion of the respondent and elicit their perception of the quality of actual and future research environment.
Our main outcome variable is the share of respondents willing to participate in the REECAP network and our hypothesis, in line with Etner et al. (2009), was that contributions in TLow should be higher than in Control and contributions in THigh lower than in Control. Results are reported in Figure 1, considering that those who did not respond to the first email were not willing to contribute to the network’s activities.
[2] Working Group #1 “Bringing experiments into agricultural policy”; Working Group #2 “Bringing agricultural policy into the experimental field”; and Working Group #3 “Outreach activities”.
[3] Participation at REECAP workshops and expertise on CAP and experimental methods.
We only find weak support for the hypothesis that the nudge inducing low research environment quality resulted in higher contributions to the public good when comparing the control group (Two-sample test of proportions Control TLow p-value=0.0748). Overall, there is no significant differences across conditions (Kruskal-Wallis equality of populations rank test - p-value=0.4632), and the small sample size makes both, a type I and a type II error, very likely.
What do we make of these results? We can be factual and say that the nudge failed to have a substantial impact on the nudgers in this instance. Alternatively, we can be optimistic and hope for a type II error, use the excuse of our small sample and consider that even just one comparison showing a (marginal) impact serves as evidence that no-one is free from being nudged for the better. While nudging is not a universally accepted practice from an ontological and ethical perspective (see for example Lepenies & Małecka, 2015), we believe understanding its potential and limitations for different groups of individuals can inform framing this debate. Guess that we did not find the Holy Grail test that allows us to conclude that individuals acquainted with nudging can be nudged, contrary to what has been found for market experience and hypothetical bias (List, 2011). [4] We still do not know for sure if shoemakers’ children go barefoot.
This is just a preliminary analysis of your responses. We are digging further into the responses to the second questionnaire (n=48) to see whether risk and optimism variables can explain the differences we have detected.
Lessons learnt
During this small experiment we have learnt some lessons on how deal with implementation. First, the experiment was pre-registered on aspredicted.org with reference #38756. This is very quick to do once you have clear ideas and it pays off, as it makes the analysis of the results easier. Moreover, it serves as a quality cue for your research when you want to publish it. We therefore strongly advice to pre-register experiments. We welcome all members of the network to reach out to us if they are interested in seeing the preregistration. We are also happy to provide any guidance on the process.
Second, deception and data protection can be a tricky animal to deal with. Although we did not deceive people, obtaining informed consent from participants before data collection or going through ethical approval should be key steps in any experiment. Here, we opted for consent after data collection and offered deletion of the data.
Thanks to all those of you who took part in our sample! We hope you value your contribution to the common knowledge effort.
Marianne Lefebvrea, Jens Rommelb & Jesus Barreiro-Hurlec
a Université d′Angers; b Swedish University of Agricultural Sciences; c European Commission Joint Research Centre
COMMENTS
We will be happy to publish your comments to this post here. Please write to jens.rommel@slu.se
Disclaimer: The views expressed are purely those of the authors and may not in any circumstances be regarded as stating an official position of the European Commission.
[4] Probably that explains why you are not reading this in AER or JEEM!
References:
Briz, T., Drichoutis, A. C., & Nayga Jr, R. M. (2017). Randomization to treatment failure in experimental auctions: The value of data from training rounds. Journal of Behavioral and Experimental Economics, 71, 56–66: https://doi.org/10.1016/j.socec.2017.09.004
Etner, J., Jeleva, M., & Jouvet, P.-A. (2009). Pessimism or optimism: a justification to voluntary contributions toward environmental quality. Australian Economic Papers, 48(4), 308–319.
Lepenies, R., & Małecka, M. (2015). The institutional consequences of nudging–nudges, politics, and the law. Review of Philosophy and Psychology, 6(3), 427-437.
List, J. A. (2011). Does market experience eliminate market anomalies? The case of exogenous market experience. American Economic Review, 101(3), 313-17
Addendum, 2020-07-03
Nudges work on Richard Thaler too
Recently, The Nudgefather mused on the question of “what’s next for nudging and choice architecture?” He did his best to ignore the pleas from Katy Milkman to write the introduction to the special issue she was organizing. But just when he thought he was out of the obligation, she pulled him back in. In a light and speedy read, Thaler lays out how Milkman’s emails used behavioral science to make him an offer he couldn’t refuse. More importantly, he shares his hopes and dreams for what comes next. Among these is the desire to see behavioral scientists involved in creating choice architecture from the start.
“Often, behavioral scientists are asked to help change a particular behavior but are severely limited in the ways they can alter the underlying environment,” he explains. “Rarely, if ever, are researchers given the opportunity to design the entire choice architecture. We get to remodel the kitchen, but not design the entire home, let alone pick the lot on which it is built.” [Organizational Behavior and Human Decision Processes]
Google Sites
Report abuse