Summary of our second ESRC Workshop on "Increasing the richness and frequency of social science survey data" (23/5/14)

Thanks to all of those who attended our second ESRC workshop on “Increasing the richness and frequency of social science survey data” in Stirling on Friday May 23rd. It was the second of our six workshops funded by the ESRC that take place in 2014/15.

We heard fascinating talks and had illuminating discussions about the present and the future of using new measures in the social sciences. The abstracts of the talks, pictures, and some links for further reading are below. Details of future workshops will be provided via the mailing list, the blog and our twitter account.


09:20-09:40: Dr. Michael Daly (Stirling Behavioural Science Centre):
Introduction and Opening Talk

You'll find the slides that Michael used for his presentation here.

09:40-10:10: Jos van den Puttelaar (Wageningen): 

Measuring choice behaviour in a simulated supermarket; a comparison of survey result s, virtual and actual behaviour.

Consumer research in retail settings was usually constricted to either questionnaires about behaviour, an assignment with one specific shelf or in a rare occasion implemented into a real store. The ideal was to be able to do consumer research in larger areas and still be able to track a consumer from step to step. And while doing that, not spending a lot of man-hours or money. For this reason we developed a computer program that we could use to simulate shopping behaviour wherever we wanted, whatever we wanted to simulate, are total control of and with fully automated tracking.

At this moment the developments are threefold. First of all on the technological side, we have a drag and drop system that can build and simulate supermarket shopping without knowledge of programming. Added to this are the implementation of several techniques like, eye tracking, facial expression analysis and automated ambiance creation. Secondly, we have done and are still busy with several validation studies. Third, the aim of the whole project, was to do research into consumer behaviour. By now, several studies have been finished and published. Next to that many students have used the tool in their BSc or MSc thesis research. Altogether the shop simulator shows to be a research tool that extends very well on survey data with actual behavioural data.

10.10-10.40: Anouar El Haji (University of Amsterdam)
Online Experimental Auctions: Measuring Willingness to Pay in the Field

Product value is often measured through non-incentivized measures, such as survey scales or hypothetical willingness to pay. The second-price auction is a well-established auction mechanism to elicit true maximum willingness to pay. However, due to practical reasons these auctions are usually only conducted among students and on a small scale. We overcome both limitations with an online platform that makes it possible to run second price auctions with a large sample of consumers. We argue that the platform offers researchers access to unique data with a high external validity.

11:00-11.45: Ernesto Schwartz Marin (University of Durham)
Citizen Led Forensics: DNA and data banking as technologies of disruption.

Governmental institutions in Mexico have officially recognised 121,683 violent deaths in the period between 2006- 2013. During the same period around 27,000 disappearances occurred, and approximately 15,000 bodies remain unidentified, as there are no national databases in the country. In Mexico, distrust in governmental authorities is the norm, since the practice of forensic science has been opaque, and has sometimes itself been used to cover the tracks of the perpetrators of grave crimes. In response to this crisis, and thanks to an ESRC transformative research grant (2014-2015), this project aims to create the first citizen led Forensic DNA database as a way of positively intervening in the humanitarian crisis currently lived in Mexico.The project will make DNA swab kits available for 1,500 people (approx 500 Mexican families), accompanied by a clear set of instructions on how to collect DNA from cheek swabbing, as well as from the personal belongings of the missing person. In the same DNA Kit we would ask the participating families to include written accounts of their case, their experience with forensic investigations (if any) and personal narratives of what they have gone through since their relative(s) disappeared The citizen led Forensic database will be designed as a mobile and Participatory Action Research device, articulated through civil society organisations of relatives of the disappeared. The presentation will cover the preliminary design of this database and the challenges and opportunities opened by its creation. The work is based on fieldwork, first hand experience and recurrent analysis of the current situation lived in Mexico.

11.45 -12.30: David Stillwell (University of Cambridge)

Title: myPersonality: Example of successful online social network data collection from 6 million people

Link: The myPersonality project website

Abstract: In June 2007 Facebook released its apps platform, and since then commercial apps like Candy Crush Saga and Farmville have entertained hundreds of millions of users around the world. With a few tweaks, academics can also create interesting experiments, questionnaires and games that attract millions of respondents. The myPersonality application collected psychological questionnaire responses from 6 million users whose sole motivation was to receive feedback on their results. Users also opted-in to sharing their Facebook profile information, including demographics, political/religious views, Facebook Likes, friendship network graph, and status updates. This talk will discuss the data available through Facebook, and give some tips on how to entertain users with academic research studies.

13:30-14:15: Dr Julia Allan (University of Aberdeen)
Capturing the real time determinants of snacking and inactivity in daily life: the SNAPSHOT study
Julia Allan

Link: The SNAPSHOT Study

We live in an environment where snack foods are readily available, and where desk jobs and labour saving devices mean that we are less active than in previous generations.   In this context, eating well and being active requires considerable self-control (i.e. being able to do/not do something effortful now in order to achieve something valuable in future). Most research linking self-control to health behaviour has focused on stable, trait-level differences between people.  However, the ability to exert self-control is known to fluctuate within people over time, as mental resources are depleted and replenished.  This means that people may be particularly likely to succumb to dietary temptation / fail to be active at moments when their regulatory resources are depleted.   In this talk, I give an overview of the Scottish Government funded SNAPSHOT study, that monitors snacking, physical activity, mood, self-control, social context, and location in real time as people go about their everyday lives.  By integrating these different measures, it’s possible to build up a rich and detailed picture of when, where and why people snack and spend time inactive.  The pros and cons of this type of real time data collection will be discussed.

14:15-15:00: Andrew Jones (University of Liverpool)
Response Inhibition in Everyday Life. Administering a Mobile Version of the Stop-Signal Task set up on an Android Smartphone.

Response inhibition (disinhibition) or the inability to inhibit behaviour that is no longer appropriate is a core component of impulsivity and executive functioning. We can measure response inhibition in the laboratory using well validated, computerised assessments such as the Stop-Signal and Go/No go tasks.  Poor performance on these tasks is linked to problem drinking in cross-sectional studies; however it is impossible to infer causality from this. Laboratory research is beginning to demonstrate that inhibition can fluctuate over time in response to internal and environmental cues and these fluctuations may increase the risk of alcohol consumption. We aimed to test this hypothesis by examining inter-individual differences in disinhibition and alcohol consumption over time. I will discuss an ecological momentary assessment study, in which we administered a stop-signal task via a mobile phone application to a sample of heavy drinkers to examine daily fluctuations in inhibition along with alcohol consumption.  The focus will be on our preliminary results, the costs/benefits of the methodology, reliability of the app and future research using these methods.

15:30-16:40: Professor Neil Stewart (Warwick University).
The Psychology of Nations and States

Link: Google Trends
Link: Mechanical Turk


People search the internet for things they are thinking about. We used Google Trends to extract search frequencies for years (e.g., 2011) and adjectives (e.g., "hot"). With the years, we construct a measure of future orientation and a measure of intertemporal span and show that both are correlated across nations with higher per-capita gross domestic product. With the adjective frequencies we construct a measure of the connotation of searches from different US states, and show that these measures are correlated with per capita gross state product and inequality. Overall, we demonstrate that constructing psychological measures from Google Trends data is possible and that these measures correlate with sensible economic measures like GDP and inequality.

16:40-17:30 Panel Discussion

Follow our Twitter account @StirlingEconPsy and the hashtag #StirBSC to follow the day's events.