Wednesday, October 29, 2014

Launch of What Works Centre for Wellbeing

Via ESRC website - full details available here 
A new centre is being set up to gather and share evidence on what works to improve wellbeing. The independent What Works Centre for Wellbeing is the latest addition to the What Works Network, launched by the Government last year to improve public services through evidence-based policy. 
Funded by the Economic and Social Research Council (ESRC), Public Health England and other partners, including government departments, the centre will become operational next spring, when grants will be awarded to universities to research the impact that different interventions and services have on wellbeing. 
The initial focus of research will be on the themes of work, communities and culture. The results will help government, councils, health and wellbeing boards, charities and businesses make decisions and choices informed by what the evidence says really matters for the wellbeing of people, communities and the nation as a whole. 
To mark the launch of the What Works Centre for Wellbeing, the Department for Business, Innovation and Skills published complementary research on employee wellbeing, what workplace factors influence it and how wellbeing affects performance.

Irish Government Economic Evaluation Service on Behavioural Economics

The Irish Civil Service established an economic evaluation service in 2012 "to enhance the role of economics and value for money analysis in public policy making". Interestingly, they have recently published a document on the role of behavioural economics in policy-making in Ireland. This is available on this link.

The Impact of Text Message Reminders on Adherence to Antimalarial Treatment in Northern Ghana: A Randomized Trial

The Impact of Text Message Reminders on Adherence to Antimalarial Treatment in Northern Ghana: A Randomized Trial

Julia R. G. Raifman 1, Heather E. Lanthorn 2, Slawa Rokicki 3, Günther Fink 1

1 Department of Global Health and Population, Harvard School of Public Health, Boston, MA, United States of America, 2 Harvard School of Public Health, Boston, MA, United States of America, 3 Department of Health Policy, Harvard Graduate School of Arts and Sciences, Cambridge, MA, United States

PLoS ONE 9(10): e109032. doi:10.1371/journal.pone

Abstract

Background

Low rates of adherence to artemisinin-based combination therapy (ACT) regimens increase the risk of treatment failure and may lead to drug resistance, threatening the sustainability of current anti-malarial efforts. We assessed the impact of text message reminders on adherence to ACT regimens.

Methods

Health workers at hospitals, clinics, pharmacies, and other stationary ACT distributors in Tamale, Ghana provided flyers advertising free mobile health information to individuals receiving malaria treatment. The messaging system automatically randomized self-enrolled individuals to the control group or the treatment group with equal probability; those in the treatment group were further randomly assigned to receive a simple text message reminder or the simple reminder plus an additional statement about adherence in 12-hour intervals. The main outcome was self-reported adherence based on follow-up interviews occurring three days after treatment initiation. We estimated the impact of the messages on treatment completion using logistic regression.

Results

1140 individuals enrolled in both the study and the text reminder system. Among individuals in the control group, 61.5% took the full course of treatment. The simple text message reminders increased the odds of adherence (adjusted OR 1.45, 95% CI [1.03 to 2.04], p-value 0.028). Receiving an additional message did not result in a significant change in adherence (adjusted OR 0.77, 95% CI [0.50 to 1.20], p-value 0.252).

Conclusion

The results of this study suggest that a simple text message reminder can increase adherence to antimalarial treatment and that additional information included in messages does not have a significant impact on completion of ACT treatment. Further research is needed to develop the most effective text message content and frequency.

Trial Registration

ClinicalTrials.gov NCT01722734

Tuesday, October 28, 2014

November 21st ESRC Workshop on Preferences and Personality


ESRC Workshop 4: Preferences and Personality (21/11/14)

This is the fourth Behavioural Science Workshop in a series of six that will take place in 2014/15. These workshops are funded by the Economic and Social Research Council. The venue is the Maclaren Suite in the Stirling Highland Hotel. There will be drinks and dinner after the days talks to which all attendees are welcome.

One of the major challenges in economics is understanding the statistical properties of measures of time, risk, and social preferences and evaluating the validity of such measures. This workshop will focus on empirical research examining economic preferences in laboratory and real-world settings

Speakers will address the reliability of traditional preference measures, their structure across demographic characteristics, innovations in measurement, and links between preference estimates and objective economic and biological measures. We have invited speakers who are engaged in the theoretical and empirical mapping of preference measures to personality traits which have been shown to have substantial predictive validity for important life outcomes (e.g. income, disease morbidity and mortality, employment). Taken together, this workshop will enhance cross-talk and expand the common conceptual ground that exists between personality psychologists and economists interested in the assessment of preferences in the UK and Europe. Furthermore, it will cultivate frontier thinking regarding the future data-collection priorities for social science in the UK and further afield.

Sign up to attend the workshop here


DAY SCHEDULE

09:15-10:00: Professor Alex Wood (Stirling Behavioural Science Centre)

Integrating personality psychology and economics

10:00-10:30: Bernardo Fonseca Nunes (Stirling Behavioural Science Centre)
Transition to retirement and home production: personality explains heterogeneous changes in housework at retirement
Abstract: Previous studies on home-production at retirement do not consider the role of individual personality traits on the time retirees devote to housework. Here we examine whether personality determines the heterogeneous changes on the time individuals devote to housework due to a transition to retirement from the labour market. We use British longitudinal data which included individuals’ personality measures, and responses about the amount of hours spent per week on housework tasks. We find a positive change in housework hours for male and female retirees. Personality traits are shown to be more relevant on the explanation of housework changes at retirement than consumption expenditures, household income, and gender.


10:30-11:00: COFFEE

11:00-11:45: Dr. Hannah Schildberg-Hörisch (University of Bonn)
How does parental socio-economic status shape a child’s personality?
(with Thomas Deckers, Armin Falk, Fabian Kosse).
Abstract: We show that socio-economic status (SES) is a powerful predictor of many facets of a child's personality. The facets of personality we investigate encompass time preferences, risk preferences, and altruism, as well as crystallized and fluid IQ. We measure a family's SES by the mother's and father's average years of education and household income. Our results show that children from families with higher SES are more patient, tend to be less likely to be risk seeking and more altruistic, and score higher on IQ tests. About 20 to 40% of this relationship can be explained by dimensions of a child's environment that are shown to differ by SES: parenting style, quantity and quality of time parents spend with their children, the mother's IQ and economic preferences, a child's initial conditions at birth, and family structure. Moreover, we use panel data to show that the relationship between SES and personality is fairly stable over time at age 7 to 10. 

Personality profiles that vary systematically with SES offer an explanation for social immobility. In a companion study, we present evidence on a randomly assigned variation in life-circumstances, providing children with a mentor for the duration of one year. Our data reveal a significant increase in altruism in the treatment relative to the control group. These findings thus provide evidence in favor of a causal effect of social environment on the formation of altruism. Moreover, we show that enriching life-circumstances bears the potential to close the observed developmental gap in altruism between low and high SES children.

11:45-12:15: Dr. Christopher Boyce (Stirling Behavioural Science Centre) 
Individual differences in loss aversion: Does personality predict how life satisfaction responds to losses versus gains in income? (with Alex Wood and Eamonn Ferguson). 
Abstract: Loss aversion is generally regarded as a pervasive bias occurring regardless of context or decision-maker. No studies have examined the relationship between personality and loss aversion. Here, using data from Germany (N = 18,039), we examine whether the effect of income losses (versus income gains) on life satisfaction differ by personality. We show that, although there are no personality differences in how gains relate to life satisfaction, when experiencing an income reduction people higher on conscientiousness (versus those lower) exhibited larger declines in life satisfaction. Similarly, those lower on openness (versus those higher) experienced larger life satisfaction falls. Our results suggest; (a) important individual differences in loss aversion, (b) personality interacts with socio-economic events to influence life satisfaction, (c) some personality traits may promote resilience in this context, and (d) income relates to life satisfaction only for individuals that experience income losses, and have high conscientiousness or low openness.

12:15-13:00: Dr. Bart Golsteyn (University of Maastricht)
Risk attitudes across the life course
Abstract: This paper investigates how risk attitudes change over the life course. Even with panel data that span several years, age patterns are generally difficult to identify separately from cohort or calendar period effects. We provide first evidence on the age profile of risk attitudes all the way from early adulthood until old age, in large representative panel data sets from the Netherlands and Germany, using a proxy variable approach to achieve identification. The main result is that willingness to take risks decreases over the life course, linearly until approximately age 65 after which the slope becomes flatter.

13:00-14:00: LUNCH

14:00-14:45: Dr. Elisa Cavatorta (King's London) 
Measuring ambiguity preferences (with David Schroeder Birkbeck).

Abstract: Ambiguity preferences are important in explaining human decision-making in many areas in economics and finance. To measure ambiguity preferences, the experimental economics literature advocates using incentivized laboratory experiments. However, in many circumstances, carrying out complex lab-experiments is not feasible. In this paper, we evaluate the ability of thought experiments and attitudinal questions to generate a behaviourally valid measure of ambiguity preferences. We find that a small set of thought experiments and attitudinal questions can serve as an alternative measure when carrying out laboratory experiments is impractical. Our results can be useful in many situations that require measuring ambiguity preferences in an easily implementable and cost-effective way, such as large surveys, field experiments, or everyday business and finance applications.

14:45-15:30: Professor Marjon Van Pol (University of Aberdeen)
Measuring time preferences: insights from the health context
Abstract: There is a relatively large empirical literature on individual time preferences for health outcomes.  This interest has been stimulated by policy concerns around health behaviours such as obesity and smoking and by the debate on the appropriate discount rate in the case of health outcomes.  It could be argued that the literature on time preferences for health has been more innovative in terms of elicitation methodologies used and methodological questions that have been examined.  This presentation will reflect on a range of measurement issues that have been observed in the context of time preferences for health including framing effects, decision heuristics and negative time preferences.  Measurement issues will be demonstrated using a number of case studies.  General lessons for the elicitation of time preferences will be drawn out.  The presentation will finish with a discussion around predictive validity: does type of outcome in time preferences tasks matter for the predictive validity of life outcomes such as health?


15:30-16:00: COFFEE

16:00-16:45: Professor Sule Alan (University of Essex) 
Good Things Come to Those Who (Are Taught How to) Wait: Results from a Randomized Educational Intervention
Abstract: We report results from a randomized evaluation of a unique educational intervention targeted at elementary school children in 3rd and 4th-grade in Turkey. The program, which lasts eight weeks, uses case studies to discuss issues related to forward looking behavior, improve the ability to imagine future-selves and evaluate different contingencies arising from different actions, supplemented by classroom activities supervised by trained teachers. We find that treated students make more patient intertemporal choices in incentivized experimental tasks. The effect is stronger for students who are identified as present-biased in the baseline. Furthermore, using official school administrative records, we find that treated children are significantly less likely to receive a low “behavioral grade”. These results are persistent one year after the intervention, replicate well in a different sample, and are robust across different experimental preference elicitation methods

16:45-17:30: Professor Eamonn Ferguson (Nottingham) 
Personality and Pro-Social Preferences 

Saturday, October 25, 2014

If you ever need to explain the availability heuristic...

...compare opinion polls about controversial topics to reality. From Ipsos MORI:

"A new survey by Ipsos MORI for the Royal Statistical Society and King’s College London highlights how wrong the British public can be on the make-up of the population and the scale of key social policy issues.  The top ten misperceptions are:
1.       Teenage pregnancy: on average, we think teenage pregnancy is 25 times higher than official estimates:  we think that 15% of girls under 16 get pregnant each year, when official figures suggest it is around 0.6%[i]
2.       Crime: 58% do not believe that crime is falling, when the Crime Survey for England and Wales shows that incidents of crime were 19% lower in 2012 than in 2006/07 and 53% lower than in 1995[ii].  51% think violent crime is rising, when it has fallen from almost 2.5 million incidents in 2006/07 to under 2 million in 2012[iii].
3.       Job-seekers allowance: 29% of people think we spend more on JSA than pensions, when in fact we spend 15 times more on pensions (£4.9bn vs £74.2bn)[iv].
4.       Benefit fraud: people estimate that 34 times more benefit money is claimed fraudulently than official estimates: the public think that £24 out of every £100 spent on benefits is claimed fraudulently, compared with official estimates of £0.70 per £100[v].
5.       Foreign aid: 26% of people think foreign aid is one of the top 2-3 items government spends most money on, when it actually made up 1.1% of expenditure (£7.9bn) in the 2011/12 financial year.  More people select this as a top item of expenditure than pensions (which cost nearly ten times as much, £74bn) and education in the UK (£51.5bn)[vi].
6.       Religion: we greatly overestimate the proportion of the population who are Muslims: on average we say 24%, compared with 5% in England and Wales.  And we underestimate the proportion of Christians: we estimate 34% on average, compared with the actual proportion of 59% in England and Wales[vii].
7.       Immigration and ethnicity: the public think that 31% of the population are immigrants, when the official figures are 13%[viii]. Even estimates that attempt to account for illegal immigration suggest a figure closer to 15%.  There are similar misperceptions on ethnicity: the average estimate is that Black and Asian people make up 30% of the population, when it is actually 11% (or 14% if we include mixed and other non-white ethnic groups)[ix].
8.       Age: we think the population is much older than it actually is – the average estimate is that 36% of the population are 65+, when only 16% are[x].
9.       Benefit bill: people are most likely to think that capping benefits at £26,000 per household will save most money from a list provided (33% pick this option), over twice the level that select raising the pension age to 66 for both men and women or stopping child benefit when someone in the household earns £50k+.  In fact, capping household benefits is estimated to save £290m[xi], compared with £5bn[xii] for raising the pension age and £1.7bn[xiii] for stopping child benefit for wealthier households.
10.   Voting: we underestimate the proportion of people who voted in the last general election – our average guess is 43%, when 65% of the electorate actually did (51% of the whole population)[xiv]
These misperceptions present clear issues for informed public debate and policy-making, which will be discussed at an event being run by the Royal Statistical Society, King’s College London and Ipsos MORI today, as part of the International Year of Statistics."
Any other good examples of this? I'm sure there's decades of American polling data to draw on.

Friday, October 24, 2014

Lecturer/Senior Lecturer Post at Stirling Economics

The Economics Division is seeking to recruit an outstanding lecturer/senior lecturer in Economics, with a focus on applied microeconomics in particular in areas such as health, education, aging and population economics. This post is offered at Lecturer or Senior Lecturer level, dependent on the experience and achievements of the candidate appointed.

Thursday, October 23, 2014

Lecture on Intertemporal Choice

I am currently giving a set of lectures as part of a module "Behavioural Economics: Concepts and Theories" in Stirling. I am posting brief informal summaries of some of these lectures on the blog to generate discussion. Thanks to Mark Egan for a lot of help in putting these together online.

Decisions that deliver benefits and costs over different time periods are central to the study of economics and a key area of interaction between economics, psychology and policy. This lecture reviews the basic discounted utility model. It examines hyperbolic discounting and dual-process accounts of inter-temporal choice. The lecture reviews domain specific discounting, children's discounting, preferences for sequences, heuristics employed in judging future utility, evidence on the power of defaults. It then examines recent evidence on neurological mechanisms involved in time preferences. The lecture concludes with a discussion of the policy issues at stake, in particular the implications for regulation of financial markets.

Readings:

Frederick, S., Loewenstein, G. & O’Donoghue, T. (2002), "Time discounting and time preference: a critical review", Journal of Economic Literature, 40: 351-401

Elster, J. (1985). Ulysses and the Sirens: Studies in Rationality and Irrationality. Cambridge: Cambridge University Press.

Fehr, E. (2002), "The economics of impatience", Behavioural Science, 415: 269-272.

Textbooks:

1. Camerer, Loewenstein & Rabin (2004) Advances in Behavioral Economics

2. Frey & Stutzer (2007), Economics and Psychology: A Promising New Cross-Disciplinary Field

3. Loewenstein (2007), Exotic Preferences: Behavioral Economics and Human Motivation

4. Shafir (2013), The Behavioral Foundations of Public Policy

5. Angner (2012), A Course in Behavioural Economics

6. Wilkinson & Klaes (2012) An Introduction to Behavioural Economics

7. Varian (2008), Intermediate Microeconomics

Monday, October 20, 2014

Mastering ’Metrics: The Path from Cause to Effect

Have not read this yet but the follow-up to the now famous primer on microeconometrics "Mostly Harmless Econometrics" is bound to be of big interest to a lot of readers here. Will do a couple of sessions on this internally when we get some copies. Have no idea what the Kung-Fu stuff is about but both authors are top of their game in terms of micro-econometrics and the reviews are very good so I am confident this will be a very useful book. 
Applied econometrics, known to aficionados as 'metrics, is the original data science. 'Metrics encompasses the statistical methods economists use to untangle cause and effect in human affairs. Through accessible discussion and with a dose of kung fu–themed humor, Mastering 'Metrics presents the essential tools of econometric research and demonstrates why econometrics is exciting and useful. 
The five most valuable econometric methods, or what the authors call the Furious Five--random assignment, regression, instrumental variables, regression discontinuity designs, and differences in differences--are illustrated through well-crafted real-world examples (vetted for awesomeness by Kung Fu Panda's Jade Palace). Does health insurance make you healthier? Randomized experiments provide answers. Are expensive private colleges and selective public high schools better than more pedestrian institutions? Regression analysis and a regression discontinuity design reveal the surprising truth. When private banks teeter, and depositors take their money and run, should central banks step in to save them? Differences-in-differences analysis of a Depression-era banking crisis offers a response. Could arresting O. J. Simpson have saved his ex-wife's life? Instrumental variables methods instruct law enforcement authorities in how best to respond to domestic abuse. 
Wielding econometric tools with skill and confidence, Mastering 'Metrics uses data and statistics to illuminate the path from cause to effect.
Shows why econometrics is important
Explains econometric research through humorous and accessible discussion
Outlines empirical methods central to modern econometric practice
Works through interesting and relevant real-world examples

Nick Chater's online psychology course

Nick Chater, Professor of Behavioural Science at Warwick Business School, is currently teaching a MOOC called "The Mind is Flat: the Shocking Shallowness of Human Psychology". The current edition of the course started last week but you can still sign up for free.

Sunday, October 19, 2014

PhD Research at the Centre

Our research centre currently has 9 PhD students who are either taking a PhD in Economics or a PhD in Behavioural Science. The centre structure is such that PhD students, research fellows and faculty work closely together on a variety of research questions. Some of the PhD students are also directly integrated into the Economics and/or Management divisions of our School of Management. Our centre meets weekly followed by an external seminar. There are also regular workshops and training events and informal peer-learning sessions including a regular STATA users group meeting. Most of our students are located in Stirling with dedicated office-space but we also have some students who are working part-time and we are willing to discuss this option. A very detailed overview of how PhD research is conducted here is available on this web page and it is worth studying carefully if you are thinking of applying to work with us. Furthermore we strongly encourage people to look at the publications page of our webpage to ensure that the type of work we are doing is interesting to you (interesting enough that you are willing to spend 3 or 4 years of your career and beyond working on similar work!).

For those wishing to secure funding to conduct their PhD from September 2015, it is worth starting this process early as funding deadlines tend to be early in 2015. The key funder for social science PhDs in the UK is the ESRC. The website for the Scottish branch of this funding is available here.  For those wishing to apply, the relevant pathways for our group are Economics, Business/Management and Advanced Quantitative Social Science. Economics does not have a residency requirement but most of the pathways are limited to UK residents and you should check this carefully to make sure you are eligible. Given the limited availability of PhD funding in the UK, self-funding is another option.


PhD in Behavioural Science

Below you can find details of our PhD in Behavioural Science from the webpage of the Centre for Graduate Study in the School of Management. Further details of PhD study in Stirling Management School are provided on this webpage.

The PhD in Behavioural Science programme is aimed at students who want to work in the Behavioural Science Centre to conduct and publish world leading research at the interface between the social sciences (such as economics) and the behavioural sciences (such as psychology). This area - which encompasses behavioural economics - is a fast growing field within the social and behavioural sciences and this is one of the only PhD programmes within Europe to have this area as the primary focus. Completing this PhD involves becoming an interdisciplinary researcher, with advanced research skills in both economics and psychology, as well as an appreciation of how assumptions, methods, and theories differ between the two fields.

The PhD research may involve the analysis of large (N > 10,000) pre-existing longitudinal datasets, quantitative field surveys, experimental designs, randomized controlled trials, or a combination of these methods. Research can be desk based or involve our public, private, or third sector partners (such as the Scottish Government, local council, job centres, or consultancy business) normally with a focus on basic science research.

In addition to the individual supervision and structured training given to all students at Stirling Management School, students benefit from being a full member of the internationally leading Behavioural Science Centre, which has developed a genuine community of closely interacting and collaborating researchers. Members come from diverse backgrounds, with some having degrees exclusively from economics and others exclusively from psychology, but all share the same passion for researching at the interface between these two areas. There is a strong culture of joint socialising (including regular drinks and meals) and collaborating - projects normally have input from several centre members, all of whom are always willing to be a part of joint research. Collaboration and the research culture are promoted with a two hour centre meeting each week. The first 30 minutes is devoted to "business", where everyone is updated on recent relevant developments and participates in shared decision making as to the direction of the centre. The second 30 minutes is slot booked by a centre member (including our students) to use as they want - commonly to put up an early research idea for feedback, results for discussion, or present a preview of a conference talk in a supportive environment. The remaining hour is used for the weekly seminar, which has more substantial talk from centre members or more commonly presentations from invited researchers from across the UK, as well as occasionally from our partners in industry and government. In addition, we run regular workshops (including currently a £28,000 ESRC seminar series) which attract the key figures in the field. PhD students are an equal part of our community and are expected to enthusiastically participate in all aspects of centre life.

Students are part of the Division of Economics, where the centre is based, which runs additional seminars and social events. Depending on the student's background and the precise research, students may be said to graduate with a PhD in behavioural science, social science, economics, psychology or another related area that accurately describes the work and skills set developed.

Potential students are very strongly recommended to carefully review the centre website, blog, and particularly the Centre Director’s informal guide to a PhD, as well as the other pages at the links below. For interesting modules to attend whilst taking the PhD, the MSc in Behavioural Science may be particularly relevant.

Useful links:
The Behavioral Science Centre webpage
Centre Director Professor Alex Wood's Informal PhD Guide
The Centre's blog, details can be found here about the work we do, plus details on our seminars and workshops
The Centre's Twitter Feed
MSc in Behavioural Science
Example thesis from centre member Dr Christopher Boyce
Contact

If you are interested in joining the centre as a PhD student you are recommended to first approach the Director, Professor Alex Wood. You are also welcome to contact at any stage the relevant Postgraduate Research Tutor Dr David Comerford . Other helpful individuals include;

MRes in Business and Management Programme Director – Dr Scott Hurrell.

Administrator – Lisa Reid.

Saturday, October 18, 2014

Life-satisfaction correlates very well with more well established measures of societal progress

In a previous post about life-satisfaction I mentioned that the five most satisfied countries in Europe according to the life-satisfaction measure from the 2012 European Social Survey were also the top five in the 2013 Legatum Prosperity Index, although the five countries were in a different order. I thought that consistency was fairly remarkable considering that the ESS measure presumably took about 5 seconds to administer, whereas the Legatum Index uses 89 different variables to create its ranking. I wondered how well this very simple life-satisfaction measure would correlate with other established measures of societal progress. The results are below.

I used data from several places: (1) life-satisfaction ("How satisfied are you with life as a whole" scored 0-10) and trust in others ("Most people can be trusted or you can't be too careful" scored 0-10; I included this since higher trust correlates with more happiness) from the 2012 European Social Survey, (2) average $ GDP per capita over 2009-13 from the World Bank, (3) life expectancy as of 2013 from the World Health Organization and (4) country rankings from the Legatum Prosperity Index. I used all 27 countries in the EES but excluded Kosovo from the Legatum ranking and life expectancy analysis because I couldn't find data for it. I also excluded a country labelled "IS" because I couldn't find what country this code corresponded to in the ESS data dictionary.


The results are pretty clear: higher life satisfaction correlates with a better ranking on the Legatum Index (R = 0.84), higher GDP per capita (R = 0.81, 0.83 for ln GDP), more trust in others (R = 0.76) and higher life expectancy (R = 0.72). For such a simple measure I find those amazingly strong associations. 





The Stata code I used is below:
use "C:\File Location\ESS6e02.dta", clear
rename idno id
rename cntry c
rename stflife ls
rename ppltrst t
drop if ls > 10 | t > 10 | age == 999  | c == "IS" 
keep id c ls t

*Collapsing the variables by country will automatically convert trust and life-satisfaction scores into their country averages
collapse t ls, by(c)

*Average GDP per capita in $ over 2009-13 taken from World Bank http://data.worldbank.org/indicator/NY.GDP.PCAP.CD
gen gdppc = 4652 if c == "AL"
replace gdppc = 45387 if c == "BE"
replace gdppc = 7296 if c == "BG"
replace gdppc = 80477 if c == "CH"
replace gdppc = 25249 if c == "CY"
replace gdppc = 18861 if c == "CZ"
replace gdppc = 45085 if c == "DE"
replace gdppc = 58894 if c == "DK"
replace gdppc = 18478 if c == "EE"
replace gdppc = 29118 if c == "ES"
replace gdppc = 47219 if c == "FI"
replace gdppc = 41421 if c == "FR"
replace gdppc = 39337 if c == "GB"
replace gdppc = 13134 if c == "HU"
replace gdppc = 47400 if c == "IE"
replace gdppc = 36151 if c == "IL"
replace gdppc = 34619 if c == "IT"
replace gdppc = 15538 if c == "LT"
replace gdppc = 47617 if c == "NL"
replace gdppc = 100819 if c == "NO"
replace gdppc = 13432 if c == "PL"
replace gdppc = 21035 if c == "PT"
replace gdppc = 14612 if c == "RU"
replace gdppc = 58269 if c == "SE"
replace gdppc = 22729 if c == "SI"
replace gdppc = 17689 if c == "SK"
replace gdppc = 3900 if c == "UA"
replace gdppc = 3816 if c == "XK"

gen lngdppc = ln(gdppc)

*Inverted Legatum Prosperity "Europe only" rankings where 1 = lowest, 27 = highest, taken from http://www.prosperity.com/
gen legatumrank = 1 if c == "AL"
replace legatumrank = 18 if c == "BE"
replace legatumrank = 4 if c == "BG"
replace legatumrank = 26 if c == "CH"
replace legatumrank = 9 if c == "CY"
replace legatumrank = 13 if c == "CZ"
replace legatumrank = 20 if c == "DE"
replace legatumrank = 24 if c == "DK"
replace legatumrank = 10 if c == "EE"
replace legatumrank = 16 if c == "ES"
replace legatumrank = 23 if c == "FI"
replace legatumrank = 17 if c == "FR"
replace legatumrank = 19 if c == "GB"
replace legatumrank = 6 if c == "HU"
replace legatumrank = 21 if c == "IE"
replace legatumrank = 7 if c == "IL"
replace legatumrank = 12 if c == "IT"
replace legatumrank = 5 if c == "LT"
replace legatumrank = 22 if c == "NL"
replace legatumrank = 27 if c == "NO"
replace legatumrank = 11 if c == "PL"
replace legatumrank = 14 if c == "PT"
replace legatumrank = 3 if c == "RU"
replace legatumrank = 25 if c == "SE"
replace legatumrank = 15 if c == "SI"
replace legatumrank = 8 if c == "SK"
replace legatumrank = 2 if c == "UA"

*Life expectancy from WHO http://en.wikipedia.org/wiki/List_of_countries_by_life_expectancy
gen lifeexp = 74 if c == "AL"
replace lifeexp = 81 if c == "BE"
replace lifeexp = 74.5 if c == "BG"
replace lifeexp = 82.8 if c == "CH"
replace lifeexp = 81.2 if c == "CY"
replace lifeexp = 78 if c == "CZ"
replace lifeexp = 81 if c == "DE"
replace lifeexp = 79.5 if c == "DK"
replace lifeexp = 76.1 if c == "EE"
replace lifeexp = 82.5 if c == "ES"
replace lifeexp = 79.6 if c == "FI"
replace lifeexp = 82.3 if c == "FR"
replace lifeexp = 81 if c == "GB"
replace lifeexp = 75 if c == "HU"
replace lifeexp = 81.4 if c == "IE"
replace lifeexp = 82.1 if c == "IL"
replace lifeexp = 83.1 if c == "IT"
replace lifeexp = 75.9 if c == "LT"
replace lifeexp = 81.5 if c == "NL"
replace lifeexp = 81.9 if c == "NO"
replace lifeexp = 77.5 if c == "PL"
replace lifeexp = 80 if c == "PT"
replace lifeexp = 70.5 if c == "RU"
replace lifeexp = 83 if c == "SE"
replace lifeexp = 80 if c == "SI"
replace lifeexp = 77 if c == "SK"
replace lifeexp = 71 if c == "UA"

pwcorr ls t lngdp legatum lifeexp, sig

scatter ls legatum, mlabel(c) || lfit ls legatum
scatter ls lngdppc, mlabel(c) || lfit ls lngdppc
scatter ls t, mlabel(c) || lfit ls t
scatter ls lifeexp, mlabel(c) || lfit ls lifeexp

Tuesday, October 14, 2014

Stirling Choice Workshop

Below is the programme for Thursday's workshop on choice organised by Dr. Danny Campbell. There are spaces available in the lecture hall if people wish to attend.  

CHOICE WORKSHOP
16 October, 2014

University of Stirling, IMB Gallery Room, IrisMurdoch Building

10:00–10:30 Welcome

Session 1: Benefit transfer and the environment

10:30 Enhanced geospatial data for meta-analysis and environmental benefit transfer: an application to water quality improvements
Robert Johnston (Clark University)

10:55 Using Bayesian methods for benefit transfer from choice experiments: results fromsimulations
Klaus Glenk (Scotland’s Rural College)

11:20 Impact of a collective bonus on farmers’ participation and acreage enrolment in an agri-environmental scheme
Laure Kuhfuss (University of St. Andrews)

11:45 Does what you know and what we tell you change the value people place on coastal flood defence?
Katherine Simpson (University of Stirling)

12:10–1:15 Lunch

Session 2: Social norms, oaths and mode effects

1:15 Using behavioural economics to improve social norms interventions: rank-based nudging
AlexWood (University of Stirling)

1:40 Social norms, economic incentives and ethical motives in choices for household recycling
Nick Hanley (University of St. Andrews)

2:30 Discrete choice experiment under oath
VerityWatson (University of Aberdeen)

2:05 Mode effects: valuation workshop versus online panel
Erlend Dancke Sandorf (Arctic University of Norway)

2:55–3:15 Tea and coffee

Session 3: Uncertainty and processing

3:15 Preferences analysis under inherent uncertainty
Michela Faccioli (University of St. Andrews)

3:40 It’s not what you ask, but how you ask it! Respondents’ ability to reflect on their use of attribute non-attendance in a DCE
Sebastian Heidenreich (University of Aberdeen)

4:05 The effect of gender on risk perceptions
Seda Erdem (University of Stirling)

4:30 Using eye-tracking methods to informdecision making processes in discrete choice experiments
Mandy Ryan and Nicolas Krucien (University of Aberdeen)

4:55 Position bias in best-worst scaling surveys
Danny Campbell (University of Stirling)

5:20–5:30 Concluding remarks

Job: Senior Programme Manager with BIT Ventures

Our MSc students should keep an eye out for these kind of posts to get a sense of the job market they'll be graduating into next year. Also bookmark this page on be-recruit.com.

Job title: Behavioural Insights Ventures Senior Programme Manager
Salary: £50-60K plus benefits
Term: 2 year contract

Details here & candidate specification here

"Nesta is looking for an exceptional candidate to lead the 'Behavioural Insights Ventures' programme.

This person will channel the world class behavioural science of the Behavioural Insights Team (BIT) into a series of commercial products designed to deliver positive social change. 

To do this they will help design a portfolio of different products, and then on a case by case basis build the right delivery capability to take these products to market – either by building teams internally, partnering with outside organisations or starting new companies.

This role has huge potential to build successful ventures and achieve large-scale social impact.  In order to qualify you will need to be able to demonstrate a successful track record of spotting commercial opportunities, and building high quality teams or partnerships to take products to market.

This is a two year fixed term contract and either full time or part time hours will be considered.

To apply: send your CV, a short covering letter and completed equal opportunities form to jobs@nesta.org.uk.
Closing date for applications: 9am Monday 20 October 2014.
First interviews will take place: Thursday 23 October 2014"

Links 14.10.14

1. Interview with Oliver Payne by the website Science Rock Stars. Oliver is the founder of behavioural insight firm The Hunting Dynasty.

2. "Large company CEOs are roughly +1, +1.5 and +0.5 SD on cognitive ability, non-cognitive ability and height, respectively"

3. A review of Cass Sunstein's books 'Why Nudge' and 'Conspiracy Theories and Other Dangerous Ideas' from the New York Review of Books.

4. On whether American schools should teach their students 'grit' from Vox & Angela Duckworth's TED talk on same.

5. A conversation with Douglas Detterman, the editor of the journal 'Intelligence'.

6. A review of 35 years of emotion and decision-making research from an upcoming issue of the Annual Review of Psychology.

7. A look at inequality focusing on the poor from Aeon magazine.

Monday, October 13, 2014

Regression Models for Categorical Dependent Variables using Stata by J. Scott Long & Jeremy Freese (3rd edition)

Scott Long and Jeremy Freese have released the third edition of their book "Regression Models for Categorical Dependent Variables using Stata", their first update since the 2nd edition came out in 2006 (here is a review of that edition by Richard Williams). Many thanks to Timberlake and Stata Press for sending the blog an advance copy. It is a beautiful piece of work which I will be using as my main Stata reference for the foreseeable future.

The new edition is a hefty 589 pages, a significant increase from the 311 pages in the 2nd edition. The book focuses on categorical outcome variables[1]or outcomes with two or more possible values. These kind of outcomes require non-linear models to properly analyze, such as Probit, Logit, negative binomial or Multinominal Probit/Logit, rather than the basic OLS which is used for linear models. When dealing with non-linear models, "the simple interpretations that are possible in linear models are [not] appropriate... Because of this nonlinearity, no method of interpretation can fully describe the relationships among the independent variables and the outcomes. Rather, a series of postestimation explorations are needed to uncover the most important aspects of these relationships. If you limit your interpretations to the standard output of estimated slope coefficients, your interpretation will usually be incomplete and sometimes even misleading" [p7]. 

The book's index is shown in Fig. 1 and described further here.
Fig 1. Index
Part I begins with a concise introduction to Stata and a review of the fundamentals of model estimation, making the book relatively accessible for researchers unfamiliar with Stata. Part II describes how to estimate and interpret binary, ordinal, nominal, and count outcomes[2]. Throughout the text the authors use clear language and many practical examples to emphasize the intuition behind the various analytic techniques without a lot of dense mathematics.

Although the book is a good introduction to Stata, the real added value comes from the enormous level of detail the authors devote to describing how to interpret (and graph) regression results using the margins command[3] and a suite of supplemental post-estimation commands created by the authors called mgen, mchange and mtable (these latter three commands are collectively called SPost13 and replace the popular earlier suite SPost 9; see here for an explanation of why the authors recommend using SPost13 rather than margins). Chapter 4 is devoted entirely to describing these post-estimation commands and chapters 5-9 contain many example applications of them for different kinds of models. The authors also provide free example data-sets and code to practise these commands on, which can be downloaded by following the instructions in the book [p13].

The authors highlight that margins excels at four things in particular [p137], all of which are described in further detail here. These are:
(1) Predictions for each observation
Margins can predict the probability of an outcome for each person in the data, taking into account all the covariates included in the regression. The predict command can do this, but margins also provides standard errors and confidence intervals.
(2) Predictions at specified values
Margins can compute the probability of an outcome at specific values of covariates while holding the others constant or at their means (or at any other value).
(3) Marginal effects
Margins can compute how changes in a covariate are associated with changes in an outcome variable, holding other covariates constant.
(4) Graphs of predictions
Marginsplot can easily graph outcome variables based on margins. Commands such as marginsplot and Ben Jann's coefplot are particularly good at turning potentially obtuse interaction coefficients into intuitive graphs such as (i) age*age (ii) gender*race (see p43) (iii) age*race (see p24-25).

This book is ideal for graduate students (like myself), who may wish to work their way through Chapters 1-4 in detail. More experienced researchers could start with Chapters 4-6, which really unpack the margins command and methods of testing and interpreting non-linear coefficients. Although the title of the book focuses on categorical dependent variables, there is plenty for researchers working with continuous outcomes to learn from it given how useful the margins and SPost13 commands are for clarifying all manner of results.

I'll be using this book as a reference for a series of posts over the next few months which further explore the capabilities of margins, SPost13 and marginsplot. These will be available under the 'Stata resources' tab on the left hand side.

[1] Categorical variables are distinct from continuous variables because they don't have any intrinsic ordering. Examples include (0 = employed, 1 = unemployed, 2 = in education) or (0 = smoker, 1 = not a smoker). An example of a continuous variable is a measure of intelligence where the scores range from low to high along a common scale (i.e. 1-100).
[2] Binary outcomes have two values such as whether a person is healthy or sick. Ordinal outcomes have more than two categories that are assumed to be ordered on a single, underlying dimension such as the answers to a survey question ranging from "Strongly disagree" -  "Disagree" - "Neither agree nor disagree" - "Agree" - "Strongly agree". Nominal outcomes have more than two categories but the categories are not ordered, such as whether a person travels to work by car, train, bus or foot. Count variables count the number of times something has happened, such as the number of months a person has been unemployed for, or the number of articles written by a scientist [all definitions taken from p8].
[3] Until now, my own knowledge of the margins command has been gleaned from online guides and presentations, so it is nice to finally have an authoritative text to refer to.

Examples of the 4 main capabilities of the 'margins' command in Stata in use with linear models

The below gives examples of the four main capabilities of the margins command in Stata, as highlighted by J. Scott Long & Jeremy Freese in the 3rd edition of "Regression Models for Categorical Dependent Variables using Stata".

I used data from the 2012 edition of the European Social Survey to illustrate these capabilities. To simplify things, I used only four variables for 1,714 individuals from Great Britain. The main outcome variable is 'ls' which is life-satisfaction scored from 0 = Extremely dissatisfied to 10 = Extremely satisfied. Gender is coded 1 = male, 2 = female. Income is grouped by decile where 1 = lowest income decile and 10 = highest income decile. There is also an ID variable.


(1) Predictions for each observation
Margins can predict the probability of an outcome for each person in the data, taking into account all the covariates included in the regression. The 'predict' command can also do this.

reg ls i.gender income
margins gender




























The average predicted life satisfaction is 7.309 after controlling for gender and income. The same thing can be computed using the predict command:

reg ls i.gender income
predict predicted_ls
sum predicted_ls










The predict command produces the same predicted life satisfaction score of 7.309, but margins also provides standard errors and confidence intervals

(2) Predictions at specified values
Margins can compute the probability of an outcome at specific values of covariates. The below code instructs Stata to compute predicted life satisfaction at 1 unit increments of the ten deciles of income, starting from 1 and stopping at 10, while controlling for gender. The 'vsquish' option is just an aesthetic change in how the results are presented and can be ignored.

reg ls i.gender income
margins, at(income=(1(1)10)) vsquish

Note: I could have instead used "margins, at(income=(1 2 3 4 5 6 7 8 9 10))".



























Predicted life satisfaction scores range from 6.82 for the bottom income decile to 7.87 for the top decile after controlling for gender.

(3) Marginal effects
Margins can compute how changes in a covariate are associated with changes in the outcome, holding other covariates constant. In other words, it can compute marginal effects.

reg ls i.gender income
margins, dydx(gender)

















Women have -0.61 lower predicted life-satisfaction scores than men after controlling for income, although the difference is not significant. Because this is a linear OLS model, this same result could have been calculated by looking at the gender coefficient in the main regression. The real value of this command is when estimating marginal effects in a non-linear model such as a Probit, which doesn't return intuitive coefficients as a default.

(4) Graphs of predictions
Lastly, margins can be easily combined with marginsplot to graph its predictions, with 95% confidence intervals included as a default.
reg ls i.gender income
margins, at(income=(1(1)10))
marginsplot























It's also very easy to graph interaction terms using this method. The below code examines whether the relationship between income and life-satisfaction differs by gender.

reg ls i.gender##c.income
margins gender, at(income=(1(1)10))
marginsplot






















This is part of a series of posts designed to highlight the usefulness of the margins command & various graphing capabilities in Stata. It draws on the 3rd edition of "Regression Models for Categorical Dependent Variables using Stata" by Long & Freese as a primary reference.