YourMorals Blog

The shutdown impasse has led to a great deal of speculation about the underlying causes of governmental dysfunction in the United States.

It has become clear that the parties have separated on moral issues. As politics has become more about engaging the mass public (as opposed to an earlier era of patronage politics), the parties have adopted increasingly divergent moral positions.

Moral diversity within parties

But the picture isn’t so clear cut as liberals on one side and conservatives on the other. While most metrics do show a better sorting of members of Congress and the mass public into parties by ideology (conservatives are much more likely to also be Republicans and liberals are much more likely to be Democrats today than has been the case historically). There still exists a great deal of heterogeneity within each party.

The shutdown fight is an interesting case of intraparty disagreement. It seems clear that Speaker Boehner would have preferred to keep the budget negotiations on the subject of entitlement reform and other long-term Republican goals. However, conflicts within the Republican Party have put him into a difficult position. The (very) conservative wing of his party has made the current imbroglio about health care.

In other work, I have generated estimates of the moral foundations at the Congressional District level. Does variation in the moral foundations of Republicans’ constituents help us to explain the current conflict?

The poison pill

One way to gauge support for the shutdown is to examine the cosponsors of Rep. Tom Graves’s Continuing Resolution (CR). Graves included language in his version of the budget resolution that would ‘defund’ the Affordable Care Act – a non-starter for the administration and senate Democrats and by all accounts contrary to Boehner’s wishes. Which moral foundations correlate with support for Graves’s CR among Republicans?

One-hundred and thirty-six (nearly 60%) House Republicans cosponsored Graves’s CR.

It turns out that even after controlling for Obama’s share of the vote in their districts (which explains a very significant portion of the variation), Republicans who represent districts that are particularly low on the Care/Harm foundation are most likely to support the CR.

The relationship (holding all of the other foundations and presidential support in the district at their mean values) [1] looks like this:

The plot shows the predicted probability of supporting the CR across the range of scores for the Harm/Care foundation in Republican districts. The points at the top and bottom of the plot show the actual positions of each district. The points at the top show the districts that cosponsored the CR and the points at the bottom show those that did not cosponsor it.

According to my model, a legislator who represents a district at the mean levels of each of the foundations and the average for Obama’s vote share in Republican districts is 60 percent likely to support the CR. An increase of one standard deviation in the Care/Harm foundation lowers the probability by a little over 10% making it dead even. A decrease in one standard deviation in the Care/Harm foundation increases the probability of cosponsoring the CR by about the same amount to 70 percent.

Given two districts that are equally Republican (gave the same proportion of the vote to Obama in 2012), variation in the Care/Harm foundation explains some of the remaining variance in position taking on the CR.

We can think about this finding in two ways. First, it might be that Republicans who represent individuals who place more emphasis on the morality of care and harm are less likely to go on the record as wanting to take away health coverage from those who will be added under the president’s new initiative. Perhaps their constituents would punish them for perceived cold-heartedness.

On the other hand, it could be that the effect is mainly driven by the fact that many Republicans represent districts whose constituents do not place particularly high emphasis on concerns about care and harm to individuals. Their mix of moral concerns tends to direct their attention toward the (perceived or actual) loss of individual freedom that is implied by compelling individuals to receive health coverage. In the weighing of costs and benefits of the new proposal, there is obviously a deep moral divide between the two sides.

Given that this is a single action at a single point in time, it is difficult to say definitively one way or the other. As with most things, the correct answer is surely some mix of the two explanations (along with a host of other unaccounted for factors).

Discussion

The above analysis suggests that some of the conflict within the Republican Party can be explained by variation in the moral concerns of constituents. There is nothing inevitable or foreordained about the relationship outlined above, but it does perhaps tell us something about a way forward.

For the most strident parts of the opposition, arguments focusing on the harm being done by the shutdown aren’t likely to be very persuasive. Interestingly, there is some suggestion in my analysis that the Loyalty/Betrayal foundation is negatively associated with support for the CR (although the effect did not reach conventional levels of statistical significance). It is possible that framing the debate in terms of patriotism (our national obligation to pay our bills) or even good-old partisanship (the need of the party to stick together behind Speaker Boehner) would be more successful.

---

[1] I estimated a probit regression with cosponsorship of the CR as the dependent variable. The model included the five moral foundations and Obama’s share of the vote in the 2012 election as independent variables.

Recently, I received news that Reid Hoffman, founder of LinkedIn, has decided to indirectly support some of the work I do with Jonathan Haidt, because he is a fan of Jon’s work.  The news got me thinking a bit about the confluence of moral psychology and what LinkedIn does.  As Ranker’s Principal Data Scientist, I will often find myself at data conferences with members of LinkedIn’s data team, which is one of the most visible and productive, given the wealth of data they have to mine.  It may seem that what they do and what moral psychologists do are separate worlds, but as I argued in my 2012 SXSW talk, there is likely to be a convergence.  Big data is merely a tool that inevitably will be used to answer the questions we care about, and for those of us with first world problems where crime is going down, war is less of an issue, and obesity is a bigger problem than hunger, the questions we increasingly ask tend to be more existential.  It is only natural that we use the tools of the age (data) to answer the questions of the age, like how we can live a meaningful life.

Unlike some who are more dogmatic in their belief, I don’t believe that data can tell you what a meaningful life is or how you should live.  In short, moral psychology cannot tell you what ought to be…but it can tell you what is.  For example, in a partnership with Zenzi Communications, we have been working on describing the narratives that people of different value types tend to resonate with.  I can’t tell you whether action or comedy movies are better, but I can provide a probabilistic fit between a person’s value orientation and the kinds of stories that they are likely to enjoy.

Similarly, companies and prospective employees are looking for such fits as well.  Zappos wants people whose values tend toward the unconventional.  LinkedIn wants people who will put their members’ first.  In moral psychology terms, Zappos is looking for individuals who are high on traits like openness to experience and stimulation, while LinkedIn might be looking for individuals who are more self-transcendent in terms of putting others (their members) first.

These values are measurable both with self-characterizations and endorsements, just like LinkedIn skills, but also through more subtle means such as music tastes or use of certain words.  Further, following on work by Shalom Schwartz, values can be organized as oppositional (see above graph using Zenzi’s value typology which is based on Schwartz’s work), so that self-presentation effects (e.g. “I’m high on all values!”) can be mitigated.  In this way, companies and employees may potentially be able to find employment matches that are not just a job, or even a career, but are closer to a calling, leading to happier and more productive employees.

We get a lot of resumes at Ranker.  They honestly all start to blend together at some point.  The employment market needs more nuance, more subtlety, and more data to help create better matches.  I don’t know if LinkedIn or some other organization will lead the charge, but I’m confident that this type of values based matching will indeed be part of the future of human resources recruiting, and it already is part of many forward thinking organizations’ processes.  Just as the internet and information technology has made matching in other domains (e.g. classified ads or home buying) more efficient, values based matching in the labor market is bound to be a problem tackled by data science teams at online companies in the near future.

– Ravi Iyer

Very few days go by without a new article describing the limits of published scientific research.  The headline cases are about scientists who plagarize or completely fabricate data.  Yet, in my experience, most scientists are actually quite ethical, meticulous, hard-working, and really concerned with finding the truth.  Still, non-scientists would likely be surprised to know that a large number of scientific studies are actually false.  An Amgen study found that 46 out of 53 studies with ¡®landmark¡¯ findings were unable to be replicated.  A team at Bayer found a slightly more optimistic picture where 43 out of 65 studies revealed inconsistencies when tested independently.  Scientific journals continue to accept articles based on the novelty and projected impact of the submission, yet simulations illustrate how the bias of journals toward publishing novel results likely leads to an environment where most published results are actually false.  My home discipline of psychology is currently doing some soul searching as it¡¯s a relatively open secret that many results are difficult to reproduce such that a systematic reproducibility project is taking place.

Crowdsourcing is, and always has been the solution.  Indeed, the phrase at the bottom of Google Scholar, ¡°standing on the shoulders of giants¡±, acknowledges that science has always been about crowdsourcing, as every scholar is collaborating with the scholars before them.  Findings are not produced in a vacuum and build upon (or challenge) previous findings.  Replication by others, which effectively crowdsources verification of results, is at the heart of the scientific method.  It is perhaps a sign of the narcissism of our age that scientists feel compelled to believe that they discover things largely independently, such that they feel compelled to attack when their findings are challenged.  Yet a willingness to be wrong about something is essential to learning, as we can¡¯t learn to walk without falling or learn about relationships without heartbreak.  When science becomes more about ego, career, and grant money, it naturally becomes less accurate.  Insisting that findings be crowdsourced solves this.  No single study, paper, or research group can prove anything by themselves.

Crowdsourcing is not simply averaging the opinions of the masses, as those who would argue against that straw man would have you believe.  Mathematically, crowdsourcing is about reducing the influence of sources of error and there is a great deal of academic research on this topic.  A good crowdsourcing algorithm does not weight all inputs equally, but instead seeks to identify clustered sources of error, which explains why aggregating across people with diverse personalities, perspectives or job functions produces better results.  Inputs need to have some signal vs. noise and need to have uncorrelated error.  The unfortunate assumption in most research is that error is uncorrelated statistical noise that can be dealt with using statistical tests.  Yet error also occurs due to the unconscious biases of researchers, the sheer number of researchers trying to find novel findings, the degrees of freedom that a researcher has in trying to prove their hypothesis, the non-randomness of sampling, and the volume of available statistical tests that a researcher can use.  Given all these other sources of error, it is no wonder that many findings are false.  A good crowdsourcing algorithm would be weighted such that true results would have to be shown by multiple researchers using multiple methods, multiple samples, multiple statistical tests, and multiple paradigms.  This requires crowdsourcing as no single person can do all this, and even if they could, they would still represent a single source of error.

Technology enables crowdsourcing to be conducted far more efficiently, as has been proven by successful science crowdsourcing projects like GalaxyZoo, FoldIt, Seti@Home, and psychology¡¯s reproducibility project.  Trends like citizen science, the quantified self, open access publishing, and interdisciplinarity improve the diversity of perspectives which mathematically improves the ability to find truth.  Every meta-analysis result and Nate Silver¡¯s success in aggregating polls in the last election take advantage of the mathematical principles that underlie crowdsourcing, specifically the certainty that aggregating across sources of error produces more truth.  In our daily lives, we all crowdsource knowledge that we are uncertain about, looking for confirmation from multiple independent sources when we are skeptical.  This same skepticism serves scientists well and scientists should embrace being wrong, confident that the broader truth will be revealed when all data is aggregated intelligently and all perspectives are valued.  Crowdsourcing is not some new technique that threatens to fundamentally change scientific research.  Rather, it is an extension of the collective effort of knowledge aggregation that is the heart of science and scientists should embrace it as such.

• norwaycasino
• swedencasino
• casinonor
• svenskacasino
• casinonorge
• Svenskacasino
• nonorge
• HelenaBremer

– Ravi Iyer

Part of my job at Ranker is to talk to other companies about our data. While people often talk about how “big data” is revolutionizing everything, the reality of the data marketplace is that it still largely revolves around sales, marketing, and advertising. Huge infrastructures exist to make sure that the most optimal ad for the right product gets to the right person, leveraging as much data as possible. For example, I recently presented at a data conference at the Westin St. Francis in San Francisco, which meant that I spent some time on their website. For the past few weeks, long after the conference, I’ve been getting ads specifically for the Westin St. Francis on various websites. At some level, this is an impressive use of data, but at another level, it’s a failure, as I’m no longer in the market for a hotel room. The data to solve this problem is out there as someone could have tracked my visitation of the conference website, understood the date of the conference, and better understood my intent in visiting the Westin. However, this level of analysis doesn’t scale well for an ad that costs pennies, and so nobody does this level of behavioral targeting.

I bring up this story because I believe this illustrates a difference between how people who think of themselves as businesspeople and people who think of themselves as technologists often think. When talking about Ranker data, I often see this dichotomy. People who are more traditionally business minded want a clear business reason to use data, while people who think of themselves as technologists seem more open to trying to envision a world where data does all sorts of neat things that data should be used for. For example, I recently graphed opinions about beer, illustrating that Miller Lite drinkers were closer to Guinness drinkers than to Chimay drinkers. As a technologist, I’m certain that a world will soon exist where bartenders can use data about me and others like me (e.g. the beer graph), to recommend a beer. I don’t worry as much about the immediate path from the conception of such data to monetization. I know that the beer graph should exist and I’m happy to help contribute to it, confident of my vision of the future.

This division between people who think like businesspeople and people who think like technologists is important for anyone who does business development or business to business sales, especially for those of us in the technology world where the lines are often blurry. Mark Zuckerberg is a CEO, but clearly he thinks like a technologist. My guess is that a lot of the CTOs of big companies actually think more like businesspeople than technologists. If I were trying to sell Mark Zuckerberg on something, I would try to sell him on how whatever I was offering could make a huge difference to something he cared about. I would sell the dream. But if I were selling a more traditional businessperson, I would try to sell the benefits versus the costs. I would have a detailed plan and sell the details.

I actually have a bit of data from YourMorals.org to support this assertion. We have started collecting data on visitors’ professions and below I compare businesspeople to technologists on two of the Big Five personality dimensions that are said to underlie much of personality: Conscientiousness and Openness to Experience. As you can see, businesspeople are more conscientious (detail oriented, fastidious, responsible), while technologists score higher on openness which is indicative of enjoying exploring new ideas and thinking of new possibilities.

The reality is that every business needs a balance between those who are detail oriented and precise (Conscientious) and those who think about a vision for the future (Openness to Experience). Often, technologists who start a company will eventually hire professional businesspeople who provide this balance (e.g. Sheryl Sandberg or Eric Schmidt). Clearly, the best sales pitch will be both detailed and forward thinking. However, if you’re talking to someone and have limited time and attention, considering whether you are speaking to someone who is more of a businessperson or more of a technologist may give you better insight into how to frame your pitch.

– Ravi Iyer

• pharmaciegenerique
• pharmaciegenerique
• frau2
• miesten
• mann
• andorra

ps. Crossposted on Zenzi Communications‘ blog here, which is using a data driven approach to improving communications strategies.