# Submission for Habilitation at Jacobs University

I submitted for habilitation in Computational Social Science at Jacobs University Bremen with a cumulative habilitation work on

## Systemic Effects in Models of Opinion Dynamics, Societal Growth, and the Wisdom of Crowds

This (click the title above for download) is an 88 page booklet presenting the results of twelve published papers focusing on systemic effects in opinion dynamics and societal growth and on the wisdom of crowd phenomenon in group guesstimation games. Read a short summary with a focus on the new results here.  Continue reading

# Meandian

In this post (in German) I demonstrated the superiority of the median over the mean in aggregating a good estimate for a guessing question which answers are bounded by zero but have no upper limit. I speculated further that the correct value often lies slightly above the median. John Doyle recently invented the meandian (see Meandian: A Measure of Location Based on Signed Rank of Deviations) as a measure in between the mean and the median. As a consequence the meandian is typically slightly larger than the median. Doyle showed on our PNAS dataset that the meandian indeed delivers as good results as the median and often better ones.

As I really liked the idea of the meandian, I quickly checked the measure on my “Weisheit der Bremer”-dataset and found:
Truth = 10788, Median = 9843, Meandian = 12170.
Deviations: |Median-Truth| = 945, |Meandian-Truth| = 1382.
Conclusion: Meandian is also good, but the Median wins in this case.
By the way: All are beaten by the geometric mean:
Geomean = 10510 (Deviation |Geomean-Truth| = 278)

# Ministers’ incumbency ends after 1.79 failures with daily failing probability 0.11%

Today, Thomas Metz made me aware of a dataset about ministers in Eastern German federal states (Bundesländer) by Sebastian Jäckle. The dataset includes the variable “duration of incumbency” in days for 291 ministers between 1990 and 2011.

I was curious to look at the distribution of duration with the intention to be brave as a physicist and infer a simple stochastic model which reproduces that distribution. I copied the duration data into a matlab vector duration, made histograms, fits for different distributions and KS-Tests. As duration is a discrete random variable (days starting from inauguration), distributions living on the nonnegative integers are the natural candidates. The classical one-parameter distributions Poisson and geometric failed to deliver fitting distributions, but the negative binomial (NB) did surprisingly well.

Normalized histogram for duration of incumbency and best-fit negative binomial probabilty mass function.

The best fit yielded parameters $r=1.79$ and $p=0.0011$. The Kolmogorov-Smirnov test did not reject that duration data came from the distribution with these parameters (p=0.32), but rejects under reasonably small changes of the two parameters. Thus, it is reasonable to assume

$\text{duration}\sim\text{NB}(1.79,0.0011)=\text{NB}(r,p).$

What model does this imply? Looking at the days in the incumbency of a minister. Let us assume that every day can either be a success or failure which happens with probability $p$. The negative binomial is the distribution of the number of successful days until $r$ failures occur (there is an extension to non-integer number of failures). Our model is thus, that a minister’s incumbency ends after a certain number of failures (what ever that means in practice). The best fit suggests that under this model 1.79 failures are allowed during a minister’s incumbency and that failures are relatively rare events happening with probabilty 0.11% every day, i.e. on average the first failure happens approximately at day 900.

# Posters vs. Talks

This is my view on organization of scientific conferences, especially the mid-size and large ones. In my opinion poster sessions are a much more efficient way of scientific communication than a bunch of 20 minutes talks. I advocate for upgrading the scientific prestige posters deliver to the author to be larger than the prestige of short talks.

# Dissent due to more links

I watched Eli Pariser’s TED talk on increasing bi-polarization of viewpoints due to automatic prioritizing of Facebook news feed items.

I think this effect is related and enforced by what is studied in a model by J.K. Shin and me in Tipping Diffusivity in Information Accumulation Systems: More Links, less Consensus (Shin & Lorenz, JSTAT 2010, arXiv).

# Body connected socket to charge cell phone

The cell phone gets more and more important for life, including critical issues such as long and short term memory, orientation in space, and payment information. Loss of battery power can thus result not only in loss of remote communication ability but also more critical things such as no access to memorized information, getting lost in space or being short on cash. Probably, dependence on mobile devices will get even more serious in the future. Thus, it can be as important to get battery power than for example to get food, shelter or sleep. A body power point (e.g. a socket) which extracts electricity out of body energy would be a necessary invention. At best it would be something biochemical such that energy is accessible also when the physical ability is already very limited.

A possible nice side effect:
Imagine you burn some calories to recharge your mobile.
I appreciate any technical information how to implement such a charge power point on human skin.

PS: The issue was also touched at xkcd.com but with an emphasis on software issues not electricity.

# Scientific Citation Markup

Citations in scientific papers are the basic ingredient to compute impact factors and eigenfactors of scientific journals. The number of citations a publications receives in the course of time is the most neutral and accepted criteria for its relevance. Thus, the number of citations a paper receives is very important for the careers of its author.

Everybody who reads or writes scientific papers knows that a citation may have very different meanings which do not always coincide with the interpretation of “pointing to relevance”, as implied by the above uses. Further on, every scientist knows that the motivations of citing a paper are not always driven by pure scientific reasons.

Ideally, all these differences in citations could be incorporated in a scientific citation markup like \cite[markup]{RefKeyForPaper} (in LaTeX citation style) or the like.
Some ideas:

\cite[negative]{PaperWithSevereErrors}
\cite[community feeling]{PopularPaperWithNoSpecificRelation}
\cite[please journal editor]{AnyPaperOfEditor}
\cite[enforced by a referee]{SuggestedPaper}
\cite[proof or evidence elsewhere]{TechnicalPaper}
(The latter can be used to (i) avoid redundancy in the literature, (ii) save space, or (iii) to obfuscate that there is none. (iii) is prominent in physics as a citation with a reference to an own paper which is “to appear”)

Imagine, what more sophisticated profiles of papers, journals and scientists one could create with it …

# Credit Ratings as Unwanted Aggregation

Just a quick thought: Shouldn’t credit rating agencies be prohibited, just for the same reason why price-fixing agreements are prohibited in a free market?
The credit rating agencies correlate investment decisions and thus diminish the power of markets, when we see a market as a mechanism to aggregate knowledge about economic performance.

The agencies need not do worse than the market itself would do, but still there is a negative effect in the correlation. The job of the credit rating agency is to assign the credit risk of firms (or states). This can also be done by the market, where creditors would than have to decide on the offers (e.g. of interest rates) on their own. Than the market would also find some quantification of the credit risk. If we believe in markets than the information about the credit risk should not be pre-aggregated.