about illusions

Andrew Gelman talked about a really old paper I did together with Anton Kühberger ages ago. It was actually the first paper / ‘real’ scientific project I was involved in.

It generated quite the buzz over its 20 year lifespan and was cited a whopping 13 times (stats look good without y-axis) …

Stats

 

 

 

Going back to it, I was happy to see that we already talked about replication (and were very reluctant to push the button harder – as we would have not been able to get through the reviews, I guess) … Things have changed.

all that mutate() and summarise() beauty

The friendly people from RStudio recently started a webinar series with talks on the following topics (among others):

Data wrangling with R and RStudio
The Grammar and Graphics of Data Science (both dplyr happiness)
RStudio and Shiny

… and many more.

Our friend Dr. Nathaniel D. Philipps also started a cool R course with videos, shiny apps and many other new goodies.

 

 

When something old …

Schulte-Mecklenbeck, M., & Kühberger, A. (2014). Out of sight – out of mind? Information acquisition patterns in risky choice framing. Polish Psychological Bulletin, 45, 21–28.

I teamed up with Anton Kühberger to write about one of our old, favorite topics: framing and process tracing …

Here is the abstract:
We investigate whether risky choice framing, i.e., the preference of a sure over an equivalent risky option when choosing among gains, and the reverse when choosing among losses, depends on redundancy and density of information available in a task. Redundancy, the saliency of missing information, and density, the description of options in one or multiple chunks, was manipulated in a matrix setup presented in MouselabWeb. On the choice level we found a framing effect only in setups with non-redundant information. On the process level outcomes attracted more acquisitions than probabilities, irrespective of redundancy. A dissociation between acquisition behavior and choice calls for a critical discussion of the limits of process-tracing measures for understanding and predicting choices in decision making tasks.

New paper on food choice and simple heuristics

We got a new paper out on how people (consumers) use simple rules to make food choices. This is work in collaboration with the Nestlé Research Center in Lausanne.
Here is the reference:

Schulte-Mecklenbeck, M., Sohn, M., Bellis, E., Martin, N., & Hertwig, R. (2013). A Lack of Appetite for Information and Computation: Simple Heuristics in Food Choice. Appetite, 71, 242–251.

Abstract

The predominant, but largely untested, assumption in research on food choice is that people obey the classic commandments of rational behavior: they carefully look up every piece of relevant information, weight each piece according to subjective importance, and then combine them into a judgment or choice. In real world situations, however, the available time, motivation, and computational resources may simply not suffice to keep these commandments. Indeed, there is a large body of research suggesting that human choice is often better accommodated by heuristics—simple rules that enable decision making on the basis of a few, but important, pieces of information. We investigated the prevalence of such heuristics in a computerized experiment that engaged participants in a series of choices between two lunch dishes. Employing MouselabWeb, a process-tracing technique, we found that simple heuristics described an overwhelmingly large proportion of choices, whereas strategies traditionally deemed rational were barely apparent in our data. Replicating previous findings, we also observed that visual stimulus segments received a much larger proportion of attention than any nutritional values did. Our results suggest that, consistent with human behavior in other domains, people make their food choices on the basis of simple and informationally frugal heuristics.

Limesurvey randomizing

It is kind of an odd problem.
For the following pretty straight forward question: How do I randomise questions within a group in Limesurvey? It seems to be really hard to find an answer.

With the help of Jonas I figured out that there is a randomisation option hidden in the ‘Advanced Settings’ section of a question. What you have to do is provide the same number (this is important) for each question in a group that you want to randomise. Limesurvey will then take care of the rest. It did not seem to work if I named the variable with a string, eg, ‘group1’ but only numeric counters work fine.

Thanks Jonas! (I would not have finished at all) …

Screen Shot 2013-06-03 at 9.37.52 PM

*apply in all its variations …

Here is an excellent stackoverflow post on how *apply in all its variations can be used.
One of the followups points at plyr (from demi-R-god Hadley Wickham) which provides a consistent naming convention for all the *apply variations. I like plyr a lot, because like ggplot, it is easy to grasp and relatively intuitive to find an answer to even tricky problems.

Here is the translation from *apply to plyr …

Base function   Input   Output   plyr function 
---------------------------------------
aggregate        d       d       ddply + colwise 
apply            a       a/l     aaply / alply 
by               d       l       dlply 
lapply           l       l       llply  
mapply           a       a/l     maply / mlply 
replicate        r       a/l     raply / rlply 
sapply           l       a       laply