[20° Forum SIS] Biometrical Journal, Special Issue on "Models for continuous data with a spike at zero"

[Marco Alfo']

###APOLOGIES FOR CROSS_POSTING####

The Biometrical Journal is a leading journal in biometry. It is
calling for papers for a special issue on "Models for continuous data
with a spike at zero".

Zero-inflation models for count data have a long history and a huge
amount of published research exists to the topic. In fact, the
Biometrical Journal has contributed numerous papers to this area since
zero-inflated count data are a prominent problem in the life-sciences.
See, for example, the September 2013 issue of the Biometrical Journal.
In this call, we would like to draw the attention of the research
community to another prominent problem in the life sciences:
continuous data with a spike at zero. The extra-zeros can arise out of
many reasons and it is not necessarily clear how they could be
handled. A frequent problem is the occurrence of data below a limit of
detection (LOD). In these cases, it might be reasonable to treat them
as left-censored. A further modeling alternative could be that of
employing so called two-step models, where at the first stage a binary
variable describes whether or not the observed value is above the
limit of detection while, at the second stage, a continuous variable
is used to describe positive measurements. In other cases, it is less
clear how to approach the problem. For example, consider the
consumption of alcohol measured on a continuous scale. Then, there is
a spike at zero representing the non-drinkers. The zeros are neither
left-censored nor interval-censored.  Here, there are at least two
problems involved. For one, the quantity of interest is bounded below,
and for two, the zeros create a single mass distribution.  A further
empirical application where extra-zeros may arise is when modeling
length of stay in hospital admissions. This is frequently modeled
following a count data approach, although the underlying quantity is
continuous.  In length-of-stay situations, day surgery patients may be
classified as having a null length of stay but, for administrative
purposes, it could be necessary for them to be considered together
with in-hospital patients whose length of stay is positive. Hence, we
have a continuous quantity bounded below by zero and a spike at zero.
In a regression context, we also think that these spikes could occur
on the left-hand side of a regression equation as well as on the
right-hand side, i.e. for the dependent (outcome) and the independent
(predictor) variables

With  this call for submissions we are looking for novel contributions
presenting
•	authentic continuous data with  excess-zeros
•	novel statistical approaches to handle these issues.
•	up-to-date case studies which address the above issues  and are of
interest for the biometric community.
.
Authors, who are not sure whether their problems and ideas fall under
the call or not, may approach the Editors with an initial draft of the
paper.

All submitted papers under this call will undergo the standard review
process of the journal. The deadline for papers for this call is May
31, 2014.

The special issue is handled through the website of the Biometrical
Journal and is handled as any other paper submitted to the journal
except that the editing and the review process is handled by the
Special Issue Editors  Dankmar Böhning and Marco Alfó.

Please submit your manuscript at is http://mc.manuscriptcentral.com/biomj


To submit to the special issue login at MC (manuscript central),
indicate at Step 1 that the submission is to a special issue and at
Step 4 of the submission process enter "ZERO I" in the Special Issue
Information field.

Submitted manuscripts should comply with the instructions for authors
of Biometrical Journal which you find on the Biometrical Journal
homepage using the link to the Author Guidelines.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20131213/19cdcf61/attachment.html>