Matthew Harding and Carlos Lamarche, "Penalized Quantile Regression
with Semiparametric Correlated Effects: An Application with
Heterogeneous Preferences," Journal of Applied Econometrics, Vol. 32,
No. 2, 2017, pp. 342-358.

The data used in this paper are all Nielsen Licensed Materials
provided by the United States Department of Agriculture (USDA) and
shared with the authors as collaborators.  The authors may not host or
distribute the Licensed Materials through the Internet or any other
medium. Hence the data cannot be lodged here.  The data are held at
Duke University by Matthew Harding, to whom enquiries concerning
access should be addressed. 

In this paper, we restrict attention to single-person households. The
sample includes 6326 female consumers and 2839 male consumers observed
during 12 months in 2010. The data records all purchases at the UPC
level for a large sample of nationally representative households.
Purchases are made in a variety of supermarkets and grocery stores and
are meant for at-home consumption. Each purchased product is uniquely
identified through its Universal Product Code (UPC), a barcode, which
is scanned at the point of sale.

We focus on six of the most essential (and salient) nutrients consumed
per month: total fat, salt, sugar, cholesterol, protein, and
carbohydrates. The data include the gender, race, income, and education
of the head of household.

Matthew Harding
matthew.harding [AT] duke.edu