listserv message detail


 

N. Peter Whitehead | Apr, 18 2017

 

A new source for science funding data

Good morning SciSIP:I may be late to the parade, but I just discovered Steve Balmers pet project to track all public spending and present the data in usable form.  It may (I have not yet tried myself) be relevant to discussions on government spending on science.  Link here:http://usafacts.org/Best,
N. Peter Whitehead, PhD, ME2npw2w@virginia.edunpwhitehead@mac.com

David Wojick | Apr, 18 2017

 

Re: A new source for science funding data

 

View Original Post

Under spending is says no data available yet, at least on my iPad.I just found a big black box -- research to support climate change adaptation. This is local and regional in nature, past climate and future projections. There is a lot going on, especially in DOD, but OMB says it is not reported separately so there is no way to know how much. Could be a big policy issue.DavidInside Public Access

Susan Fitzpatrick | Apr, 18 2017

 

Re: A new source for science funding data

 

View Original Post

Fascinating.   One quick look – family expenditures – health is the highest, by quite a bit.    Susan M. Fitzpatrick, Ph.D.President, James S. McDonnell Foundation Visit JSMF forum on academic issues: www.jsmf.org/clothing-the-emperor SMF blog  www.scientificphilanthropy.com      

C. Scott Dempwolf | Apr, 24 2017

 

Re: A new source for science funding data

 

View Original Post

I use USASpending.gov.  A little unwieldy but comprehensive and updated regularly - most agencies within 30 days;  DOD within 90 days.  It contains all federal contracts and major subcontracts and you can filter for research contracts.  For me it is a good supplement to NIH, NSF, and SBIR; and captures research spending by other agencies.Scott

David Wojick | Apr, 25 2017

 

Re: A new source for science funding data

 

View Original Post

Thanks Scott,
How/where do you filter for research contracts? I do not see it.
David
At 11:51 AM 4/24/2017, you wrote:
I use
USASpending.gov.  A little
unwieldy but comprehensive and updated regularly - most agencies within
30 days;  DOD within 90 days.  It contains all federal contracts
and major subcontracts and you can filter for research contracts. 
For me it is a good supplement to NIH, NSF, and SBIR; and captures
research spending by other agencies.
Scott
On Tue, Apr 18, 2017 at 4:35 PM, Susan Fitzpatrick
<susan@jsmf.org> wrote:

Fascinating.   One quick look – family expenditures – health
is the highest, by quite a bit. Â 

Â

Susan M. Fitzpatrick, Ph.D.

President, James S. McDonnell Foundation

Visit JSMF forum on academic issues:

www.jsmf.org/clothing-the-emperor

SMF blog 

www.scientificphilanthropy.com
  

Â

Â

Â

Caroline Wagner | Apr, 25 2017

 

Re: A new source for science funding data

 

View Original Post

Just to be clear – these are NOT NEW sources of data. These data have been around for years, and have been Congressionally mandated to be shared since at least
the 1990s.
Caroline Wagner
 
 

Jeffrey Alexander | Apr, 25 2017

 

Re: A new source for science funding data

 

View Original Post

I'll add that the USASpending.gov data on contracts is drawn from the Federal Procurement Data System, which is notoriously error-prone (although it has improved in recent years). In particular, the classifications of contracts are often faulty, especially in that contracts that are not research-related get classified as research contracts due to insufficient attention during data entry.If you're not looking at these datasets at a granular level, but more for general trends, they are quite sufficient.-jeff

Gary Anderson | Apr, 26 2017

 

Re: A new source for science funding data

 

View Original Post

I wanted to second Jeff’s note of caution regarding FPDS data. A while back we had a contractor take a look at FY11 FPDS data and we are currently taking a deep dive into FY14, FY15 and
FY16 data. Contracting officers enter detailed data on each contract action into the FPDS system. For the four years that we have looked at there are approximately 350,000 contract actions in the FPDS data. For each action, contract officers choose a product
or service code which specifies the stage and area of research and development obligation, a description of contract requirement, the amount obligated (or de-obligated) and numerous other details regarding the contract and contractor.
 
At a high level, there are at least two concerns.
 
o   Although classified as R&D within the FPDS system, these data include obligations on activities NOT classified as R&D using standard definitions. In FY16, for example, nearly 5%
of (net) obligations were for “Commercialized Products” which are not typically classified as R&D. The percentage of these “non-R&D” obligations varies substantially by research area and funding agency. 
 
o   Even if using the FPDS data only to look at broad trends, analysts would be wise to take great care in interpreting the results. For example, in FY16, according to FPDS, the Department
of Justice accounted for 62% of total(!) Federal energy R&D obligations. These obligations are predominantly on building and facility energy efficiency improvements which are classified as either Advanced Development or Engineering Development within FPDS. 
Additionally, USDA reported 46% of its contractual R&D obligations as defense related procurements.
 
This is definitely work in progress. All comments, questions and insights are welcomed.
 
 Gary W. Anderson, Ph. D.
Senior Analyst
National Science Foundation
National Center for Science & Engineering Statistics
4201 Wilson Blvd., Suite 965
Arlington, VA 22230
703.292.8572
 
 

From:
Science of Science Policy Listserv <SCISIP@LISTSERV.NSF.GOV> on behalf of Jeffrey Alexander <jeffalex@STANFORDALUMNI.ORG>
Reply-To: "jeffalex@stanfordalumni.org" <jeffalex@STANFORDALUMNI.ORG>
Date: Tuesday, April 25, 2017 at 2:29 PM
To: "SCISIP@LISTSERV.NSF.GOV" <SCISIP@LISTSERV.NSF.GOV>
Subject: Re: [scisip] A new source for science funding data

 

I'll add that the USASpending.gov data on contracts is drawn from the Federal Procurement Data System, which is notoriously error-prone (although it has improved in recent years). In particular, the classifications of contracts are often
faulty, especially in that contracts that are not research-related get classified as research contracts due to insufficient attention during data entry.

 

If you're not looking at these datasets at a granular level, but more for general trends, they are quite sufficient.

 

-jeff

 

 

On Tue, Apr 25, 2017 at 1:56 PM, Wagner, Caroline S. <wagner.911@osu.edu> wrote:

Just to be clear – these are NOT NEW sources of data. These data have been around for years, and have been Congressionally
mandated to be shared since at least the 1990s.
Caroline Wagner
 
 

C. Scott Dempwolf | Apr, 26 2017

 

Re: A new source for science funding data

 

View Original Post

All good points Gary and Jeff.  Glad NSF is looking at this from a data quality perspective.  I use the data to augment network and 'event-sequence' models of innovation ecosystems, and I add it after the core networks have been constructed from patent, NIH, NSF, SBIR and other data sources.  The models are resilient to modest levels of potential mis-codings you described.On Wed, Apr 26, 2017 at 3:07 PM, Anderson, Gary W <ganderso@nsf.gov> wrote:

I wanted to second Jeff’s note of caution regarding FPDS data. A while back we had a contractor take a look at FY11 FPDS data and we are currently taking a deep dive into FY14, FY15 and
FY16 data. Contracting officers enter detailed data on each contract action into the FPDS system. For the four years that we have looked at there are approximately 350,000 contract actions in the FPDS data. For each action, contract officers choose a product
or service code which specifies the stage and area of research and development obligation, a description of contract requirement, the amount obligated (or de-obligated) and numerous other details regarding the contract and contractor.
 
At a high level, there are at least two concerns.
 
o   Although classified as R&D within the FPDS system, these data include obligations on activities NOT classified as R&D using standard definitions. In FY16, for example, nearly 5%
of (net) obligations were for “Commercialized Products” which are not typically classified as R&D. The percentage of these “non-R&D” obligations varies substantially by research area and funding agency. 
 
o   Even if using the FPDS data only to look at broad trends, analysts would be wise to take great care in interpreting the results. For example, in FY16, according to FPDS, the Department
of Justice accounted for 62% of total(!) Federal energy R&D obligations. These obligations are predominantly on building and facility energy efficiency improvements which are classified as either Advanced Development or Engineering Development within FPDS. 
Additionally, USDA reported 46% of its contractual R&D obligations as defense related procurements.
 
This is definitely work in progress. All comments, questions and insights are welcomed.
 
 Gary W. Anderson, Ph. D.
Senior Analyst
National Science Foundation
National Center for Science & Engineering Statistics
4201 Wilson Blvd., Suite 965
Arlington, VA 22230
703.292.8572
 
 

From:
Science of Science Policy Listserv <SCISIP@LISTSERV.NSF.GOV> on behalf of Jeffrey Alexander <jeffalex@STANFORDALUMNI.ORG>
Reply-To: "jeffalex@stanfordalumni.org" <jeffalex@STANFORDALUMNI.ORG>
Date: Tuesday, April 25, 2017 at 2:29 PM
To: "SCISIP@LISTSERV.NSF.GOV" <SCISIP@LISTSERV.NSF.GOV>
Subject: Re: [scisip] A new source for science funding data

 

I'll add that the USASpending.gov data on contracts is drawn from the Federal Procurement Data System, which is notoriously error-prone (although it has improved in recent years). In particular, the classifications of contracts are often
faulty, especially in that contracts that are not research-related get classified as research contracts due to insufficient attention during data entry.

 

If you're not looking at these datasets at a granular level, but more for general trends, they are quite sufficient.

 

-jeff

 

 

On Tue, Apr 25, 2017 at 1:56 PM, Wagner, Caroline S. <wagner.911@osu.edu> wrote:

Just to be clear – these are NOT NEW sources of data. These data have been around for years, and have been Congressionally
mandated to be shared since at least the 1990s.
Caroline Wagner