Read on Twitter
1/ We can't wait for the @NSERC_CRSNG fellowship stats release to start this
discussion:
In 2018, 2019, & 2020, PDF success rates for Ecology & Evolution were 30%, 37%, & 26%, respectively.
In 2021, they are <20% and might* be as low as ~8%.
But COVID, right? No. A thread. https://twitter.com/AllomeDNA/status/1357478372364935169
discussion:In 2018, 2019, & 2020, PDF success rates for Ecology & Evolution were 30%, 37%, & 26%, respectively.
In 2021, they are <20% and might* be as low as ~8%.
But COVID, right? No. A thread. https://twitter.com/AllomeDNA/status/1357478372364935169
2/ *based on the number of awards last year, though this fluctuates from year to year
If an 8% success rate sounds familiar, it's because that's what it was overall back in 2012 (ie. Harper, Death of Evidence rally in Ottawa, birth of @E4Dca). Dark days. https://www.cbc.ca/news/politics/death-of-scientific-evidence-mourned-on-parliament-hill-1.1218019
If an 8% success rate sounds familiar, it's because that's what it was overall back in 2012 (ie. Harper, Death of Evidence rally in Ottawa, birth of @E4Dca). Dark days. https://www.cbc.ca/news/politics/death-of-scientific-evidence-mourned-on-parliament-hill-1.1218019
3/ The low success rate of 2012 wasn't an anomaly, but hinted at a downward trend (source: https://www.universityaffairs.ca/opinion/the-black-hole/come-on-nserc-really-youve-completely-missed-the-point/)
250/1169 (2008)
254/1220 (2009)
286/1341 (2010)
133/1431 (2011)
98/1254 (2012) = 8%
How could NSERC have improved this measly rate?
250/1169 (2008)
254/1220 (2009)
286/1341 (2010)
133/1431 (2011)
98/1254 (2012) = 8%
How could NSERC have improved this measly rate?
4/ NSERC introduced a new rule reducing the number of times an individual can apply from 2 to 1, effectively doubling the success rates without increasing $$$. This move was widely criticized for its shadiness and potential negative knock-on effects (eg. https://www.universityaffairs.ca/opinion/the-black-hole/come-on-nserc-really-youve-completely-missed-the-point/).
5/ After evaluating the PDF program in 2013 ( http://tinyurl.com/8kegkyl7 ), success rates rose due to both 
applicants (1x rule) AND 
fellowships. From https://www.universityaffairs.ca/opinion/the-black-hole/nserc-turning-ship-around/:
110/802 (2013)
130/629 (2014)
180/585 (2015)
180/579 (2016)
199/594 (2017) =33%! Everyone
ed.
applicants (1x rule) AND fellowships. From https://www.universityaffairs.ca/opinion/the-black-hole/nserc-turning-ship-around/:110/802 (2013)
130/629 (2014)
180/585 (2015)
180/579 (2016)
199/594 (2017) =33%! Everyone
ed.
6/ This trend continued through 2020 (see #1) 
. But how much of the improvement was real ($) vs. artificial (rules)?
For the 2021 competition, NSERC completely removed the limit on the number of times one can apply. This is how it should be (see #3). HOWEVER...
. But how much of the improvement was real ($) vs. artificial (rules)?For the 2021 competition, NSERC completely removed the limit on the number of times one can apply. This is how it should be (see #3). HOWEVER...
7/ Given that the eligibility window for a PDF is ~3 years (final year of PhD + 2), one would predict that this rule change TRIPLES the number of applicants (not even including those with leave extensions).
Indeed, for E&E at least:
2020: 31 applicants
2021: 90 applicants = 3x
Indeed, for E&E at least:
2020: 31 applicants
2021: 90 applicants = 3x
8/ If* the number of fellowships is stable, this puts us at an 8% success rate. Again, we know from word-of-mouth and forums that it's at least <20% for E&E. We haven't seen rates this low for many years.
*see #2
*see #2
9/ Open Qs:
How much of the change in success rates is real ($) and how much is due to changing eligibility rules?
Depending on this, what are the consequences of such low PDF funding rates on the Canadian research environment?
The
population and # PhDs are always 
.
How much of the change in success rates is real ($) and how much is due to changing eligibility rules?
Depending on this, what are the consequences of such low PDF funding rates on the Canadian research environment?
The
population and # PhDs are always .
10/ Open Qs (cont'd):
How long can we afford to wait to find out if the low success rates this year are anomalous? Yr-to-yr variability makes it difficult to detect a hint of trend w/o >3-5 years data. That's >3-5 cohorts potentially disproportionately leaving academia &/or E&E.
How long can we afford to wait to find out if the low success rates this year are anomalous? Yr-to-yr variability makes it difficult to detect a hint of trend w/o >3-5 years data. That's >3-5 cohorts potentially disproportionately leaving academia &/or E&E.
11/ Unfortunately, underrepresented groups tend to be hardest hit by employment gaps and financial instability, and diversity tends to suffer when fewer opportunities are available. If you're aware of stats on this re. NSERC, please share. Transparent data on this is needed. #EDI
12/ FYI there is often* a gender bias in success rates. For E&E:
2019: 3x success rate for men (57%!) compared to women (21%). 0% for nonbinary/unknown individuals
2020: 2x higher success rate for men (38%) compared to women (19%). NA for nonbinary/unknown
We need better.
2019: 3x
success rate for men (57%!) compared to women (21%). 0% for nonbinary/unknown individuals2020: 2x
higher success rate for men (38%) compared to women (19%). NA for nonbinary/unknownWe need better.
13/ Given all of this year-to-year variability in success rates + narrow eligibility windows, it's important not to judge ECRs based on whether they got PDFs (or any other one thing for that matter). There are MANY excellent researchers and ultimately A LOT of luck is involved.
14/ There's also this.. https://twitter.com/LucyRodina/status/1357387707643740161?s=20
15/ This 
was motivated by my 
over seeing my brilliant early career colleagues getting shafted by this apparent stroke of bad luck. The 1st yrs post-PhD are critical for a scientist's career. Not just in academia. Postdocs show independence & provide time to publish PhD work
was motivated by my over seeing my brilliant early career colleagues getting shafted by this apparent stroke of bad luck. The 1st yrs post-PhD are critical for a scientist's career. Not just in academia. Postdocs show independence & provide time to publish PhD work
16/ So I wanted to start a conversation to learn more and try to understand whether this signals a major problem for science or just a sad day for many ECRs. I fear the former. Input most welcome!
@NSERC_CRSNG @ChiefSciCan #AcademicChatter #AcademicTwitter #PhDchat
science or just a sad day for many ECRs. I fear the former. Input most welcome!@NSERC_CRSNG @ChiefSciCan #AcademicChatter #AcademicTwitter #PhDchat
17/ NSERC stats:
https://www.nserc-crsng.gc.ca/_doc/Students-Etudiants/2019StatsScholarships_e.pdf
https://nserc-crsng.gc.ca/_doc/Students-Etudiants/2020StatsScholarships_e.pdf
+ more available on the web.
https://www.nserc-crsng.gc.ca/_doc/Students-Etudiants/2019StatsScholarships_e.pdf
https://nserc-crsng.gc.ca/_doc/Students-Etudiants/2020StatsScholarships_e.pdf
+ more available on the web.
