Neither confirm nor deny

I was recently listening to a Radiolab podcast on the history of the phrase "can neither confirm nor deny", formally known as the "Glomar Response". If you have not yet heard this episode, I highly recommend it. The episode includes an interview with the CIA lawyer hired to come up with the phrase.

Although the Glomar Response now comes in a variety of forms, the original phrase coined by the CIA in the 1970s was as follows:

We can neither confirm nor deny the existence of the information requested but, hypothetically, if such data were to exist, the subject matter would be classified, and could not be disclosed.

This got me thinking: how prevalent is the Glomar Response in Canada's Access to Information Act (ATIA) system and which agencies use it the most? As someone who files a lot of ATIA requests for their research, I've received my fair share of Glomar Responses, usually from national security and intelligence agencies.

To answer this question, I used the Government of Canada's statistics on the administration of the federal Access to Information and Privacy Acts. Although these statistics go back to 2006, the government seems to have only started collecting data on the Glomar Response in 2014.

You can find a copy of the raw data I am using on my github page.


#Load in data
ncnddf <- read_csv("")

#Filter out agencies that never invoked ncnd response
ncnddf2 <- ncnddf %>%
  filter(requests > 0)

#aggregate total number of times ncnd response invoked by agency
ncnddf2 <- aggregate(ncnddf2$requests, by = ncnddf2['agency'], sum)

#plot results
ggplot(ncnddf2, aes(x=x, y=reorder(agency, +x))) +
  theme_minimal() +
  labs(caption = "") +
  scale_x_log10() +
  labs(x = "Number of requests (log scale)",
       y = "",
       title = "'Neither confirm nor deny' responses to ATIA requests, 2015-2019",
       subtitle = "By agency",
       caption = "Source: Statistics Canada")

Alex Luscombe
Alex Luscombe
PhD Candidate in Criminology