If there is one thing authoritarian governments hate, it is statistics. They despise anything that has to do with ratios of numbers because they bring daylight to the environment they thrive in: ignorance, confusion and deceipt.   The government of Eritrea is no different: since it has come to power, the only statistics it has released is the percentage of Eritreans who voted for the referendum.  As for the rest of the vital stats--like the percentage of Eritreans living in poverty, the percentage of Eritreans afflicted with AIDS, the percentage of Eritreans who are displaced--it is all guesswork because the government doesn't even release the most basic data in establishing a denominator: a census. 

How does an authoritarian government measure its legitimacy and popularity?  Typically, it is by manufacturing numbers.  The "99.99%" approval/confidence rate is a mainstay of every totalitarian regime.  Likewise, in Eritrea, this number or its verbal equivalent--"overwhelming majority", "all Eritreans excepting for a handful"--is bandied about to paint a picture to the world that the Eritrean government enjoys nearly total and absolute support from the people.  This claim is then "substantiated" by anecdotal evidence like the number of people attending festivals, the size of the population who attended a demonstration, the number of people visiting Eritrea, the small size of the members of the organized opposition.   Given that the demonstrations are organized by the PFDJ embassies and the people are attending this "mandatory" rallies under duress, can they be shown to be genuine support for the government?  Is attending a festival because one wants to dance and listen to music a real measure of support? If an Eritrean ambassador gathers people, reads a resolution and then asks only those who disagree to stand up meet the test of a "free and fair" measure of popular opinion? Isn't there a better way to measure popular opinion, at least in the Diaspora?

The slogan of the PFDJ is "ignorance, confusion and deception", ours is "inform, inspire, embolden."  With that in mind, we share with you our findings..

Polls and Surveys

So far, we know of no scientific poll or survey that has been undertaken to measure the popularity of the PFDJ.  For a survey to be scientific, we must have the ability to extrapolate the results we measure from a small size (the sample) into a larger group (the universe, or population.)  For that to happen, we must make sure that the sample is representative of the whole--quantitatively and qualitatively--and that every effort has been taken to reduce bias, including in the wording of the questions.  

If one takes a look at the discussion pages of many of the Eritrean websites, it is evident that most fail this test.  This is either because their sample size is not representative of the universe or because there is obvious bias in their questions or because they are perceived as being pro-government or pro-opposition.   One website that has taken great care to maintain a "neutral" stance is Asmarino.com.  ("Aynfeli, Aynfelali.")  It is true that, as of late, it has been criticized as being part of the "traitor camp" but that is only by people who believe that a medium should not offer a platform to anyone except supporters of the government.  It displays articles that are pro and against the government and, unlike other websites, it has no editorial.   This is why, we think, its discussion board, "Talking Point", is as close as one can get to a microcosm of Eritrea.  

There are, of course, limitations.  First of all, there is the "Internet bias": the sample will disproportionately represent those who are comfortable with the medium in terms of privacy protection and technical savvy.  Thus, for the sampling to be valid, we must assume that the views of those who write on the Internet is no different from those who don't.  Second, Asmarino's language of choice is English or Tigrigna and can be said to suffer from some selection-bias in that it doesn't reflect the views of Eritreans whose only means of communications is Arabic.  Again, for the sampling to be valid, we must assume that, when it comes to expressing opinion about the government, the views of those who write in Arabic is not substantially different from those who write in English or Tigrigna.  Actually, if we use the Eritrean websites that also use Arabic as a "control" to validate the results from Asmarino, we will find that the readers of those websites tend to be overwhelmingly against the government.

Data Source and Tools Used

The listener comments posted at Asmarino.com's "Talking Point", following Russom Mesfun's interview with Ambassador Gerahtu, Dan Connell and the South African students, were used as the sample size.   The sample size is 309 postings (through February 17, 11:30 AM posting time.) From each interview, the following were extracted:

(a) the country where the post originated, as indicated in the headers and flags of nations;

(b) gender of the respondent (from the names and pseudo names, as well as from the language usage if the post was written in Tigrigna)

(c) the language of the post (In addition to Tigrigna writeen in Ge'ez alphabets, Tigrigna written in Latin alphabets was counted as Tigrigna)

(d) sympathies of the writer as pro or anti-GoE.  In almost every case, this was very clear from the post.  Where it was not, the response is counted as "undeclared."

(e) A "validation variable" was used to exclude commentary that was irrelevant (one-to-one correspondence between two writers.)

Here are the results.

1) The listeners who responded were from these broad regions. 

2) More specifically, the following countries. 

3) There weren't that many identifiable females. 

4) Substantial usage of the Tigrigna language was apparent (considering this is Internet data). 

5) Over all, less than one in three of the listeners who responded were sympathetic to the Government of Eritrea. 

6) The Gerahtu interview generated most interest. 

7) It is bad news for the GoE everywhere but Europe appears to be abandoning ship faster than North America.

8) Not enough (declared) female respondents to make a robust inference but the anecdotal indication is that the GoE fares better with the gentler sex.

9) Disapproval rate is three times higher among the Tigrigna users compared to the English users.

10) The Gerahtu interview appears to have generated the greatest resentment towards the GoE; Connell's the least (but still a +50% majority). The latter is to be expected since the interviewee was a foreigner.

11) There is a glaring paradox here. The sympathy for the PFDJ in the USA tend to be relatively higher compared to everywhere else except in the UK (at 50%.)  Whether the support in UK is genuine or an aritifact due to the small size is unknown.