In reference to the recent entry on misperception of minorities, John Sides sent me the following data on the estimated, and actual, percentage of foreign-born residents in each of 20 European countries:
The estimates are average survey responses in each country. People overestimate the % foreign born everywhere, but especially where the percentage is close to zero. This is consistent with the Erev, Wallsten, and Budescu findings about estimation of uncertain proportions.
John writes,
I don’t know where the Harper’s Index got its estimate, but I have been working on a paper with Jack Citrin, also of the Berkeley polisci dept, that looks at attitudes towards immigrants in Europe and uses the multi-nation European Social Survey (http://www.europeansocialsurvey.org/) as data. Two of the questions on the survey were:
“Out of every 100 people living in [country], how many do you think were born outside [country]?”
“Compared to other European countries of about the same size as [country], do you think that more or fewer people to come and live here from other countries?” (response categories: far more, more, about the same, fewer far fewer)These questions occur (in the above order) amidst a battery of questions of immigration, so it is possible that people have been “primed” to think that immigration is a big issue or problem just by being asked all of these questions. Leaving that possibility aside for the moment (my guess is it’s not a big confound), the results suggest that overestimation is the norm. I’ve attached a .wmf file with a graph of the actual percent foreign-born in each country and the estimated percent from the ESS. (Btw, if you have any suggestions on improving the graph, we’re open to them. The editor of the journal where we submitted the paper wanted a bar chart so I tried to make a nice one. I know you have an interest in graphical presentation.)
You will see from the graph that the UK finding you cite in the blog entry is quite representative. Only the Germans and the Swiss come close to the right answer. We also find evidence that overestimation, not surprisingly, is associated with more opposition to immigration. (There’s a possibility of endogeneity there — maybe disliking immigrants drives overestimates — but nevertheless.) However, misperception’s effect is conditional: only if you think that your country receives about the same, more, or far more immigrants than other countries (see the second question above) is the absolute level of misperception significant. In other words, it’s both this “relative” perception and the absolute (mis)perception that matter.
We do not analyze in any great detail the causes of misperception. This is our brief footnote on the subject:
“At the individual level (in a model we do not report here), misperceptions are driven by several factors. Not surprisingly, education and political engagement tend to render estimates more accurate. This suggests that misperceptions derive in part from an availability bias: the more prominent are immigrants in one’s immediate environment or social networks, the greater the over-estimate. Misperceptions are positively related to economic dissatisfaction as well, which implies that estimates are driven not only by availability but also by anxiety.”
Obviously, that’s pretty speculative. We could (and may) write another paper on the subject. In any case, I thought you might be interested in those results.
P.S. That’s the end of John’s message. I have nothing further to add, except of course that I’d like to see the details of the analyses, especially the interactions of interest. (“Those who are themselves immigrants, who have immigrant friends, and who live in urban areas tend to over-estimate immigrant numbers.”) I’d expect that much could be learned by estimating these interactions using multilevel modeling.
Some potential covariates from a European :)
How does one estimate? If you walk around Spain or Italy, you'll find many African street vendors hawking their goods laid in sheets on the sidewalks. For that reason they're highly visible. You won't see that in Germany or in Switzerland.
Secondly, how does one know that someone is foreign-born? Looking different (culturally nonassimilated) is a heuristic people use. Again, you find a difference between Germany, where the immigrants often come from other parts of Europe and don't look particularly different, in comparison to southern Europe, France, Holland and UK where they often come from other continents or former colonies.
Another comment from Europe: I'm a bit sceptical about the availability bias. I would have thought that this would have been seen in a positive correlation between actual proportion and amount of mis-estimate, but I can't see it from eyeballing the graph. This may just be a perceptual problem, though.
I wonder if John S. could extract the information to look at this. If the data can be separated into regions within each country, then the proportion of immigrants in the rgion could be used as a covariate. A bit of care might be needed, though: capital cities tend to have more immigrants (a Brit living in Helsinki writes…), and they are not representative in other ways either, of course.
Bob
Sorry to put style over substance for a second, but: it would be nice to see this graphic re-made as a plot of perceived versus actual percentage, with points labeled by country.
Phil,
I agree completely.
Regarding Aleks' comment:
Fortunately, the OECD data we use break down the immigrant population by nationality. So I examined the association between the level of misperception and the percent foreign-born from various regions (Middle East, South Asia, Africa, etc.). It appears that misperception is highest in those countries with a larger number of African immigrants (e.g., France and Portugal). The relationship is by no means perfect (the r2 in a bivariate regression is .50), but there appears to be some confirmation that "visible" immigrant populations generate more misperception.
Regarding Bob's comment:
There is a negative correlation between the extent of misperception and the actual proportion of foreign-born. This arises because Germany, Switzerland, and Luxembourg have the highest proportion of foreign-born and relatively less misperception. Our comment about availability was driven more by the positive, individual-level relationship between misperception and whether the respondent has friends who are immigrants or lives in an urban area. The idea, which is speculative to be sure, is that these respondents "see" more immigrants in their daily lives and thus are more susceptible to availability bias.
Bob's comment about region is also a good one. Unfortunately, the survey we employed does not contain information about the region where the respondent lives. (The measure of urban/rural described above comes from respondents' self-reports.) We would of course love to have additional contextual measures at lower levels of aggregation.
Regarding Phil's comment about the graph:
The original version of this graph was precisely the plot you suggest. However, we were asked to change it to a bar graph by a journal editor. Alas!
1. Something wrong with the graph, Germany has 10% according to EUROSTAT not 20%.
2. "I'm a bit sceptical about the availability bias. I would have thought that this would have been seen in a positive correlation between actual proportion and amount of mis-estimate, but I can't see it from eyeballing the graph."
I would say that is an excellent proxy for dislike of immigrants. Same way leftwingers underestimate taxes more than rightwingers.
3. I am sure they do overestimate. However people also do not respond to what they are asked but what they think they should have been asked. I am sure most would include second generation immigrants, and in Swedens case just including the second generation you get up to 12% from 16%. Than again much of those are nordic, probably not included. Would be very nice to ask:
how many muslim immigrants and how many Western European immigrants, compare the overestimation.
To respond to Tino's comment about the estimate for Germany. Our data are from the OECD (URL below). Essentially, their dataset contains a count of the number of residents in each country, broken down by their country of origin. To calculate the number of immigrants, we simply summed up the number of residents and substracted the native-born. The exact figure for Germany depends on how one counts those whose country of origin was "unknown." Deleting them from the analysis lowers the estimate of the proportion foreign-born in Germany from 19% to 11%. (Fortunately, no other country has such large numbers of unknowns.)
Using the 11% estimate means that Germans are actually overestimating the number foreign-born as well, though not as dramatically as do those in other European countries.
Ultimately, the true percent foreign-born probably falls somewhere in between. Thanks for drawing our attention to this.
And here's the URL for the data:
http://www.oecd.org/dataoecd/18/23/34792376.xls