Ethical cars

June 28, 2017

The first completely autonomous machines that will invade society as a whole might very well be self-driving cars. With “completely autonomous” I mean that these cars will perform their duties without any interaction with their owners, making their own decisions. Obviously, there is great commercial value in such transportation devices. However, allowing them to take responsibility for their own actions in the real world may involve considerable risk. For how can we be ensured that the decisions of these cars are in alignment with what we humans find morally acceptable?

A typical scenario that I get confronted with, is a self-driving car which has to swerve in order to avoid hitting a dog, but if it does that, as a consequence, hits a human. While obviously we would prefer the car to avoid hitting both dogs and humans, if there is no choice but to hit one of them, we would like the car to then choose to hit the dog. A potential solution to this scenario would be to outfit the car with ethical rules along the lines of Isaac Asimov’s three laws of robotics, e.g., with a rule that says “do not harm humans” given priority over a rule that says “do not harm dogs.”

However, the specification of such rules is not a trivial matter. For instance, it is logical that a rule would state “you have to obey the laws of traffic.” This would entail that the car is not allowed to drive through a red light. But what if the car stops for a red light, while a traffic warden motions it to continue driving? You may update the rule stating that an exception is made for directions given by traffic wardens. But what if there is no traffic warden, the car has stopped for a red light, and a police car sounding its horn is coming from behind and cannot get past unless the car drives forward a bit (through the red light) to park to the side? You may update the rule even more to take that situation into account, but is this then covering each and every situation in which the car is allowed to break the rule that it should stop for a red light? Probably not.

The point is that human drivers ever so often break the rules of traffic to avoid a problematic situation. You are trying to pass another car which drives fairly slowly, and suddenly that car speeds up. You can still get past, but you have to drive faster than the speed limit for a few moments. So that’s what you do. Or you are at a crossing in a deadlock with two or three other cars. One of them has to break the rules and start moving, otherwise they will all be stuck there forever.

The point is that human drivers improvise all the time. They know the traffic rules, they have been trained to recognize safe and dangerous situations, and know how to anticipate on the behavior of other drivers. And sometimes they bend or break the rules to avoid problems. A self-driving car that cannot improvise is dangerous. However, a consequence of the need for improvisation is that any rules that we would want to impose on the car, it should be able to break. The only alternative would be to envision each and every situation in which the car could find itself and specify the exact behavioral rules for it to deal with all those situations. Clearly, that is impossible.

So how do we get a car to behave like a responsible driver without laying down an endless list of rules? The answer is: by training it. First, we let the car drive in a highly realistic simulation, punishing it every time that it causes a situation that is undesirable, and rewarding it when it manages to perform well. A learning structure can incorporate the lessons that the car learns, thereby bringing it ever closer to being a model driver. Once it is perfect or almost perfect in the driving simulation, it can be let loose on the road under the guidance of a human, continuing learning. In the end, it will behave on the road as well as, and probably a lot better than, a good human driver.

How will such a car deal with a choice between hitting a human or a dog? It is likely that similar situations will have cropped up during the training process — maybe not with exactly the same race of dog and the same human as in the real situation, but as the car has been trained instead of having been given specific rules, it has the ability to generalize, and it will make the choice that is closest to what the training would have rewarded while avoiding choices that the training most likely would have punished. In other words, it will choose to hit the dog to avoid hitting the human, just as it would likely hit a cat, a moose, a badger, or a duck in order to avoid hitting a human.

It might, however, in a situation where someone pushes a mannequin in the road, hit a dog to avoid hitting the mannequin — not because it thinks the mannequin is a human, but because the situation of hitting the mannequin more closely resembles hitting a human than the situation of hitting a dog does. If we do not want the car to make that choice, we should ensure that its training regime includes situations in which it has to deal with objects that resemble humans but are not humans. This, however, could lead to a situation in which it chooses to hit a highly inert human to avoid hitting a dog. That’s the problem with allowing a car to make its own choices based on how it is trained: you can probably always find an exceptional situation in which it is not doing what we hoped it would do. The same is true for humans, of course, and in the end the self-driving car will probably still be a much safer driver than any human.

So if one wonders how we can be sure that the ethics of a self-driving car will be acceptable to us humans, the answer is that we can only draw conclusions based on observations of how the car deals with tough situations. We will not be able to open up the car’s brain and examine some kind of ethics module to read how it will deal with situations that come up. Therefore there is no way for us to be “sure.”

We can only draw comfort from the fact that if at some point the car takes a decision that we find doubtful, we can punish it and it is likely to make a different decision when a similar situation comes up again. It will be less stubborn than the average human in that respect.

Diversity interlude

April 24, 2017

I don’t know if I am yet done with the topic of diversity. The discussions about diversity are currently rather intense, and I feel I still have to say quite a bit about it. In general, I have noticed that my position is that diversity is based for at least a considerable part in biology, while those who are on the side of social engineering are of the opinion that it is all culturally determined. In that respect, I discovered a fascinating series of documentaries by the Norse comedian and sociologist Harald Eia, called “Hjernevask“. I haven’t yet completed watching them, but up to now I am pretty surprised about some of his discoveries. The two things that stand out to me are: (1) the role of biology in determining differences between genders and races is much bigger than I had previously assumed, and (2) evidently in “enlightened” western countries such as Norway the thought that differences can be the result of biology is actively shunned even by people whose job it is to know better. Heartily recommended.

Diversity III: Diverse equalities

April 8, 2017

This post is a follow-up to my post on the wage gap, and my post on the glass ceiling. In these posts I showed that there is no reason to assume that gender discrimination is at work in determining the salaries or promotion chances of women.

In The Netherlands, “equality” is generally seen as an ultimate good. Which is why underrepresentation of women and minorities in professions and organizations is often seen as inherently wrong. Certain political movements have made fighting inequality into a main feature of their program. For instance, minister Jet Bussemaker is quite miffed about the fact that only 18% of full professors in The Netherlands are female, and she therefore forces universities to appoint more women to such positions. For 100 new female full professors she makes extra money available for the next five years, and on top of that she also wants universities to put women on 200 chairs that are currently held by men.

Basically, this means that Bussemaker wants to change the top selection criterion for appointing someone to the position of full professor from “we want the best person for the job” to “we want someone who has no Y-chromosome.”

For me, a major question is why she (and with her many others, such as professor Derks) believes that the underrepresentation of women in this area is the result of “inequality.” The straightforward line of thinking that is commonly held, is that 50% of society consists of women, and thus, if women were treated equal to men, 50% of full professors would be women. Naturally, such a thought process is too simplistic. It is a silly notion to divide Dutch citizens according to particular criteria such as gender, immigration status, age, sexual orientation, or physical attributes, and then expect that the percentages you end up with are reflected by a certain profession. As I showed in the previous two posts, if you take into account unequal participation in the job market, the 18% female full professors that we have now are close to what you may expect if men and women have equal opportunities (assuming that in academics the statistics for job participation are about average for The Netherlands).

The point is that what minister Bussemaker and others like her want to see are not equal opportunities, but equal outcomes. What they fail to acknowledge is that if you confront 100 men and 100 women who already hold a good job with the prospect of another job with more status, offset by a greater workload and higher responsibility, 80 of the 100 men will accept that job, while 80 of the 100 women will walk away. In our society where people are given a lot of freedom of choice, women tend to make different choices than men, and thus equal opportunities do not lead to equal outcomes.

Assuming that minister Bussemaker does not want to take away freedom of choice, there are only two ways open to get a higher percentage of female professors: either making men less ambitious, or lowering the requirements for women (and only for women) to get the job. I have no idea how you would accomplish the first, so I am not surprised that she chose the latter.

Yes, I know that there are female associate professors who are ready to become full professors. But there are far more male associate professors with the same ambition. In an equal-opportunities environment, which we have now, you end up with more men being appointed than women, because with the same quality there are more men than women available for the job. When I say that Bussemaker lowers the requirements for women to become professors, I am not saying that that puts universities in a position where they are forced to promote female associate professors who are not ready to become full professors — I am only saying that these women no longer have to compete on a level playing field with men. It is possible for them to do a worse job than some men would do, and still be chosen for promotion over those men.

Professor Derks states that women do not want to adapt to the standards set by men. I assume that she means that many women want to be full professors in a part-time job with a nice balance between their work and home life. This sounds a lot like having your cake and eating it too. In practice, as with most high-demand jobs, being a full professor entails sacrificing much of your home life for your work. It is really a tough job. Maybe that is a standard set by men, but it means that if, as a woman, in an equal-opportunities environment you want to compete with men, you have to meet that standard. By replacing the equal-opportunities environment with an equal-outcomes environment, women can get away with doing a worse job than men.

In sports, women usually do not compete with men, as men tend to be physically stronger and faster. Consequently, female sports are seen as inferior to male sports. By creating a special academic division for women to compete in, where standards are lowered, you create a situation in which female academic achievements are seen as inferior to male academic achievements. If I were a female professor today, who clearly got the job because she competed successfully on a level playing field with men, I would not be happy about that.

The 100 extra positions that Bussemaker creates are filled by women only, which leads to a situation where some top-quality men see full professorships going to women who, even if they do a good job, do a worse job than these men would do. And if the rest of Bussemaker’s plans get executed, they know that their chances to ever be promoted are diminishing rapidly, as they have to compete for a decreasing number of positions. If I was a young man who was contemplating a career in academics today, considering the unequal treatment of men and women and the lowered career prospects, I would probably choose a job outside the academic world, regardless of how much I love scientific research. All in all, the policies of Bussemaker may do serious harm to the quality of Dutch academics.

In conclusion, the belief that forcing universities to appoint more female professors will resolve some sort of societal injustice is misguided. In fact, it creates a societal injustice by lowering the standards that women need to meet to become full professors, while increasing them for men.

I wish to point out that there is another solution for the problem that certain female associate professors are not getting promoted, even though they are ready for it. That solution is to make the promotion to full professor an automated process when you meet certain criteria. This is the case in many countries outside The Netherlands, such as the UK. If you do the work, you get the position. I am in high favor of that.

Making the promotion to full professor based on individual accomplishments rather than the academic structure with only a limited number of available chairs, has the added benefit that there will be more full professorships in The Netherlands, and thus the requirements for the job may be slightly lowered for men as well as women.

Diversity II: Diverse choices

April 6, 2017

This post continues my previous post on the topic of diversity, in which I discussed the first part of the sentence “Women still earn less and are slower to be promoted” (lifted from an article in De Volkskrant). In the present post, I delve into the second part. Is it true that women are slower to be promoted, and if so, is that (as is often suggested) the result of gender discrimination?

What I assume the author is referring to with the second part of their statement is the fact that the representation of women in higher positions in companies and organizations is considerably less than the representation of men. In the academic world in the Netherlands, for instance, only 18% of full professorships are held by women.

There is no debate about the fact that women are underrepresented in higher positions. What is up for debate is whether this is the result of gender discrimination. Despite the often-expressed but seldom-substantiated notion of some sort of “glass ceiling” that keeps women down, I would argue that the state of affairs is mostly the result of personal choices.

In The Netherlands, almost every job can be done part-time. In practice, we find that men usually take full-time jobs, while women predominantly choose to work part-time. According to the Dutch Central Bureau for Statistics (CBS), in The Netherlands in 2017, only 33% of working men held a part-time job, compared to 78% of working women. The ministry of OCW (Education, Culture, and Science) reports that, on average, women have a work-week of 25 hours, while men have a work-week of 36 hours, i.e., on average women work 1.5 days less per week than men. It also reports that only 71% of women between 20 and 65 years of age have a job, as opposed to over 82% of men.

Referring to statements of professor Derks in the interview I mentioned above, working part-time has a big influence on career opportunities — thus, if you are interested in having a career, you should realize that working full-time is more or less a necessity. The detrimental effects of working part-time are two-fold:

First, when working part-time you simply accomplish less in your career in the same number of years than someone with the same abilities who works full-time. This leads to a lower status among peers, and a less impressive CV.

Second, professing the desire to work part-time shows a lack of ambition and motivation, which consequently leads to a reduced willingness of an employer to hand out a promotion.

The second effect is in line with differences between the average man and the average woman in how they view work (see, for instance, the report from Monsterboard). Typically, men are more driven by status than women: men far more than women seek a higher salary, possibilities for personal growth and development, independence, and responsibility; women far more than men seek possibilities to work part-time, possibilities to work from home, and a nice balance between work and home life. You may translate that as “on average, women are less invested in their work than men.” That cannot help in having a career. The employer who says: “What I really want from my personnel is that they focus on their home life” has not been born yet.

Note that by no means I am saying that all women lack the motivational drive, the capacity, the experience, and the willingness to have a career. Obviously a considerable number possess these attributes, which is why we see such women in higher positions. They made the choice to invest in their work rather than their personal life, just like many men do.

It is a great benefit of our society that one can actually choose how to balance work and home life. But when someone chooses to be less invested in work, he or she should expect having a hard time getting into higher positions. That holds for women as well as for men. Since, in general, women tend to be much less invested in work than men, it is not surprising that in higher positions, the number of women dwindles. Whether or not the prevalence of women in part-time jobs fully explains the low number of women in higher positions is unknown. Saying that it is the result of gender discrimination, however, is unwarranted.

The only extensive scientific research on the topic that I could find is an article by Williams and Ceci of Cornell University in PNAS, who found that for tenured assistant-professor positions in biology, engineering, economics, and psychology in the US, if you compare female candidates with male candidates of equal quality and matching lifestyles, the chance for female candidates to be hired is twice as high as the chance for male candidates. The conclusion is that gender discrimination is at work here, but it is to the considerable advantage of women.

I wish to finalize this discussion with a simple calculation. Knowing that 82% of men have a job against 71% of women, and that 67% of working men have a full-time job against 22% of women, if we assume that there are about the same number of men and women between the ages of 20 and 65, that means that 78% of the people who work full-time are men, against only 22% women. If we compare that 22% with the 18% of women with full professorships in The Netherlands, it strikes me that these numbers are pretty close. So, while 18% sounds low, it is actually close to what you can expect if you take into account the generally accepted assumption that working full-time is a requirement for getting a high-end job.

In summary, there is no reason to assume that the underrepresentation of women in higher positions in companies and organizations is the result of gender discrimination. Still, there are political forces at work that try to enforce placing more women in higher positions. These forces are particularly strong in the academic world, where universities are ordered by ruling politicians to appoint more women as full professors. What I think about those forces I will discuss in a follow-up post.

Diversity I: Diverse wages

April 1, 2017

The term “diversity” refers to the uniqueness of individuals. In policy making, the theme of “diversity” refers to the inclusion of underrepresented groups of people in particular functions or domains.

In recent years, the theme of “diversity” has infiltrated many aspects of society, and now has serious impact on policies in governmental matters and professional life. I am quite wary about the effects of the diversity discussions. It is such a complex theme, however, that it is hard to formulate statements on it, without running serious risks of finding oneself under a bombardment of accusations of being some kind of –ist or –phobe, which then puts an end to any chance of getting into a civil parlay.

I think that the discussions around this theme, now they are widely politicalized, will only intensify in the coming years. As a scientist, I have the responsibility to deal with these matters as objectively as possible. As I am increasingly involved in policy making and providing advice in matters of policy, it is important to me to be able to express opinions in this domain that are as close as possible to the truth, without letting emotions and political pressure get in the way.

Today’s ranting is instigated by an interview in the Dutch newspaper De Volkskrant with Belle Derks, professor of Social and Organizational Psychology at Utrecht University, specialized in “gender equality.” I had many thoughts while reading this interview, and I might get back to it in the coming days. For now, I want to zoom in to the second sentence of the main text of the article, which is “Vrouwen verdienen nog altijd minder en klimmen minder snel op naar hogere functies” (“Women still earn less [than men] and are slower to be promoted”). For clarity, I wish to point out that this is not a statement of professor Derks as far as I can see, but part of the introduction of the journalists.

The quoted sentence is factually correct, but it suggests something that is factually wrong, namely that women get discriminated as far as their wages and promotion opportunities are concerned. With respect to wages, one only needs to read the report of the CBS (the Dutch Central Bureau for Statistics, a governmental organization that prepares reports on all aspects of Dutch social and economic life) on this matter, which is available from their website. This report investigates specifically gender differences in salaries in The Netherlands, and is updated up to the year 2014. In this report one can read the following:

For jobs in companies, the hourly salary of women is about 20% less than that for men. For government jobs, the difference is 10%. However, these numbers are not corrected for differences in actual work, i.e., it just averages the hourly wages of all women that have a job, and all men that have a job, regardless of position, education, or responsibilities. If you correct the differences for about 20 factors, which include age, experience, and position, then the percentages drop to 7% and 5%, respectively.

Does that mean that gender is the explanation for the remaining salary differences? The CBS report states that there is no reason to suppose that gender discrimination is at work here, as there are quite a few factors that can explain the remaining differences, which they were unable to take into account for lack of detailed information. These are, among others, motivation, job level, and secondary employment conditions such as exchanging salary for extra vacation days. As the report states that women are far more likely than men to work part-time, one might expect that in particular women would take the opportunity to sacrifice some salary for extra spare time, though the report does not delve into that.

As there are plenty factors available that may explain the remaining small differences in average wages between men and women, the CBS found no indication that gender discrimination is an explanation for these wage differences. I think that settles the matter.

What I found particularly interesting is that the report also made clear that up to 36 years of age, women actually earn a higher salary than men for doing the same work. However, again this has nothing to do with gender, but with the fact that among younger people, women have, on average, a higher education than men. Between 36 and 45 years of age, there are no significant differences in salaries between men and women. Over 45 years of age, men tend to earn more than women, for which an easy explanation is that in those age groups, men tend to have a higher education and more job experience than women.

How these trends will be extrapolated to the future is hard to say. If many women between 30 and 40 years of age are losing interest in work and accept lower salaries for, for instance, a more extensive home life, the picture might remain as it is now. If women keep focusing just as much as men on their careers and earnings, clearly in about 30 years, women will on average earn more than men in all age groups.

The discussion above has contextualized the first part of the sentence “Women still earn less and are slower to be promoted,” and refutes the suggestion that women earn less than men because of gender discrimination. The second part needs another discussion, which I will get to later.

For now, the conclusion is that in The Netherlands a small gender wage gap exists, but that there is no reason to think that the explanation for it has to do with gender discrimination.

The Witness

December 25, 2016

The Witness (2016) is a game made by Jonathan Blow.

I hesitated when I wrote that first sentence, because I have a feeling that calling The Witness a “game” is not doing it justice. However, I don’t think we have a better name for what it is, so let’s call it a “game” from now on, realizing that the term “game” is expansive and that The Witness is located somewhere at the boundaries of that expanse.

On the surface, The Witness is a collection of line puzzles sprinkled over a nice-looking 3D environment, namely an island with several different biomes, such as a mountain, a desert, an orchard, a jungle, and a village. Some of the puzzles are placed in obvious series. By solving seven of these series of puzzles, you open up a final area where some difficult puzzles are found, which you solve to end the game.

Some people stated that without the 3D environment, The Witness could be a smartphone game, and they are right, up to a point. You can play The Witness as a collection of puzzles, and when you are done with those, finish the game. If that is what you are looking for, and you are not enticed to look beyond those line puzzles, then the game has little more to offer you than the average smartphone puzzle game.

Most of the puzzles are actually obnoxiously simple, and you solve them in mere seconds time. The ones that have you stuck for a while longer are almost always those that unexpectedly introduce a new element, or puzzles that are placed in the world by themselves, or puzzles in the island’s village. The puzzles that introduce a new element are rather clever in the sense that when you see a new element, you are probably not going to solve the puzzle at all, as you have no idea what the new element means. However, you also quickly realize that somewhere on the island there is a series of puzzles that explain this new element. And the wonderful thing about these “explanation puzzles” is that they do the explanation purely intuitively. There are no help texts in this game. The puzzles have been carefully constructed in such a way that almost everybody will intuitively grasp what they have to do to solve them, and thus add to their knowledge of the game.

For the few puzzles that provide a tough challenge: you actually do not have to solve any of them to be able to finish the game. If you say: “screw these difficult puzzles, I just want to see the ending” and you rush through the puzzles that are easy enough, you will get access to the final series of puzzles (you only need to solve about 50% of the puzzles to get there). Those final puzzles can get pretty difficult, but you can find solutions online, so you enter those solutions and then experience the final ending — which will be a huge disappointment. The ending just makes the camera traverse the island, out of your control, after which you get dumped at the start of the game, with all the puzzles reset. That’s it. That’s why you went through all that trouble. I can imagine someone writing a review that states “Some of the puzzles were fun, but they got quite tedious after a while. The puzzles near the end were more interesting and challenging, but probably too hard for the average player. The ending is stupid, and a slap in the face of all who went through the trouble of solving all these puzzles. 2 stars out of 5.” (I have seen reviews like this.)

But here is the thing: The Witness is not about the puzzles. The puzzles are just a mechanism. They provide something for you to do to get a sense of progression. But they are not the core of the game. There is so much more to experience in the world. There are so many ideas placed in this game, most of them “hidden in plain sight.” You may miss these ideas, but once you catch some of them, you start seeing more. Fortunately, the game has ways built in to intuitively point out the most obvious of these ideas to you. But they are just the start.

I can’t say more about this, as you have to experience these elements of The Witness for yourself. The whole point of the game is to allow you to discover them. The Witness is one of the few games of which I say: “Make sure that you do not read anything about it before you play it, this game is best played without foreknowledge, and should only be discussed with others after you finished it.”

Let me just say that the game contains a lot of clever tricks. Many of these appear to be no more than just “clever,” and I expect that some of them indeed are nothing more. But others have layers of depth to them. They raise questions, which the game sometimes — but not often — provides an answer for. To give you an idea, without spoiling anything, here are some of the questions that you may ask:

  • Why does the sun not move?
  • Why is there a single rain cloud in the sky?
  • Why does the player have a shadow but no hands?
  • What happened to the occupants of the village?
  • Why is the game called “The Witness”?

These questions are all directly related to what you see in the game. But they lead to deeper, philosophical questions, which the game shares viewpoints (not answers) on if you are willing to look for them. However, besides these deeper layers, there are also game elements that are just there to screw with the player. In some cases I got the feeling that Jonathan Blow included specific features just to raise a middle finger to a particular kind of game player. In this sense, The Witness not only contains a game, but also a meta-game, a meta-meta-game, and probably even a meta-meta-meta-game.

I realized that the experience that The Witness provided me with, in many ways resembled my first reading of Douglas Hofstadter’s masterpiece Gödel, Escher, Bach: An Eternal Golden Braid. This book is quite good at making you understand intuitively the complex ideas that it discusses before bringing up the actual ideas themselves. It also has so many layers of depth, that when you are sensitive to these things, you make new discoveries every few pages, sometimes thinking “that’s clever,” other times thinking “that’s deep,” and yet other times thinking “what am I supposed to think about that?” The point is that you feel that Hofstadter placed each and every word in that book deliberately, and that you are always missing elements just because you glanced over them — even though you get the overall intentions of the book.

The difference between The Witness and Gödel, Escher, Bach is that Hofstadter makes the overall purpose of the book explicit in the final chapters, while Jonathan Blow never states in clear terms what he means with the game. In that sense, Gödel, Escher, Bach is a scientific work, while The Witness is more a spiritual work — even though it radiates the feeling that Jonathan Blow heavily prefers science to spirituality.

I can’t say more about The Witness without running the risk of spoiling anything. Let me just say that I think it is a game that should be played by everybody — even though not everybody will get out of it what I got out of it, it manages to use the medium “game” as a way to let the player discover new ideas, which is a rare thing indeed. Moreover, it has a highly accessible interface, so that young and old can play it, even if you have very little previous experience with games.

I stated before that games can be an art form, though they seldom rise to the level that you expect of high art. The Witness, however, does.


ADDITIONAL (January 1, 2017): I got into a brief discussion on The Witness on VideoGameGeek. I gave a fairly lengthy interpretation of the game, including my personal answers to the questions stated above. In it, I refer to the excellent video made on the game by Joseph Anderson. For those who are interested in my comments, you can find them here. Note that my comments and the video mentioned both contain very heavy spoilers, and I urge you not to consult them unless you have already finished the game. I know this warning is given by many for almost any review/discussion, but it is really serious this time, as what the game offers you is the discovery of perspectives, and you cannot experience that discovery if you just listen to or read about someone else’s discoveries.

I couldn’t fail to disagree with this less

November 28, 2016

Some people think that brutal murder may be justified in certain situations. Do you think this applies to the following circumstances? (multiple answers possible)

  • Career advancement
  • Defense of family members
  • Drug-induced rage
  • Honor killing
  • Jihad
  • Entertainment
  • Bad parenting
  • None of these

Would you check one or more of the ones apart from “none of these”?

Personally, I would ask the interviewer what he means with “this”. It seems that with “this”, he is referring to the fact that “some people think that brutal murder may be justified in certain circumstances” — grammatically, that is what “this” means in this case. But maybe, just maybe, he is referring to “brutal murder being justified in certain circumstances.” If he means the first, I would definitely check at least two of the options. If he means the second, of course I would check only “none of these”.

So what if you ask this question to 10,000 people, and 100 of them simply refuse to answer the question, while 2,500 check one or more of the suspicious options? Is the interviewer then allowed to conclude that:

  1. only 100 people had problems with this question;
  2. 9,900 people interpreted the question as “Do you personally think that brutal murder is justified in certain circumstances?”; and
  3. therefore 25% of all people would condone brutal murder.

If you, like me, think that the interviewer did a bad job, you might ask him: “If you really wanted to know whether people condone brutal murder in certain circumstances, why did you not ask them that? Why ask such an ambiguous question?”

The interviewer then answers: “If I would ask people directly whether they condone brutal murder, everybody would say ‘no’. So I could not ask that. I added “none of these” to clarify the question.”

To which I would say: “If you know that nobody would check any of the possibilities if you would ask them directly whether they condone brutal murder, doesn’t that mean that you already know that almost nobody condones brutal murder? And if you think that the question needs clarification, don’t you yourself agree that it is badly formulated? Moreover, why do you think that adding “none of these” as an option is clarifying anything?”

Of course, this interviewer seems to believe that quite a few people actually do condone brutal murder, and are just not willing to admit that. So he feels he needs to trick them into admitting what they actually think (according to him) by asking a question that can be interpreted in multiple ways, whereby people interpret it in one way while he can claim that it was meant another way.

Questionnaire-based sociological research is always a bit flaky, and therefore needs to use questionnaires that are as unambiguous as possible in order to draw any conclusions at all. Questionnaire-based research that uses vague and misleading questions is worthless. A sociologist should not try to play the television lawyer who asks questions in a roundabout way and then, when the accused accidentally seems to admit being guilty, shouts out “A-ha! He fell for my little trick!”

Yet the European Commission has published a report that is based on such a vague and misleading questionnaire. This report is on “Gender-based violence” and draws the conclusion that 25% of European citizens condone “sexual intercourse without consent” (i.e., rape). It posed the following question to the respondents:

“Some people think that sexual intercourse without consent may be justified in certain situations. Do you think this applies to the following circumstances? (multiple answers possible).” I.e., the exact same question as I pose above, with “sexual intercourse without consent” substituted for “brutal murder”.

Since for this question 27% of the respondents checked one or more of the options besides “none of these”, the report concludes that “more than one quarter of respondents (27%) [is] of the opinion that there are situations where sexual intercourse without consent is justified”. Not that “27% of respondents think that there are situations where some people (not necessarily they themselves) believe that sexual intercourse without consent is justified”.

When confronted with the ambiguity of this question, the primary investigator who had formulated it, responded like I wrote above: that this was the only way to get people to answer anything else than “of course rape is never justified”, and that they should be able to understand the actual intent of the question, regardless how convoluted its formulation was, because he added the “clarification” of having an option “none of these.”

Is it likely that this question was interpreted differently by different respondents? Yes, it is very likely. All of the questions that come before it are clearly asking the personal opinion of the respondent, in forms such as “Do you agree that…” This particular question, which comes at the very end, is the only one that is suddenly talking about the opinions of other people. The fact alone that this question was not formulated as “Do you think that sexual intercourse without consent is justified in certain circumstances” while all the other questions follow a formulation like that, clearly indicates that this question should not be interpreted as the respondent being asked to give their own opinion.

The Belgian newspaper “De Morgen” noted that the Dutch translation of the questionnaire differed for the Flemish (the Belgian Dutch speakers) and the Dutch. In the Belgian version, “Do you think…” was translated with “Denkt u…”, which is more or less a literal translation. For the Dutch, it was translated as “Vindt u…”, which means “Are you of the opinion…”, which suggests a bit more that the sentence could mean “do you personally agree…” (still very vague, though). The net result was that significantly more Belgians checked one of the suspicious options than the Dutch did. Considering that there is very little difference in morals between Belgians and Dutch citizens, this demonstrates once more that this question tends to be interpreted by different people in different ways.

So what can be concluded from the questionnaire? The Dutch paper “De Volkskrant” argued that formulating questions in such a roundabout way is a common method in sociology to ask about unpopular opinions. And that while the report can be criticized, in light of the other findings of the report the numbers were not unbelievable. They state: “if you ask in the correct way, many people are shown to have pretty nasty ideas about applying sexual force.”

This conclusion is hogwash. If questions on a questionnaire are dubious, nothing can be concluded from them. The fact that the resulting numbers, when viewed in the light of your own beliefs, seem to support those beliefs, doesn’t change the fact that the numbers are meaningless.

I checked the complete report, and contrary to the statements of “De Volkskrant”, the other numbers are actually not supporting the findings for the last question at all. For instance, the question that asks whether the respondent finds domestic violence (which includes rape of a partner) acceptable, is answered by 96% of the respondents with “unacceptable in any circumstances” (that is, against women — evidently only 94% find it unacceptable against men). Raping of a spouse is a form of domestic violence; this means that if you do not condone domestic violence, you do not condone raping a spouse. How can we rhyme the fact that less than 4% of respondents condones rape of a spouse, with the proposition that 27% of the respondents condones rape in general? Is raping a spouse for 23% of the respondents worse than raping a stranger?

The final statement of “De Volkskrant”, that starts with “if you ask in the correct way…” makes me wonder what they think the word “correct” means. For me, the only “correct” way to ask questions on a questionnaire is making sure that the questions are unambiguous.

In software engineering, we use the term GIGO: “Garbage in, garbage out”. I assume this applies to sociological questionnaires too.

Lack of progress

November 19, 2016

In recent weeks in the media it seems new terminology has been introduced to refer to “making big strides forward” in a certain area (such a battling traffic jams or dealing with climate change). They call it “making a quantum leap forward”. When I first saw this term, considering that when we talk about the “quantum scale” we mean the scale at which quantum effects start to play a role (less than 100 nanometers), I thought the people using it were sarcastically trying to express that no measurable results had been achieved. However, in layman’s terms, it seems that “quantum” now stands for “huge”. I am no physicist, but I find this painful.

The digital overlords are here

May 6, 2016

I read a very nice statement of prof. Pedro Domingos of the University of Washington in the Dutch newspaper NRC of May 4, 2016. In answer to the question “What do you tell people who are afraid that self-learning computers are getting so smart that they will take over the world?” he said: “Computers are stupid and they already took over the world. It would be better if they would be smarter.” That’s going to be my stock answer to this question from now on.

21st century skills

May 2, 2016

Computational technology causes the world to change rapidly.

Almost 30 years ago I got my first job as a computer programmer. At the time, only larger companies with a big administrative overload used computers. Or rather, “a computer”, because it was rare for a company to have more than one. There were no personal computers, no Internet, no mobile phones. People still used typewriters.

In those 30 years, the way people work and live has undergone huge changes. That is exceptionally clear when looking at the kind of work that people do. Mailmen, for instance, delivered the mail twice per day when I was a kid — now they deliver mail twice per week, which means that the contingent of professional mailmen has been decimated. Bank offices are closed because banking can be done much easier online. Information desks can be manned by digital avatars or be replaced by online information systems. Large department stores go out of business because people make their purchases online, leading to an enormous decline in the need for having salespeople. And though this has currently caused a small increase in the demand for people who work in transportation, we can see self-driving cars on the horizon, replacing the need to have any chauffeurs at all.

These are all “low profile” jobs, but “high profile” jobs aren’t safe either. I have taught programming to professional journalists, who told me that computers are taking over large parts of their jobs, writing basic articles and doing automated background research — they wanted to take my courses because they realized that without skills in digital technology, they would be out of a job in a few years time. Programs have been developed that take over a menial but oh-so time consuming part of lawyers’ jobs, namely researching case histories. Computers can write music, produce paintings, and even sculpt — why would you have someone hammer away at a block of granite for six months when a 3D-printer can produce a sculpture with a few hours of work? Even designing and running scientific experiments has been offloaded to computers in some research domains.

In the 30 years in which I have been a professional worker, I have seen the job market change from hardly incorporating computers at all, to a situation in which the need for human employees has been reduced considerably — regardless the job. And that change has not come to an end yet.

This does not mean that there is no place for humans in the job market. It does mean, however, that only humans who can make contributions that a computer has a hard time making on its own, can be assured of a job. In the near future, employability will be invariably linked to the ability to integrate the power of humans and computers in a way that enhances both of them.

The problem is that to be able to use computers to improve the quality of one’s work, it does not suffice to be able to use a word processor or spreadsheet. One should actually be able to expand the capabilities of computers from the perspective of one’s chosen profession. For example, a journalist who can only run a fact-finding computer program that someone else wrote, is not needed. However, a journalist who is able to expand a fact-finding program so that it can come up with facts from new sources, is an asset.

To be able to employ computers in such a way, one needs the skills to think and solve problems like a computer programmer. Having taught students computer programming for many years, I know that this does not come naturally to most. To acquire the necessary skills, students need to spend several intensive courses on the topic.

Considering the fact that universities and colleges are supposed to prepare students for the job market in which they have to function for 40 or more years, and considering the fact that in the very near future (if not right now already) the ability to incorporate the power of computers in any job is a necessity to being a valued worker, one would expect that “computer programming” is one of the basic courses that any student needs to take. Unfortunately, it is not. Typically, basic required courses are “scientific writing”, “philosophy of science”, and “statistics”, but “computer programming” is still seen, by most education programs, as an optional skill. It is not.

In my view, any course program that does not make “computer programming” a required course, is doing its students a disservice, as it is not preparing them for the job market. Actually, I would prefer it if secondary, or even primary schools would incorporate such courses, as programming skills tend to be easier to learn at a younger age. The reason is that they need a particular way of creative thinking, which is harder to acquire when one is already used to solving problems in the reproductive ways that are normally taught at schools.

All students, regardless of their chosen topic, need to learn how to program. Not because we should raise a generation of computer programmers — professional programming is a specialization that only a few people need to be able to do. But the ability to create programs provides students with the skills to think and solve problems like a computer programmer, to gain insight in the possibilities and limitations of computers, and to leverage the power of computers in a particular domain in a uniquely human way.