Is it Micro? Is it Soft?

Bad Hacker, Good Hacker


Junk Science


You too can predict the future. You too can amaze your friends and confound your enemies and become a famous soothsayer. For example, you may decide to become an expert at predicting the gender of the unborn. You start by unexpectedly approaching expecting women and ask them some irrelevant questions, such as her favorite color and her date of birth and her husband’s date of birth. Then you pretend to go into a muttering trance, perform some scribbly calculations, look towards the heavens and proclaim that the child will be a boy.


Of course you will be right, half the time. When you are right, you will be remembered.  When you are wrong, you will be mostly forgotten. Word of mouth will spread your reputation far and wide, and once you are famous, you can charge for your services (and provide a money back guarantee).


Just like soothsayers, men and women of “Science” want to predict the future. They have devised many a theory, many a mathematical model that provide insights into how things work. This knowledge is then used to predict what will happen in the future. It is possible to predict the relative positions of the Sun and the Moon and the Earth for hundreds of years to come, with uncanny accuracy. Yet, an accurate forecast of the weather for a few days is quite elusive.


Science is precise, pure and rigorous. Science is about discovery. Science is the quest for discovering laws of nature that have existed forever. Newton “discovered” (not invented) the laws of motion, which relates motion with force and time. Einstein discovered the theory of relativity, showing that Newton’s discovery was not as precise as it was previously thought. The invention of the computer or the invention of the web is not science; it is a combination of technology and engineering. However, these inventions are heavily influenced by prior scientific discoveries such as electromagnetism, solid-state physics, discreet mathematics, the theory of computability and so on.


Mathematics is the foundational tool of science. Mathematics provides the methodology to analyze, compose, correlate, model, predict and express phenomenon of physical, chemical or natural origins. From gravity to relativity, from tornadoes to electromagnetism, natural and artificial phenomenon can mostly be described in mathematically. For “softer” areas of science (e.g. human behavior, social interactions, psychology), the branch of mathematics called statistics is particularly useful. Unfortunately, statistics is one of the most misused forms of the scientific process. Juxtaposition of statistical data and suppression of key issues lead to very interesting examples of flawed scientific reasoning. Statistics, coupled with misinterpretation of data, using irreproducible experiments, disregarding peer review, passing of fiction as fact and a whole bunch of innocuous yet powerful techniques are used in the promotion of pseudo-science. Pseudo-science looks like science, feels like science, but is completely “junk science”.


Statistics in fact can be misused or even misinterpreted easily. The fallibility of sloppy statistics leads to the popular phrase “Lies, damn lies and statistics”. For example, about a decade back a careful analysis of traffic accidents showed that about 70% of the car accidents happen when the driver is within 10 miles of home. This factoid led to the speculation that people are careless when they are close to home (familiarity breed complacency). Some incorrigible wag announced its better to drive only when far from home, or even better, live away from home. While the statistic is true, the interpretation is not. The reason why most people have accidents close to home, is that most of the time, they are close to home. Similarly, consider the following, possibly true statement: "96% of streets in the city are under-lit and, moreover, 88% of crimes take place on under-lit streets". This statement was intended to imply that under-lit streets are dangerous. In fact, if you think about it, you would realize that under-lit streets are actually safer (else, at least 96% of crimes would be committed in under-lit street).


Science relies heavily on experimentation. Scientific experiments have to be reproducible experimentations. Experimental methodology is laid out very carefully and have been tested and verified over the years for accuracy and infallibility. When these rules are not followed the results obtained should be deemed as junk. For example, in testing of food for taste, scientific rules say the taster sample must be large and the test must be “double blind” in order for the results to be scientific. For example, if you want to determine whether Coke tastes better than Pepsi, you cannot just ask your three good buddies. You will have to collect a large sample of tasters, from different backgrounds (racial, social, economical etc). To ensure the test is double blind, you will mark containers of Coke and Pepsi as A and B and then get some people to server them to the tasters. Neither the server nor the taster may know the identity of A or B. That is called double blind.


Pepsi runs a much-hyped taste test called the “Pepsi Challenge” which shows “more people prefer the taste of Pepsi over Coke”. The tests are not double blind, and since Pepsi runs the tests, somehow Pepsi always wins. Maybe Pepsi has an edge because the tasters are asked to sip rather than gulp. (Pepsi is sweeter than Coke, and sweeter stuff is better sipped than gulped).  Off course, if Coke ran the tests, Coke would win. This is why following the scientific process is so very important and pitfalls are so elusive or subtle. A little overlooking somewhere and the results are pure junk.


Similar, but even more ruthless rules apply to testing of medication. These tests not only have to be extremely fair and very thorough, but also must overcome the “placebo effect” and have to overcome the bias of the manufacturer of the medication. Medicine testing involves very large samples, elaborate testing, control groups, placebos and all kinds of costly procedures. Many promising drugs fail the tests.


The mercenaries who promote all kinds of natural remedies do not have to go through such tests, and the use of pseudo-science in that market is rampant. Testimonials from people cured by a magic pill are the worst form of junk science. Any white pill will cure any disease for some people.  Hence if John’s cancer went away because he took some herbal remedy, it does not mean Jane’s cancer will react in any similar way. The alternative medicine market today is flooded with wild claims, testimonials, references to scientific studies, claims such as “used in Europe for decades” (which even if true, of course means nothing) and promises of better health and longer life. Of course, nothing has been verified in even the simplest of scientific procedures and such products are almost always snake oil.


The junk science parade is never ending in the useless gadgets marketplace. Gizmos that work miracles have always been around and none of them work. Little filters that when plugged into car engines “gives you up to” double the gasoline mileage. Metallic widgets that claim to extend the range of your radio or cell phone. Plastic strips that stop snoring. Magnetic pads that you wear to get a real boost in physical abilities. All junk.


The pioneers of junk science were the tobacco companies. Using questionable methods, substandard rigor, glaring disregard for peer review and various other pseudo-science tricks they routinely proved that tobacco use is quite harmless and in fact may have beneficial health effects. Eventually they had to abandon that stance. However the retaliation has been tit for tat. The second hand smoke studies, which le to massive grassroots rebelling against smoking, are yet another pioneering example of junk science.


Junk science is a very effective tool in the political propaganda wars and activism. If you want to promote your agenda, lets say called mypetidealism, run a study that shows mypetidealism has wonderful advantages and effects and the lack of mypetidealism leads to hunger, famines, depravation and such. Using a few techniques from the junk science war-chest you can make a very convincing case that you are right. For example, while global warming is a very plausible phenomenon, and there is some evidence that the earth is warming up, there is no conclusive proof that global warming is a reality (it is a hypothesis, not a fact). However, promotion and statistics and other convincing arguments have elevated global warming to the stature of the gospel truth.


A set of very effective tools in the arsenal of junk scientists is the use of one or more fallacies. Fallacies are very plausible, logical statements that are not at all true. For example, countries where there are McDonalds restaurants, also have high life expectancy. In fact, life expectancy improves in a country a decade after McDonalds opens restaurants there. From this you can deduce that McDonalds food leads to better health.


Junk science is a dangerous, evil yet convincing variant of the scientific process. It may be hard to detect, but if it is not published in peer reviewed journals and tested by independent scientists, it does not work.


Partha Dasgupta is on the faculty of the Computer Science and Engineering Department at Arizona State University in Tempe. His specializations are in the areas of Operating Systems, Cryptography and Networking. His homepage is at




Partha Dasgupta