The scientists are also very well capable of producing falsehoods.
Falsehoods can hide in statistics, in methods, in systematic noise, and survivor bias.

Even without fraud many scientists will often overlook:

• the problems with their statistics,

• the methods that they use. Like the instruments that they use. Or the calculations that they use.

• the systematic noise caused by the way they measure.
  Or noise caused by the environment. Or noise caused by the way and timing of measurement.

• Survivor bias. A lot of data and problems with the data will not reach the researcher. Or they will scrap it off, because the data that is not good enough. The latter happens more often than you think, because every person in the research chain can remove data. From the student, laboratory worker, local scientist, project scientist, project manager to the director. Each person will add their bias to an observation, and avoid problematic data.

• Using words as the explanation instead of the description of an unexplained problem.
  In the middle-ages it would be the devil, or a miracle. But we do not have a good explanation of what causes gravity (in my opinion).
  We have good formulas, but what is causing gravity is often just the word "gravity".
  Or a more mystical: "The bending of space/time".
  But what would be wrong with stating: "Tendency towards depression is caused by the placebo effect".

Fraud

With fraud, this picture gets a lot worse. One or all people in the research team will now deliberately work towards a certain result outcome.
This is often standard procedure for certain political or commercial related research, where the outcome determines the distribution of a lot of money. Many people think that it is ok to bend the rules a little if it is beneficial. The amount by which they bent the rules depends on the rewards. If you do not know that this is standard procedure, you should start investigating it better.
Often some scientists will make conflicting statements, because the conclusions are in conflict with the real world.

The goal of scientific fraud is to create an certain outcome out of the data, while making it seem that the outcome is scientifically valid.
So they work the outcome that they want.

1. Statistics. Lying with statistics is easy.
   In the medical industry the most common method is to reassign problematic cases to other categories. And then make those categories seem irrelevant.
   Like: There were 3 deaths during the testing of Cocaine, but these people were later diagnosed as addicts. So we are certain that the deaths were caused by the addiction.
   Or: During investigating the death we saw that the bullets had not caused any lead poisoning. So we can not conclude that the death was caused by the bullets.

2. Methods and calculations.
   John measured the hammer to see if the safe was strong enough. And noticed no problems. Our new car model reduces the number of deaths by 100% (of 1 person), as we had one less fatal accident than the older model.

3. Systematic noise / bias.
   Our product does not cause cancer. ..Because we only measured the first week after consumption.
   Our test among students showed that most people that used our medicine are very healthy
   Lying down while shooting in random directions is safe. So this is what people should do during shootings.
   Crossing the street is dangerous, we should forbid people from crossing the street.
   The satellite shows that earth is getting bigger. Because the distance between earth's surface and the satellite is getting smaller.

4. Survivor bias.
   5 of 6 scientists confirm that Russian Roulette is safe and profitable.
   After punishing doctors for reporting problems, the safety of our medicine has improved immensely.

5. Reverse logic.
   Drinking alcohol causes depression.

6. Changing or misrepresenting history.
   Everyone who has no access to modern medicine, is no longer alive.
   We remove the measured historical data and replace it with our models.