Controversial is also way of collecting data by social scientists. Their main tool is a questioner and an experiment; both when analysed deeper show a lot of disadvantages. The main proved is that people when they know that are �investigated� usually do not behave as in reality, but try to achieve, what they consider, an �ideal person�. People try to hide their defects and exaggerate virtues � �we want to see ourselves better, than we are in the reality�(1). P. Cross carried an experiment in which he asked USA college teachers: if all college teachers were sorted with respect to their teaching skills, so that in the middle of the list would be an �average teacher�, were would you been above or bellow him?. This opinion pool showed that 94% of American college teachers consider themselves as better that an average college teacher(2).
Unfortunately this optimism is against the laws of statistics. No more than 50% are better, and no more than 50% are worse than average (in this meaning of the word �average�), but in the experiment only 6% of teachers admitted that they are worse than average teacher is, so the relative error of this measurement were about 730%!(3) Well, I do not think we can talk about any outcome (in terms of its truthful) if the uncertainty of measurement is so big. In this case error could be easily calculated, although in most of cases we have no opportunity to verify obtained results, that is a strong reason why we should be more sceptical when analysing any statistical data.
The next problem that sociologists face is that they can never ask questions everyone in the population, but they every time have to assemble a representative test group. The question how to choose people to this group, what criteria should they fulfil, is still open. Every time an experiment is made, it is done on a test group, although scientists usually generalise the results on the whole population. Anyway what is �true� for the test group, is not certainly �true� for the rest of society.
The modern trend in social sciences is to present results in numbers � to quantify them � because numbers are considered more trustful, and data presented in charts look clearer. Numbers have that advantage over language that are not ambiguous � �two� means always �two� and never �three�. Unfortunately when we give assign a meaning to numbers they starts to be no less ambiguous than a language. Data showing the percentage of unemployed people are worth nothing as long as we do not know the exact definition of unemployment used in the measurements. Well known is that the meaning of word �unemployment� is usually changing before and after every election. In this way statistics are very often manipulated. Although, we are used to compare levels of �unemployment� in different times and in different countries, the worst is often that some conclusions are drawn from those comparisons. Therefore, I do not think we should consider works based on statistics as believable, though we cannot also say that they are complete fiction.
Uncertainty appears also when collecting quantified data. When we ask people to mark on a scale one up to ten, how stressful job do they have, because of different perceiving the numbers by people, we get distorted image of reality. For some ten on a scale may be equivalent to stress on a school exam, and for some a stress similar to this you fell when your both parents die.
Looking from that point of view we can consider the majority of statistical, and sociological works as just worthless, although there is some truth hidden in them � even theories dealing with sociological topics frequently turn out to be useful. The problem is that we cannot easily distinguish between the truth and false, and even we are not able to assess the level of deformities made during data collection and data analysis process.
The consequences of it is for example that some billion dollar advertising campaigns, made by the best specialists, based on the most modern theories, just bring no effect on selling level of advertised product, while other made against the trends sometimes become great successes highly and positively affecting consumers.
Paradigm shifts in our knowledge arise also because we cannot separate the truth from the falsehood. We believe in theories that others prove them false. In the social sciences they happen all the time, and therefore they are not so big sensations like in natural sciences. Sometimes contradictory theories exist even simultaneously, just like those two concerning our psyche. The first, modern says: �psyche is determined mainly by genes�, and Freud�s � older but still in use says: �psyche is determined by a childhood�. There are lots of arguments pro, and against both - they can exist together because the truth is somewhere in between, although we do not know where.
More spectacular are paradigm shifts in natural sciences. They alike prove imperfections of our knowledge but in non-social fields. I can give at least two up-to-date examples of paradigm shifts in natural sciences. First the recently disproved physical principle that an electron is the lowest electric charge freely existing in the universe, and second the discovery proving that neurones similarly to other cells can reproduce themselves in the adult age. Many physicians and biologists still can�t believe that they were in so deep mistake.
Otto Neurath�s compares knowledge to a ship, and scientists to sailors that continuously repair it by replacing old used up planks(4). Most of planks decay with time � people become aware that they were in mistake � that they confused truth with false. Drawing further conclusions, we can say that we confuse truth with false all the time, being not aware of it. Well, it is not a revolutionary statement � Aristoteles already said �To err is human�(5).
Number of words: 1231
