“I can calculate the motions of heavenly bodies, but not the madness of crowds.”
– Sir Isaac Newton
In TOK, we have been looking at the area of human sciences and looking over many aspects. Is it a science? What makes a science? We conducted studies of our own choice in class and found ways to present our data. The key part of our presentations however, were to make actual errors in our methods and to show unreliability, as a way of finding even more ways to be reliable and as a practice of really examining data. Lolo and I conducted a survey asking people what school house they were in and if they were allergic to anything. Our presentation can be found here. The errors in our presentation were that we were assuming a certain type of causation which was already unlikely and our sample size was fairly small and not selected randomly.
There are many factors that contribute to (or take away from!) reliability and the certainty we can have in all experiments, not just human sciences. For instance the sample method and size. If the sampling method isn’t controlled or isn’t random there may be room for possible bias or discrepancy in the data. Additionally if the sampling size is too small, it becomes difficult to generate an analysis that could work for a larger population. There are also factors concerning the relationship of causation. Two variables you are looking at may not be actually related to each other. For instance we found that raven house was more likely to have an allergy but realistically there is not likely a causation between house and whether people have allergies. Types of questions can be a factor in reliability or certainty. Some questions may be leading or the way they are asked may provoke the interviewee to answer in a certain way. Going along with this there is also the possibility of the observer effect wherein people act differently based on whether they are being watched or not. Finally the actual presentation of data can impact reliability or the certainty we may have, if the wrong types of charts are used or data is misrepresented.
There are ways however that scientists can increase the reliability of their claims. The main way to do this is to increase the amount of data you collect. By getting more samples / trials then you can conceivably find more repeating data and be able to draw a more solid conclusion. In all cases science wise, the best thing to do is collect a lot of data and ensure questions and sampling are unbiased and that the presentation of data is appropriate.