• Question: Is it hard to collect the correct data, and how hard is it too keep personal opinions out of it?

    Asked by storz001 to Christine, Edd, Jess, Nicolas, Zara on 16 Jun 2011.
    • Photo: Nicolas Biber

      Nicolas Biber answered on 15 Jun 2011:

      There are many things you have to consider when you are planning data collection. Before you start collecting data you need to be very sure what question you wish to answer with the data you obtain, otherwise you might end up collecting data that serves an entire different purpose. For instance if you are collecting data to find out how many dung beetles you find in a certain area and you thoroughly search the area for a whole summer, and then it turns out that you actually wanted to know how many dung beetles there were in that area in the autumn. Perhaps the most tricky bit is when you are actually collecting data and you are EXPECTING a specific result, you might catch yourself cheating so you get the result you want. This can be very subtle, especially when you collect data at random, sometimes you will ‘choose’ a ‘random’ area that has the most dung beetles in it, even though it shouldn’t matter. Finally when you report your research you separate ‘results’ and ‘discussion’. The discussion serves the purpose to interpret the results, this is where you state your opinion. The results section should only show the data and calculations that were made with the data, but often it is very hard to keep your own opinion out of the results section.

    • Photo: Zara Gladman

      Zara Gladman answered on 15 Jun 2011:

      Nicolas has already given a very good answer for this! 🙂 Here are some more thoughts….

      I agree that you must be very clear about what exactly you’re aiming to find out before you go and collect your data. Sometimes you will only have ONE chance to collect the data you need… for example, my friend is studying a colony of seabirds that only come to Scotland to breed once a year, betwen May and August. The rest of the year, they are in Africa, and so she can’t collect data on them!! (unless she goes over to Africa… but I don’t think that’s going to happen!). This means that my friend has to be very organised and decide before May which questions she wants to ask, and therefore what kind of data she needs to collect.

      Sometimes planning an experiment is easy on paper but when you actually try to collect the data in the field, it’s a lot more difficult than you imagined! I tried to do an experiment last summer to look at how crayfish impacted plants. I had everything planned out, but when I actually went out to my study site I found that the water was too cloudy to actually see the plants properly! Also, there were so many different species of plants that it was really difficult to identify them.

      Data collection may also be difficult because of the weather (especially in Scotland!!). Some days I have to cancel my fieldwork if it’s been raining too much because the river is deep and dangerous to get into.

      As for personal opinions, like Nicolas said, you can give your own ideas about what you think your results mean in the ‘discussion’ section of your paper (most scientific papers have an introduction, methodology, results and discussion section)… but these should be sensible ideas, based on evidence. A scientific paper must be scientific – you are not allowed to use your opinion on religion or philosophy to explain the results!

    • Photo: Christine Switzer

      Christine Switzer answered on 16 Jun 2011:

      Great answers already. It helps me to think about data and analysis as evidence. Rather than think about facts versus opinions, think about evidence and theories. The closest you can get to an opinion is a “hypothesis” which is a statement about what is happening in your system that you set out to demonstrate. You might have noticed that none of us are using the word prove to describe this process. In science, prove is a very strong word and it used rarely. Even if I demonstrate a hypothesis, I don’t consider it proven because someone else might come along with fresh ideas and a different set of tests that show it is not true. It takes a really long time and a lot of evidence for something to be considered proven.

    • Photo: Edward Codling

      Edward Codling answered on 16 Jun 2011:

      As others have already said, science is ‘evidence-based’ which means we need to have evidence – usually data that matches the idea we are trying to put forward.

      However, often a certain data set could correspond to different possible theories or explanations. The scientific opinion then centres on the interpretation of the data (not the actual data itself).

      For example, in my studies of animal movement it is often possible to observe a particular type of movement path that looks like a ‘fractal’ (the path looks the same if you zoom in or out). There have been lots of arguments between scientisits about what causes the movement path to look like this – we agree on what the data looks like but not on how the data is created!

    • Photo: Jessica Chu

      Jessica Chu answered on 16 Jun 2011:

      Lots of good answers. Likewise – if I want to see certain results that my plant DO affects the cancer cell lines then I try to pick at bits but the later on when I take a look again I have to be quite critical about my own work because it might be nothing or I will have to plan to repeat the experiment to see if I can get the same interesting response again!
      Because I do at least 3 consistent repeats then it kind of shows this probably does happen. It is important to choose the right experiment to do so you are not just doing something for the sake of it but it will actually tell you some really important things about what you’re working with.

Comments