Qualitative versus quantitative research in Science Education. One can argue for a method or another, or you can go on by not choosing. By not choosing I mean that you use both methods, in a mixed method setting.
I have mostly come in contact with qualitative method in my undergraduate and graduate training. This means that quantitative methods will be something new when I continue my training in higher education. If I have some training in qualitative methods I will be, and feel, relatively confident in using them. Therefore, I will see quantitative as something that exists but I see it at a distance, a method for others or other research areas. However, in my training towards the licentiate degree (educational specialist in US), I have read some articles who use quantitative methods, in this way it becomes more familiar what it is and how to use it. I’ve come to understand that it is ok to mix methods.
When it comes to collecting data it is usually not a problem, rather the opposite, to much data to analyse…in some way one might have to sort the data to find a glimpse of what can be valuable to look deeper into. Here one probably react to that all data is not analysed. Yes, one often has to sort the data several times to find the data that is worthy to analyse deeper. This is one flaw of qualitative methods. With quantitative data you can enter your data into a computer program where the analyse is made, here you can use huge amounts of data, if not all of it, because it is sorted and analysed with algoritms and computer power. However, it can be as strenuous as poking through the data in a manual way because you have to make sure, in some way that the analysis is correct. There are several methods to do this and they all have their pros and cons.
However, you could do this by hand if you like it. Just the way the woman who found evidence of Alfred Wegeners theory of continental drift, Marie Tharp. The mass of data was so huge it was written on a 5000-foot scroll (1524 meter). Read more of a remarkable woman of science here. She recalculated the sonar data several times, all by hand. The men in charge did not believe her, until Jacques Cousteau filmed the valley.
Just like qualitative research you need to code your data for the computer to be able to analyse it for you. Computer programs like SPSS (Statistical Package for Social Science) works similar to word processors like Microsoft Excel, rows and columns where you need to put your data. After you have your data sorted in a way so the computer can make sense of it you can do your first exploratory analysis. Here you try to explore what the data tells you. Confirmatory analysis tells you if you found what you intended to find. To make sense, you simply get some pie-charts, graphs or histograms that you can look at to find if you can make sense of your data. You can also make some analysis regarding variability or other measures to get a feeling if you are on the right track of analysing your data. Especially in flexible designs this initial analysis if of importance to try to find out if you have the right data or if you have to adjust something to get a more clear picture of your research object. It is perfectly ok to do this in flexible designs. However, you need to record your changes in such a way that you can explain to someone else what you have done in order for it to be apparent and clear how you did your research. Fixed designs often is reported with quantitative data. For instance, pre- and post test results that can be held against each other to try to find out if your treatment had any effect.
One analysis called exploratory data analysis (EDA) tries to simplify analysis by showing the data in pictures where you summarize the main characteristics of your data, often with visual methods. In this way EDA can tell you something about your data without the use of hypotheses or modelling. The hypotesers can be formulated after the EDA. Confirmatory data analysis (CDA) can be made after the use of EDA to confirm conclusion, in such a way that it complements the EDA analysis. However, the research community seems to not be in agreement on which quantitative methods to use is accepted. You probably need to ask a specialist or read Journals where you intend to publish your results in order to find which methods is appropriate and accepted in their eyes.
A very common test of your data is the significance test. Have you got a significant result? p < 0,05 ? The statistical significance is a test that tells you how likely you got the difference by chance alone or if it is not really a difference in the population that you got your results from. How plausible is it that you got no difference between your population. Simply put, it rules out the validity threat that the result could be due to random variation in your sample, rather than differences in the population. There are a normal distribution to most things. There are controversy to the statistical significance test as well. It is not related to the size or importance of a effect or relationship and the chance of a statistical significance increases with increased sample size . Therefore, there are a lot of different tests to try to rule out faults in conclusions or to increase the evidence of findings. Standard deviation and different cross tabulations to measure relationships between two or more variables. This is a lot to read up on if you are into doing something with quantitative analysis.
Real World Research second edition by Colin Robson
Wikipedia on Exploratory data analysis https://en.wikipedia.org/wiki/Exploratory_data_analysis