How to process data

Data processing for scientific papers: Tips and tricks for statistical data analysis.
est. 2004
Redaktion | 09.09.2019 | Lesedauer 3 min

Data processing is the part of your research where all the data you have collected will be turned into useful information that will enable you to find out if you can verify or disprove your hypothesis.
Of course, all research is unique, even within the same subject and even more so between different subjects. So there’s not a “one size fits all solution” but there are basics that you should know and take care of.

First Step

The first thoughts concerning data processing are required when you are still working on the design for your research. Make sure that relative to the phenomenon you want to deal with, the sample you have chosen is big enough to produce significant results, or that it is composed in such a way that will allow you to generalize the results of your research.

The choice of your sample is dependent upon your hypothesis, and both of them will decide in the end whether or not the effort of collecting data will lead to a reasonable and satisfying outcome. The reliability of your results also depends as much on the whole design of your research as it does on the sample. It goes without saying that the better the sample, the more reliable and more valid your results will be.

If you are not sure about how to formulate your hypothesis and how to choose a sample, find support and inspiration here and/or talk to one of our experts. They would be more than willing to assist you in developing the details of your research design. If you do everything right from the start, you will easily be able to collect the data you will need and processing them will be a mere formality.

Once you have collected your data, it will have to be conveyed into a form that will allow processing–so we are talking about creating input for an IT-based analysis here. Make sure that your data is complete, if not, see if you can complete it. If you have been working with research tools like questionnaires that allow open answers, you will have to find a way to encode the answers to make them quantifiable.

Second Step

The next step then is data-processing. It is the operationalization of the different computer applications of which software is composed. You may not need them all, but you should decide what exactly you need to find the software that is most likely to produce the results that you are looking for.

Normally, data-processing starts with a simple frequency counter to create a set of univariate data, then it continues with relating different variables to create bivariate or multivariate data, according to your research design and your hypotheses.

For most research projects it is important to find out about the average frequency of cases (connection of two or more variables), and their lowest and their highest frequencies.

The output of the processing serves as a report of your data’s potential to answer your questions, to give a statement about your hypothesis, still coded in a set of figures or maybe in graphs, awaiting your interpretation.

Do not forget that you have to consider two different levels of interpretation here: firstly, there’s the evidence concerning your topic or the answers you were looking for. Secondly, you will have to find out about the significance of these answers by relating them to the whole of the sample and/or the assumed whole of possible cases, i.e. the phenomenon as a whole. In general, the software used for data processing is creating these figures automatically, except for the last one, so you just have to make sure that they will be included in your interpretation of the results.
Also, be aware that if you want to deal with data you have to think about its security and its storage as well as about possible infringement of participants’ personal rights through your research.