Hi Florentina, I couldn't agree more that data interpretation is as important as data analysis, in fact, I would say critical in developing sound conclusions for one's research.

It all comes down to the type of research conducted, i.e. quantitative vs. qualitative and the epistemological stance used in the research (e.g. positivism - testing of theory OR interpretivism - building theory)... While a quantitative , deductive approach some would argue is easier to deal with in terms of data interpretation as it comes down to interpreting the quantitative data/numbers (usually obtained through surveys) and the statistical tests performed on this data, with qualitative data analysis and interpretation the process is not a straightforward one, it is an iterative process, going back and forth from data (i.e. interview or focus group data) to the researcher's interpretations.  

In both cases however, either quantitative or qualitative research, once data is analysed and presented, there is a crucial stage of the process: the Discussion (OR ‘as you go’, i.e. iterative process in qualitative research).

In this section, you now compare and contrast your findings with the existing literature. In essence, you bring the key literature elements back to see if your findings are supporting, similar or different in some way – this is a critical discussion (critical writing). This then informs your recommendations and conclusions.

Hope this is helpful.  There are many good textbooks in Research Methods that can put you in the right direction around data interpretation depending on the type of research conducted and research paradigms.

Best wishes,

Anca

Hi Florentina, I couldn't agree more that data interpretation is as important as data analysis, in fact, I would say critical in developing sound conclusions for one's research.

It all comes down to the type of research conducted, i.e. quantitative vs. qualitative and the epistemological stance used in the research (e.g. positivism - testing of theory OR interpretivism - building theory)... While a quantitative , deductive approach some would argue is easier to deal with in terms of data interpretation as it comes down to interpreting the quantitative data/numbers (usually obtained through surveys) and the statistical tests performed on this data, with qualitative data analysis and interpretation the process is not a straightforward one, it is an iterative process, going back and forth from data (i.e. interview or focus group data) to the researcher's interpretations.  

In both cases however, either quantitative or qualitative research, once data is analysed and presented, there is a crucial stage of the process: the Discussion (OR ‘as you go’, i.e. iterative process in qualitative research).

In this section, you now compare and contrast your findings with the existing literature. In essence, you bring the key literature elements back to see if your findings are supporting, similar or different in some way – this is a critical discussion (critical writing). This then informs your recommendations and conclusions.

Hope this is helpful.  There are many good textbooks in Research Methods that can put you in the right direction around data interpretation depending on the type of research conducted and research paradigms.

Best wishes,

Anca

Dear Anca,

Thank you very much for your answer. It is helpful!

But is the interpretation stage an independent process or an optional one? We just discuss at the end of our researches our results comparing and contrasting them with existing ones?

I think we have to state clearer our researches relation with the existing theories and I think we have to do it earlier in our research process. What do we do? Test an old theory, see if one applies in a certain setting, modify or complete one, discover a new theory etc.

 I think we have to be more rigorous about data interpretation. We have many rules about how we have to collect and analyze data, about how to make it valid etc. But when it comes to interpretation - anything goes if we discuss it?

I like the idea of the 'art and science' of the interpretation of data. While we must obey, rules, conventions etc., we also need to use a lot of our thinking about the epistemological context of the finding within the academic literature. I have some links to books most formative to my education in analytical interpretation:

https://www.amazon.co.uk/Statistics-Science-Learning-Books-Carte/dp/0133860825/ref=pd_lpo_sbs_14_t_2?_encoding=UTF8&psc=1&refRID=FS63Y2K7QE7NS14Q04JH

https://www.amazon.co.uk/d/cka/Structure-Scientific-Revolutions-Thomas-S-Kuhn/0226458083

https://www.amazon.co.uk/d/Books/Conjectures-Refutations-Growth-Scientific-Knowledge-Routledge-Classics/0415285941/ref=pd_sim_14_4?_encoding=UTF8&psc=1&refRID=DF8E0EMC8E4Z2BZ1R0JE

We learn a lot about data collection methods and techniques and about data processing and analysis. There are lots of books and articles about those research stages (being it qualitative or quantitative). How about data interpretation? Isn't it the last stage of the research? Why do we neglect it?

I'm not sure it is neglected or there are other explanations.  One explanation I can recalled is each interpretation is dependent on the specific empirical results after the data analyses in which these results are different from one research to another. However, one of the viable ways to learn data interpretation is to perform literature review on articles focusing on their discussion section i.e. immediately after the result findings section.  That's why to some researchers / scholars, discussion section is one of the most difficult sections in thesis / article to write as it is based on your unique result findings.

Do we always formulate new theories or we do that only in grounded theory studies?

Depending on your research problem, research objectives & research questions.  Quantitative research tends to test / confirm certain theories / constructs quantitatively whereas qualitative research tend to probe deeper to find out the underlying insights in which grounded theory approach can be used to discover new theories.

Do we test old theories or we do that only in quantitative studies?

Again depending on your research problem try to address, its research objectives & research questions.  Quantitative studies can be used to test certain theories.

Do we apply old theories in new situations or do we modify or complete old theories?

Again depending on your research problem, objectives & questions as mentioned above.  In quantitative research, your newly developed conceptual framework / research model needs to underpin on previous / existing theories / theoretical frameworks as you can't develop your framework / model out of the vacuum.  Depending on your research empirical results - based on that you can modify / complement previous theories.

Are there studies where we do not employ theory at all like when we ask people who they vote for?

Even though we ask people's opinion e.g. through certain questionnaire which is part of your methodology based on your conceptual framework / research model - those framework / model should underpin on previous / existing theories / theoretical framework which can serve as "swimming life jacket" to keep you afloat.  Without underpinned on any relevant theory / framework, your conceptual framework / research model might be questioned by thesis examiners or journal reviewers.

Thanks for response @Han

The data interpretation stage is NOT an independent process of data analysis neither is optional. In fact, as I mentioned before, and as Han Ping Fung notes above, it is an important stage that flows from data analysis and presentation of results.  The 'Discussion' stage IS the interpretation stage.

Now, again, this delimitation between results and discussion in qualitative research is a bit fuzzier than it is in quantitative research with a 'Findings and Discussion' section often seen, as opposed to clear cut sections on 'Results' and 'Discussion' and this is because, as I explained before, the interpretation process in qual research is an iterative one.

You are 100% right, any research should be clear from the outset about its aim, objectives and the methodology used to achieve these.  Thinking about these in the data interpretation stage is too little too late.  I have seen a number of studies/articles that present a 'Research Methods' section describing just 'the methods' used in data collection and analysis.  Much more thought needs to be given to one's research philosophy, how they go about generating that knowledge (testing or building theory), and methodologies which will then inform the methods used in the research.  Although used interchangeably, the terms 'method' and 'methodology' are two different terms with different meanings. It all comes down to how you design your research  which will then have implications on how you will analyse and interpret the data (the attached slide might be useful to visualise these differences).

The rigor in data interpretation comes from staying true to the epistemology and methodology employed and showing rigor in the discussion of results and an ability to make a sustained argument on the basis of appropriate evidence (linking this back to previous literature).

 Dear Florentina Scârneci-Domnișoru, I agree with Anca Cristiana, Yallop Han Ping Fung, that  data interpretation depends on the type of research done -qualitative or quantitative- or the logic of research ,and  the kind of theories tested .In my Ph.D. research , "Students& Radical Social Change" (Calcutta University 1991)which   has been published in Book form by The University of Burdwan(See Anirban Banerjee:"Students& Radical social Change , The University of Burdwan,2003), I used Marxian methodology , which is  a variant of positivist methodology. Mine was a  hypothesis testing study .So, I used the technique of social survey, the type of sampling adopted being stratified random sampling,.I tested a number of theories, like alienation and generation gap. To interpret data , I used tables and graphs.

Anirban Banerjee

The University of Burdwan    

I think the philosophy of quantitative and qualitative research is clearly different. The theory in the quantitative approach intends to be tested whether appropriate or not, accepting or responding to hypotheses. Hypotheses are built on theory. In contrast to the qualitative approach, which makes the theory as a guide not to test its truth. Even if later the results obtained did not mean to test the theory but tend to produce something that can be a modification theory, or also in the early stages can be a proposition. Interpretation is one way to understand and interpret the information gained during the research. Our sensitivity is certainly needed in addition to experience in using qualitative approaches. Our interpretation must be understood as full of value and not value-free. Thus, we do not assume that our interpretation is constructed from "something" we have ever read, open from research information. It's not easy, but when you're used to it, it's easier. Thanks.

Hi Florentina. 

The Data interpretation is a crucial part of research, mainly what it refers to the reliability and Validity of Data. I recomend you the publications of Dr. David L Morgan in wich you can check the specif aspects that you have to consider for data interpretation. I also recomend you our Book: Data Analytics Applications in Latin America and Emerging wich is available in: https://www.crcpress.com/Data-Analytics-Applications-in-Latin-America-and-Emerging-Economies/Rodriguez/p/book/9781498762762

Hi Florentina,

I think you made a very good point. First, while data interpretation might also be a separate part of qualitative and quantitative studies, it is an inseparable part of all data collection and data processing. Whatever we do with data is inevitably based on some understanding of what they might mean (i.e., interpretation). Even in quantiative research, wher data processing is apparently based on formalised rules, there is a number of decisions where the meanign of data to the investigator shapes the processing.

Second, the textbooks on qualitative research I know usually avoid the point or work with illustrations, and for good reason. I dont think our qualitative research methodology has already reached a stage where we can provide rules saying: "this is the text, these are the circumstances of its production, these are the circumstances under which you interpret it, ergo this is the meaning of the text". Interpretation is assigning a meaning to a text, and so far this is by and large a personal act. We do not know how we arrive at the opinion that a particular text has this meaning and no others. Unless we do know this, we cannot formulate rules for interpretation.

Thi is one of the reasons why we put quotes from texts in our publications. We want the reader to know how we arrived at interpretations from text, and we cannot do this except by providing examples.

There are workarounds. I think that interpreting data collectively (either by having agroup of researchers discussing texts or by having parallel interpretations and comparing them) is a good approach because explaining differences in interpretation forces one to explicate the reasons why one arrived at a particular interpretation (at least as far as thi is possible at all). But these are workarounds.

knowledge is situated. research interpretation depends on the knowledge you have on the subject and the insights you can draw from the the data.that is research or data interpretation.if you are not situated or anchored in any theory or field of inquiry interpretation becomes a mountain.

Hi Florentina, I agree with you but I understand the analysis itself should be very complex and include the interpretation as the core part. As Peter wrote, the knowledge on the subject makes the real value added of every research. More precisely - knowledge of statistics, goes  along with knowledge of subject and context.

It is difficult to come up with a book on data interpretation because every research is different and one cannot even anticipate the results in advance. After testing the hypothesis (if applicable), it could be accepted or rejected. It is after the statistics have been generated by the tools, that we then start interpreting in the context of the research objectives, research questions and conceptual framework.

Dear Florentina,

Thank you for asking this important question.I think that data interpretation is the most important part of the research.If you cannot interpret your data your research falls flat.But I agree with Otete that it is difficult to teach data interpretation since the requirements of every researcher is different.

ANIRBAN BANERJEE

Professor of Sociology

The University if Burdwan

INDIA

Since u r talking about data interpretation for teaching purposes, you should give due attention to it from very beginning...Anca has explained your query quite nicely covering the underpinnings of research methods....I wud just want to add lil to that...

Even before reaching discussion, we pass through the process of data interpretation (partially)...how? Ok, in qualitative research, let's say u r doing a phenomenological study and u r collecting data from a participant who is tsunami victim...now while getting data from such participants, there will be several occasions when there will be moments of crying, tears, silence, anger (towards the system or so), sadness with some words or no words being uttered at all...it means what? No data? There is amazing data here which u need to understand and interpret then n there itself in a very careful way which in fact will be tougher than data interpretation carried out in the end (which is usually done in standard conditions) which u r referring too...same way, even in quantitative research, while making the research framework itself, we need to ask ourselves imaginably that what type of data will expectedly come from which set of variables... Thinking about such data (which is still not collected), whatever framework u have, u can try to imagine what will be the probable level of explanation or probable r- square (obviously after referring to some related meta analysis or some other relevant/similar studies)....it will also give u an idea/direction pertaining to  whether something useful (variables, items for certain variables, or new supporting theories, etc) can still be added.... so, in summary, data interpretation might begin much before reaching the formally worded 'discussion' section....I hope it helps...tq