I think of "big data" as wide but very typically very thin -- in this case, "scraping" is a good metaphor, as in "scraping the surface.' In contrast, qualitative methods, and especially enthnography, may be narrow but they go deep.

Why bother with ethnography?

Not sure other RG members share the same thought of your students' comment above. Personally I find that ethnography is a subset of qualitative research that have its own way to contribute to knowledge i.e. through immersion, observation, interaction & practically live among the subjects to extract insightful information & knowledge which are not possible via other methods. Ethnography research's knowledge contribution is not based on the quantity of sample size & data analysis techniques as per quantitative research.

Not all problems have big data to aid deliberations.

It's a trade-off between reliability and specificity but it all depends on your research objectives and actual research questions.

Big data may not be congruent with one's hypotheses or other empirical segments; while smaller samples may accentuate findings toward current and future conclusions; as well as supple,etal research questions.

I think all of the answers that have been given above are excellent and I agree with these. However, I have had some students who just shrug and say give me the numbers as I can use this to make decission about my actions with confidence and ethnography does not give me this. Do you have any thoughts?

I think of "big data" as wide but very typically very thin -- in this case, "scraping" is a good metaphor, as in "scraping the surface.' In contrast, qualitative methods, and especially enthnography, may be narrow but they go deep.

Big data does not necesarilly means good data. And for some purposes the size of a sample does not imply robustness in the results. Also, numbers do not imply objectivity.

Paul M.W. Hackett One activity I've used a number of times is to make students (whether they shrugged at your explanation or not) work on small and extremely localised problems that directly affect them, not as a group but individually. So, they would not just be researchers but also potential beneficiaries of the research findings. In one place I worked parking was a perennial problem so I told them that after many months of observation, experiments, interviews, etc., I had found the perfect parking spot for any time of arrival on any day. Students would be interested but of course, I wasn't going to share. This is not a case where they can just go and scrape "big data" or even do an in-depth literature review. Rather it might be a bridge into some of the other research designs and methods (including ethnography).

"...why bother.." is inherently a biased question, for any subject/item. once again, i say what is neeed is a course in philosophy of science and/or epistemology. i did not get this until phd program and would certaily have appreciated it in bs or masteresprogram; unfotunately, all i got in resercah courses at the time (bs and ms) was the experimental/quantitative. with such lessons, they would not {have to} ask, "" why bother..."

If students are thinking that they can just use quantitative data--give me the numbers--to make decisions about which they can be confident, then they really need to learn more about the problems with quantitative approaches. Analysis of large data sets often sacrifices detail and complexity, meaning that it is much more difficult to know what people actually do think. Forced choice questionnaires often put subjects in the position of having to make a choice that they wouldn't normally make in answering a question--their actual thinking is considerably more vague or complex than the questionnaire can capture.

I use, and teach, both qualitative and quantitative methods. I'm an ethnographer by training, but I have administered questionnaires and done work with statistical methods. There are advantages and disadvantages to both approaches. When my own students ask similar questions, I often respond that qualitative approaches like ethnography are often the means by which we can generate good questions for quantitative approaches. Without qualitative research, I think many quantitative studies would suffer from asking the wrong questions (well, many already do despite the presence of qualitative research). This is particularly tru when trying to do cross-cultural research. Ethnography is one very powerful way of developing enough understanding of a context to ask questions that are meaningful.

Paul M.W. Hackett

I think all of the answers that have been given above are excellent and I agree with these. However, I have had some students who just shrug and say give me the numbers as I can use this to make decission about my actions with confidence and ethnography does not give me this. Do you have any thoughts?

I would counter this first by saying that this student's statement displays a general misunderstanding of inferential statistics. You can infer information about the whole population from a good sample, but you cannot re-apply this information to individual cases. This is essentially the same as what is commonly called stereotyping. In inferential language, it is called an ecological fallacy, and it is rejected in science just like stereotypes are rejected in political discourse.

In this vein, ethnographic research can actually give you more confidence in making decisions, if you're lucky enough that someone has conducted an ethnographic study that covers your exact issue at hand. Given enough similarity to your situation, ceteris paribus, reading an ethnographic study can give you an understanding that will allow you to anticipate things with a much higher level of accuracy than pure numbers will. One example for such an issue that each student should be able to relate to is the school grading system. At first glance, it reflects the students objective knowledge, but we all know that oftentimes there are infinitely more factors at play, such as the student's reputation, test aptitude and even things as simple as handwriting aesthetics, which are obvious to the ethnographer, but near impossible to put into numbers. Good students (subconsciously or consciously) are those who have all of these factors working in their favor, and students can significantly improve their grades if they understand and exploit this.

Aside from this, it is really important to reiterate that quantitative methods only offer an extremely selective view on social realities. As a simple example, you can ask someone about their life satisfaction in a survey, and you get a number between 1 and 10. Or you can lead a 10 hour interview with the same person and understand every facet of that person's biography, social environment, personal preferences etc.. The same is true for business decisions (which may appeal more to some of the respective students). You can ask someone for the reasons for their business success within 5 answer choices, or you can lead weekly interviews with them over the course of 5 years and understand every detail of how they grow and advance their business. The result won't be significant at the p=

Many perspectives can inform a deeper and wider approach to solutions that are sensitive to human needs.

Statistics can be a valuable tool of researchers to understand trends of health and illness behaviors. However, I have found that they have their limitations in answering why people do what they do in their everyday lives. There are no shortcuts to understanding. Ethnographic methods hold the promise of getting close to what the participants find to be important. If one approaches these studies from a humble perspective as a "listener and learner" (from the Spradley perspective), the participants enjoy being the "experts" about their lives. Spending prolonged time with these participants provides opportunities to develop trust and gain access to the meaning making in their lives. I have always found surprises in my ethnographic work. Things are not always what they seem on the first glance. This inductive process engenders respect for the actual resources and challenges of the participants. Hopefully, this leads to the development of solutions that are grounded in their realities.

John Traphagan-I so agreee with this approach. I used a real life example, that actually occurred to help students understand/see why often numbers are not insufficient and can lead to errors in making decisions.

Many yeasr ago, I was conducting program evaluations, including one of a teen pregnancy program .Results from quantitative factor collection (records/not questionnaires), showed the progam was a total failure among latino girls-(recent immigrants from mexcio, specifically)b ecause they were not meeting some of the most important/significant program goals: (no) additinal pregnancy, use of birth control, and lack of participation in education (going to school, completing hs or ged.

Further investigation (mine) also showed this group was different in three areas: had lower grade /level school achievement, were older, and married, with working spouse. A smaller study-qualitative interviews with this group showed that: they were family planning (another pregnancy); catholic- no birth control, and responded that they were raising very young children at this time and could/would return to education (goals) when they children were older.

These cutural factors/differences would never have been discovered w/o qualitatice methods. (we/agency would have lost some funding). I am sure there are many instances where such further analysis indicates very important information that is missed so often with just, ..."give me the numbers" and big mistakes are made when used ..."to make decisions."

The fastest answer is to turn this back on the student. So, how does the student envision the research to go when the study subject is a small tribe in the Amazon that has never encountered Twitter, and maybe the survey monkey can be roasted and eaten. Good, yes?

That can be followed by questions like how representative do you think twitter use is amongst the local demographic. Do you really think that all age groups, income levels, gender (or gender identities), ethnicity, are all equally represented through something like SurveyMonkey? Please discuss, and broaden to include global dynamics.

A twitter sample is just as likely to include a Tibetan monk as the president of the United States?

This can then be a useful discussion on clearly defining research goals and what can be done with data. So my population is "all people who use Twitter" and my sample (might be) influenced by the frequency of use from different groups within this population. I can now say that 23% of Twitter traffic is about coffee, so what does that tell me about the market prospects for coffee, or the demographic of coffee consumers, or anything.

Not bother, use it. Well applied ethnography gives a lot of knowledge of context. In social sciences even autoethnography, by definition one single propositus, ads valuable information, like any case study.

I did my postdoc at a Population Studies Center and was brought in precisely because the demographers wanted interactions with ethnographers to help with identifying questions that are appropriate for different cultural settings. They brought in several cultural anthropologists over the years.

Even if it's not necessrily ethnography, statistics /quantitative is very often not enough--why many researchers opt for a mixed method approach--

I am not going to be part of quantitative v qualitative research debates. To me, before one goes to that far, it is important to understand the ontology and epistemology of your research. What kind of information do you need and how to get those information. Once that is understood, it will answer the questions why ethnography. I find its not worth to justify the answer without explaining the root cause.

Dear Rajib, I totally agree with you. understanding and responding to the ontological and epistemological issues regarding your research are essential.

Perception doesn't have high importance in scientific research. Still if we go for schedule rather than survey, we can avoid biasness so small but acute is more important.

hmm?? peception is not ethnography

Anjay, I'm not sure I understand your point here. How do you define scientific research? And I'm not sure how perception could not have high importance in scientific research. If you think about the fact that all research involves the perception of the researcher, then it automatically should have high importance. And with ethnography, much of what we do is try to understand the perceptions of others. I really don't understand the point here. Can you elaborate?

I would refer them to the ergodic theorem which, as Molenaar and Campbell (2009) write, addresses the following question: "Given a particular set of selected variables, under which conditions will an analysis of interindividual variation--an analysis in which information is pooled across subjects--yield the same results as an analysis of intraindividual variation?" (p. 112). What this question essentially alludes to is the fact that we simply cannot generalize from the group (in terms of group means) to the individual. Of course, all qualitative researchers are familiar with the statement "due to the small sample size, these findings are not generalizable," but this idea equally works in reverse. With larger and larger participant pools, it makes it harder and harder to notice "the common patterns as well as the uncommon patterns and exceptions" (van Geert, 2008, p. 189).

Ethnography is all about subtle nuances. Not only is human behavior vibrant, but it is idiosyncratic. Big data necessarily smooths such idiosyncrasies, and worse, often asks us to overlook the 'black swan,' negative cases that might refute our theories when they turn up as outliers in our data set.

A small sample of ethnography can be taken as a Case. In educational perspective, cases are significant in that from the cases, larger longitudinal studies can be carried out. Ethnography can include a large sample or a small sample in the context of school organizational setting.

I use Tracia Wang in my teaching. She is an excellent example of contemporary ethnography, I agree with you.

You may widen the concept of ethnography from Social Sciences to the part of "social" Medical Sciences have. Case studies are in the core of medical reports. Small samples cover a very ample ground in Medical Research. Include social data in any Medical research ads new meaning to many situations. And the perception of the clinician is still a strong consideration in the medical reality.

All that is Ethnography. So you better bother with it...

Ethnographic and other qualitative approaches, almost all with small samples, have their merit in the possibility of inspiring one to ponder new variables, choose between multiple possible variables, conceive associations and even create whole models. Its their open-ended nature that is the key for that. In some cases, well-crafted research can also find counterexamples that put into question assumptions or models.

But that is where their utility stops.

If one is looking for knowledge where reliability and accuracy, however great or small, can be clearly established, where the results can be translated into clear-cur predictions, and/or where models can be simulated in computers to explore scenarios, then it is mandatory to go for quantitative approaches.

Paraphrasing Thomas Alva Edison's " one percent inspiration and ninety-nine percent perspiration ", some 1% of all research should be ethnographic/qualitative and 99% quantitative.

P.S.

Any and all qualitative studies that forgo the very concept of variables or models are, by definition, useless.

positivism in the 21st century is, "by definition, useless." and one man does not decide the purpose of qualitative (or quantitative for that matter) research. pick up a book on the issue that goes back several decades.

No one mentioned "positivism", which is, indeed, useless. Has been know to be so since the first half of the 20th Century at least (see Popper and Lakatos). But this has no bearing whatsoever on the issue of quantitative vs. qualitative studies. Period.

It might be true that "one man does not decide the purpose of qualitative (or quantitative for that matter) research ". But it is definitely true that one man (or woman, for that matter) may state it. Just because one says that the sky is blue does not mean that it is so by his or her decision. And, for the record, usefulness is given by utility in the prediction and/or control of the observable world.

Positivism, on the other hand, can be boiled down to the belief that the "real world" is directly observable through specific methods. In that last regard, most of the true positivists out there are those who insist on qualitative approaches as being the best or even the only way of actually assessing reality. The rest of us know that we are only proposing models that are valued not by their truthfulness, for that is impossible to assess, but rather by their utility.

thanks for the lesson- i suggest you bone up on philosophy of science and epistemology-- i don't give lessons except to my research students in the classroom. qualitative research and studies are and have been accepted in the science world for over 50 years and one man's opinion, especially flawed will never change that-- there's always one in the crowd and as a matter of fact, i know several colleagues in the "hard" sciences who consider psychology (any branch) a non science and have laughed at it (my specialty along with psychiatry).

Bruno Campello de Souza identifies usefulness with prediction and control but I only partly agree. What happens with understanding? I think that in social sciences we must seek to understand the phenomena and not only describe and control it. Predictability might be the main issue in natural or hard sciences but not in social sciences.

But, of course, this is just my point of view.

Antonio M. Oller-Marcén proposes "understanding" as a part of "usefulness", along with prediction and control. However, he fails to define exactly what does he mean by his proposition.

I can agree with the idea if by "understanding" one refers to the cognitive ergonomics of scientific models, i.e., that they should be as easy to learn, test and deploy as possible (though no easier than that, as Einstein put it). But that is basically Ockham's Razor.

Any other meaning of the term must be objectively defined before it can be evaluated as being or not a valid addition to prediction and control in the comprising of usefulness.

Unlike what happens in hard sciences, when we are dealing with social sciences, for example, we can aspire to some kind of teleology.

When can "define" the term "understanding" as finding reasons, motivations, etc. for those involved in a phenomenon... in social sciences, I repeat. This is where, in my opinion, qualitative approaches are more helpful than the quantitative ones.

Mary-Helen Castanuela , here is another philosophical lesson for you: stating as an argument that "qualitative research and studies are and have been accepted in the science world for over 50 years" is a well-known fallacy referred to as argumentum ad populum (look it up). Basically, it is a gross misdirection where one addresses the popularity of a proposition rather than its validity.

Anyhow, I entered this discussion stating the value of qualitative methods in inspiring one to ponder new variables, conceiving associations, creating whole models and producing counterexamples that falsify assumptions. I did say that the value in question is limited to such things, but it is far from being something to scoff at.

As to a large portion of Psychology being a non-scientific, the infamous "replication crisis" is a strong indication that the field is indeed rife with bad science. The exceptions seem to be the sub-fields of Cognitive Psychology, Personality Psychology (Big Five) and, in spite of the name, Experimental Philosophy, all of which are quite quantitative.

It is quite obvious that qualitative methods and models are not valid foundations for science, though they can and should be used as an invaluable element of creativity that can both broaden and deepen the scope of knowledge. Just use them for inspiration, not for building.

"Teleology" in science is quite dangerous. Just look at Nazism and Communism for examples.

Science should be practiced and used under moral and ethical constraints. In other words, it should submit to them. It must never be the other way around.

again Bruno Campello de Souza is the one lacking "understanding" if he doesn't even know the meaning of the word along with little knowledge of science. such narrow mindedness belongs in the dark ages. i repeat-go back and study philosophy of science and epistemology

I have no quarrel with Bruno Campello de Souza's basic point that the strengths of qualitative research for generating hypotheses are quite different from the strengths of quantitative research for testing hypotheses.

But I fail to see how this relates part of the replication crisis. As far as I can tell, the core of that "crisis" is the inability to replicate many studies that involve social psychology experiments. Hence, it counts agains the practices of quantitative rather than qualitative researchers.

Indeed, I do not know of anyone who claims that the results of specific qualitative studies are strictly replicable, nor should they be if the goal is to generate insights that are potentially worthy of further study. And that further study does not need to be limited to quantitative research, because there is obvious value when a variety of equalization studies on a topic all reach similar conclusions. But perhaps finding a repeated pattern is evidence for replicability?

For me the main difference between quantitative and ethnographic research lies in the nature of cases. Quantitative cases are a number of values for each variable. The advantage is that cases are all comparable as they are perfectly analogous in their structure. The disadvantage is that there is a lack of flesh around the bones, the variables and their values. Qualitative cases cannot be compared mechanically. They are like persons with all their attributes. You gain an entirely different insight into reality. You can become aware of a lot of details, without even noticing the bones. Both ways for representing social realities are attractive, but on opposite grounds. The coexistence of both types of research seems necessary to me. Reading a text like Alice Goffman's "On the run", which represents an ethnographic extreme, I am always thankful, when she opens toward quantitative research through a number of quotations. The observation also works the other way round: it is always an attractive moment, when a quantitative researcher inserts a reflection in a qualitative way. There is a simple reason for this ambivalence of social science. The tension between quantitative and ethnographic is a basic expression of the two faces of social research. In the long run we social researchers do need both.

I am reading and writing an educational ethnography from this juncture;

"Term ethnography has come to be equated with virtually any qualitative research project where intent is to provide a detailed, in-depth description of everyday academic life and practice in higher education. Sometimes referred to as "thick description" -- a term attributed to anthropologist Clifford Geertz (1970s) early writing on an interpretive theory of culture (e.g., see The Interpretation of Cultures, first published as a collection in 1973). The use of term "qualitative" is meant to distinguish this kind of social science research from more "quantitative" or statistically oriented research. The two approaches, (quantitative and qualitative) while often complementary, ultimately have different aims. Nethnographic (Virtual/Online) ethnographic approach describes a qualitative, interpretive research methodology uses Internet-optimized ethnographic research techniques to study academic social context in online communities of practice (eLearning, Google Classroom, Moodle) etc and how they can blend pedagogy, andragogy and in real time learning management system (LMS) practices. The study helps researchers listen in on web conversations in order to learn about what actually drives higher education services and understand the inner nature of consumer behaviour. Consumer behaviour in this case refer to those consumers using our e Learning Services and software just like any other fashion industry does. While an ethnographic approach to social research is no longer purely that of cultural anthropologist, a more precise definition must be rooted in ethnography's disciplinary home of sociology that branched from anthropology. Thus, ethnography may be defined as both a qualitative research process and method and product whose aim is academic cultural interpretation. Ethnography falls in what we call a 'Mixed Research Methodology' @

to equate "popularity" with "...accepted by the scientific world" is a clear indication of a lack of knowledge and understanding o the philosophy of science

I think we are facing a new era. The digital and social media time. We all have to "surfing" the digital wave. So as a researcher my choice is take the digital media big data. I see no trouble doing this, dear Professor Campello.

Paul M.W. Hackett Hi! This is a good question. In some types of research you have no access with quantitative data and you cannot perform a survey. My research subcject are drug dealing organizations in Mexico. The ethnographic study based on long-term relation and winning mutual trust increase your chances for in-depth answers.

Best for you

p