Fieldwork

From CS260Wiki
Jump to: navigation, search

Lecture Slides

File:Cs 260 fieldwork lecture.pdf

Extra Materials

A Global Graveyard for Dead Computers in Ghana - Pieter Hugo, New York Times

Discussant's Slides and Materials

File:Cs260-slides-FieldWork.pdf


Reading Responses

Charlie Hsu - 9/30/2010 12:31:44

Thick Description

Let me first qualify this reading response by stating that I had a relatively difficult time understanding and following it. I could grasp at some of the main ideas, but I believe I lost some of the big picture among the dense writing. I did some external research: I looked up "thick description" on Wikipedia and eventually found a summary of the paper here: http://academic.csuohio.edu/as227/spring2003/geertz.htm . The summary provided some good background to set the paper in context, as well as cleared up some of the points I had trouble understanding.

This paper is about Clifford Geertz's attempt to apply "thick description" towards an interpretive theory of culture and anthropological efforts. Thick description is a method of describing events with an attempt to provide enough context such that the absolute meaning of each event is understood: one example used in the reading shows context that differentiates many types of winks from involuntary eye twitches, both the same physical event, but with different meaning. Geertz concludes a few things about interpreting culture: the knowledge of culture grows in spurts, not far from the "ground", building off other studies, and that cultural theory is not predictive. Ultimately, he states that cultural analysis is in constant danger of losing touch with the "hard surfaces of life", political, economic realities, and that to defend against that, cultural analysis must always address these realities and offer a comprehensible, meaningful frame.

This paper's contribution to human computer interaction was not immediately clear to me. The theme of "Fieldwork" offers some clues: "thick description" is clearly better than "thin description" when performing user studies and evaluation. Design researchers should focus on not only the events and quantitative analyses of their proposed designs, but also the motives, the qualitative analysis, the "meaning" behind actions and realities that Geertz mentions in his paper. Perhaps design and human computer interaction are somewhat similar to developining cultural theory: one might say that since human computer interaction involves human likes and dislikes, social and cultural consequences, physical and technological limitations, and much more, the need to place HCI analysis into a meaningful frame is extremely important. HCI's interdisciplinary nature might require "thick description" in its research for some contributions to carry real meaning.

The paper also made me question some traits of HCI research that I hadn't though to before. Geertz states that cultural analysis is neither predictive, nor does it make giant leaps, instead opting to spurt into a "disconnected yet coherent sequence of bolder and bolder sorties. Studies do build on other studies, not in the sense that they take up where the others leave off, but in the sense that, better informed and better conceptualized, they plunge more deeply into the same things." Are these traits true for HCI research as well? Predictive is interesting to think about: certainly, many things implemented in the world today are a product of HCI research. Though many systems in HCI research fail, a great deal of concentration in HCI research goes towards applicability towards real humans and real computers, implying a focus on predictiveness. I feel that HCI research is indeed also typified by the quote above: HCI research does not always simply continue on the same path of research that a previous contribution proposed, but uses the findings to motivate new forays into the constant problem of determining how to make human computer interaction more efficient, effective, and pleasing.


Ethnographic Approach to Design

This paper describes the use of ethnography in design and HCI research. Ethnography is a research strategy that attempts to understand people and the meaning of their actions; it is done in the subject's natural setting, from a holistic point of view, descriptively, and values insights into the subject's own point of view. Ethnography was introduced primarily because of computer technology escaping the confines of lab environments, into the outside world: home, office, and recreational environments. This required designers to understand the "everyday realities of people working within those diverse settings." The paper then describes some ethnography techniques, their uses, and their applications.

I found much of this paper relevant to the user study techniques we learned in CS160. It seems many of the now seemingly essential parts to design (task analysis, contextual inquiry, interviews, prototyping with users) all fall under this umbrella category of ethnography, understanding the user and his/her needs and the realities behind the user's actions. It was interesting to see how many of those singular techniques we learnt in CS160 also had other related techniques, covered in the paper. For example, the master-apprentice contextual inquiry model is one extreme of a spectrum that has a formal, structured interview at the other extreme.

Some of the techniques I hadn't considered or studied before were interesting as well. Self-reporting techniques, I realized, might be an excellent way to asynchronously collect data on users in the case of a study over an extended period of time. Profiles and scenarios offer us the possibility to explore nuanced differences within a target population; previously, my work in CS160 had felt as if we were attempting to create a general picture of our target population, but I realize now that nuances within a target population increase the number of potential opportunities we have to design a better product.

Overall, I did not have any disagreements or concerns with the paper's claims. I feel this is a solid, overarching overview of ethnography in HCI design, clearly motivating the need user-centered design and offering a litany of techniques and advice for effective ethnography.


Airi Lampinen - 9/30/2010 21:42:27

Geertz's text "Thick Description: Toward an Interpretive Theory of Culture" is an introduction to ethnography and, more widely, to anthropological research and thinking. Geertz discusses the many complexities of reaching the object of ethnography which he defines as "a stratified hierarchy of meaningful structures" in terms of which behavior and different types of action are "produced, perceived, and interpreted, and without which they would not in fact exist" as well as the aim of anthropology, defined in the text as "the enlargement of the universe of human discourse".


Geertz points out how anthropological data really is researchers' constructions of other people's constructions of what is going on and how it, thus, always consists of layers of interpretation instead of getting at the thing itself. While I do agree that this point is especially true of ethnographic research material and while I have written enough of field notes myself and read those of others to know how hard it is to capture observations in any communicative way, I think the interpretative nature of research material is true of a wider range of data than the anthropological. Even when we are doing the "cleanest " of user studies, say strictly structured experimental studies, interpretations affect the research process in so many ways. So while Geertz makes a valid point for the field of anthropology, it is useful to bear in mind that some aspects of it are not unique to the field.

The name of the text, thick interpretation, refers to a comprehensive style of describing the phenomenon observed. Instead of listing only the "cold facts", an ethnographer is challenged to capture the context and the meanings attached to those cold facts. This is very difficult but also the very thing that justifies the effort that making ethnography requires. Geertz makes this point by citing Thoreau's words that "it is not worth it to go round the world to count the cats in Zanzibar" - we need to go deeper than that when doing fieldwork. The challenging aim is to "draw large conclusions from small, but very densely textured facts".


The second reading Blomberg, Burrell and Guest's "An Ethnographic Approach to Design" continues the introduction to ethnography but applies the knowledge to design and to how ethnography can be used to inform design processes. The authors list natural settings, a holistic as well as descriptive approach as the main principles guiding ethnographical inquiries.


While the text goes over some fairly basic starting points to ethnography and other qualitative research methods, its biggest merit for the sake of this course is the discussion on the reasons why and the ways in which such work can be of use in design and development. As tired as I am of hearing the buzzword "innovation", I especially liked the PARC slogan stating that "innovation requires an understanding of the present". While it can be difficult to make room for acquiring such knowledge in fast-paced design processes, it can make a big difference.


Luke Segars - 10/3/2010 11:40:03

Thick Description

This essay, and excerpt from The Interpretation of Cultures, provides a wordy description of the difficulty and importance of ethnographic studies. One of the primary components of ethnography that the author discusses is the idea of “thin” (objective description of the event) vs. “thick” (interpretation of an event in its relevant social context) description of events embedded within a particular culture. The author states that the production of thick descriptions is the contribution of the field of ethnography but, as shown by the author's sheep-thief story, can require substantially more study and investigation than a simple statement of events. Furthermore, understanding a macro-culture is often not enough; Jonestown is not America, and descriptions of an event occurring in a particular environment may well require an understanding of the culture in a particular town, village, or habitat.

The author focuses on two main (and surely important) points throughout his essay: good ethnography will generate “thick” descriptions, and generating such descriptions can be incredibly difficult. The author makes an effective distinction between “thick” and “thin” descriptions through the use of his sheep-thief story. Generating such a telling story clearly required an understanding of the individuals in the story as well as an understanding of what they might consider fair treatment, retribution, and so on. Perhaps the most interesting part of the essay was when the author discussed how difficult and precise arriving at such a description can be. Understanding the culture that a particular event occurs in is, for ethnography as well as anything else, a difficult but critically important accomplishment. In engineering or HCI, for example, designing a seemingly useful device for third-world countries where the inhabitants have no desire to use such a device marks a misunderstanding between cultures that automatically guarantees that a project, despite its technical or academic beauty, will fail.

He also compares the field of ethnography to the other sciences, distinguishing it by the fact that most ethnographic studies are investigating events that are “if not wholly nonexistant, very nearly so.” Essentially, achieving success as an ethnographer requires that you become a historical detective, tracing back events and culture to roots that may have, if not totally, largely eroded with time. This situation is perhaps not as relevant to modern engineering since the cultures that are being designed for are (hopefully) still in existence, but nevertheless summarizes the difficulty for ethnographers in particular.

While the author does provide a comprehensive view of the field, he fails to do so with any sort of brevity. Approximately half of his essay, for example, is dedicated to describing how difficult it can be to produce “thick” descriptions for a given cultural environment. In a sense, he point actually loses some power by being distributed over a large number of pages. The statements that the author makes are both interesting and seemingly truthful, but the number of times that he repeats similar statements weakens his message. He makes one statement (“The more I

manage to follow what the Moroccans are up to, the more logical, and the more singular, they
seem.

”) that actually summarizes a number of previous pages and could have been used to communicate the same important idea in far fewer words.


Luke Segars - 10/3/2010 13:48:07

An Ethnographic Approach to Design

This paper does an excellent job of describing how to evaluate a particular subculture and how the results of that evaluation impact design requirements. The paper starts off by defining some of the central tenants of ethnography that make it valuable both as a design and experimentation technique. The author then describes a number of experimental techniques (interviews, diaries, remote monitoring) and discusses how these techniques can be used to supply important knowledge to the design process.

The ethnographic methods described in this paper were very useful to read about. The author described each technique with the perfect amount of detail – covering important uses and pitfalls without drawing out the message unnecessarily. A number of suggestions for interviewing, for example, were things that were new to me but very logical once I heard them. The number of techniques discussed is a statement in itself to the difficulty of conducting research where human behavior is one of the most significant variables. One of the most valuable pieces of advice that I got from this essay was this: “a general guideline for determining when enough interviewing has been conducted is the point when responses to questions cease to be novel or surprise the researcher.” From an ethnographer's perspective, this is a simple means of determining when a particular psyche or cultural perspective is understood and even predictable.

In my understanding, data collection is one of the phases of the research process where human-based research can often be substantially more difficult than other types of research. Interviewing (now possible remotely) can be a hugely powerful technique for gathering information, but still has a number of downfalls that often keeps it from being useful in all situations. The paper makes the important note that what people say they will do and what people actually end up doing can be very different – this implies that directly requesting information from an individual can lead to skewed results. Therefore, the particularly time-consuming and often difficult technique of observation comes into play. “People-watching” is a common form of informal observation, although direct participation (becoming a “participant observer” by the author's term) can also be necessary to perceive a reaction to a particular event. Beeper studies were also briefly mentioned as a hybrid approach between passive and active observation.

Overall, this paper was very informative and provided a number of insights at the right level of detail. In fact, the only section that I found to fall a little short was the conclusion – considering the startling amount of content that was covered in the paper, a summarizing birds-eye view of the topics discussed would have been very effective in helping to tie them all together. Outside of this single downfall, I found this paper to be eye-opening, and I feel significantly more prepared to conduct research after reading it than I did before.


Kurtis Heimerl - 10/3/2010 16:23:30

Thick Description, Clifford Geertz, in The Interpretation of Cultures: Selected Essays.

I have no clue what this paper says, in all reality. Ethnographic studies are amazing to me. It's like a dark art, they sit in a hut somewhere and fantastical writings piercing to the heart of the issue suddenly appear. Every time I sit in a village I just eat biscuits and waste people's time. Somehow, the conclusions that arrive are irrefutable, despite the evidence just being a long string of anecdotes. You trust the researchers to be honestly evaluating the question, "what the fucking is going on here?"

What's even more amazing is how ethnographers view people like myself. We're the dark artists, making things appear with equations and coming to sweeping conclusions. My favorite example of this was in Jenna Burrel's class (iSchool). She's a top-tier ethnographer, but when faced with the cell phones in kerala paper (a deep, profound economics paper detailing adoption of cell phones in the Indian state of Kerala) her only response was, "I wish I could do something like this."

With that in mind, I love that HCI at least attempts to bridge this void. If the two fields would stop trying to publicize themselves and actually work to creating knowledge, I think a multi-paradigm analysis always wins. Evaluate the qualitative and quantitative, and the story will be clearest.

An Ethnographic Approach to Design This chapter attempted to describe ethnographic methods, with a focus on two particular questions: how does one gather data and how does one communicate finding to designers.

The latter question is not in the scope of this class. We're researchers, we are both the ethnographers and the designers. I've never found much value in "profiles" for that reason. They're contrived, you can't publish them, and it's just a proxy for our own person user model as constructed in our head.

The first question is something I've worked on for a while. The HCI class at the iSchool focuses much more strongly on methods than the CS one, and so I have a moderate background on these topics. I conducted a set of snowball focus group interviews in Uganda, and it took me a while to know exactly what that meant. I feel like these methods are important, and I wish we would have spent more time on them. Perhaps that's just a bias from my field, rather than the "if you build it they will come"-style of HCI.

I'm bummer that I'm going to be missing this lecture, I think I'd have a lot to contribute.


Krishna - 10/3/2010 16:28:35

Thick Description: Toward an Interpretive Theory of Culture

Clifford Geertz

The author provides us his views on how to interpret culture, why it is important to go beyond experimental science to construe it and that any attempts to do so should not be a search of laws but a search for meaning(pp 4). He says ethnography, the science of studying culture, is not just about fieldwork and ethnographers should desist themselves from making phenomenalistic judgements about what they observe - in other words, he says, their efforts should focus on interpreting the semantics of their observation and the contexts that could have generated it - what he terms as "Thick Description".

He prefers a more semiotic focus to the study of culture, observations can then be considered as a melange of symbols, and expects the results of an ethnographic study to be a "stratified hierarchy of meaningful structures" that describe these symbols.He says, describing culture as a set of rules marries extreme subjectivity to extreme formalism(pp 10) - with uninterpretable consequences. He explains this, as I understand, by comparing the view that a musical composition is a set of rules that effect a conceived tonality to the interpretation of the same by the composer - which would mainly be contextual. As an ethnographer trying to understand the composition, it is important to understand and cognize the context.

He says cultural analysis should be about guessing these meanings and contexts, evaluating them and drawing explanatory conclusions from them. But how we can make these latent contextual judgements when we are offered only facetious observations and information is not evident from the essays, at least to me. Furthermore, any prior understanding of culture can help make better guesses at these contexts - which, in effect, creates a cycle. There is also this facet of any such cultural study being microcosmic and questions on how one can assert the generalizations inferred from such microcosmic studies. The author concurs and says that any such study is characterized by its circumstantiality and thus any interpretations, as a result of the study, should not be liberally generalized.

The take away from an HCI perspective would be - it is hard to understand what goes behind user behavior (or) it is important to understand the context behind observed behaviors to make sense of them. Mere statistical interpretations is not deep enough and doesn't give any intuition on the context behind user's actions. According to the author, interpreting these contexts involves a lot of guesswork and evaluation of the guesses by trying to fit them with the observations - this troubles me.

An Ethnographic Approach to Design

Blomberg, Burrel and Guest

The authors provide a detailed introduction to ethnographic inquiry, its relevance and application to HCI research. The argument for ethnographic inquiry seems to be simple, the design of interactive technologies need careful study of the daily lives of the prospective users of the technologies.The authors discuss the principles of ethnography: ethnographic inquiry should be conducted under natural settings, the designers should observe user behaviors as they normally occur - induced experiments are also ok, if necessary; the inquiry should be holistic and efforts must be made to understand the context in which activities occur; the results of the inquiry should be descriptive and not tend to evaluate the observed behaviors - it should be unbiased; the focus and perspectives offered by the research should come from the community being observed - as such, the language and technicalities used to describe the findings should directly come from the community under study.

The authors mention that the research should be well planned in terms of formulating the objectives, devising a sampling strategy and selecting appropriate methods to represent the results - in most cases, research objectives are fuzzy and tend to evolve as new facts emerge and any changes in the objectives should be reflected in sampling and methodologies. The authors mention that sampling strategies fall under two categories - probability and non probability; in the case of the latter the sampling is done such that all possible variants are uniformly represented thus enabling generalization. It may not be feasible to create such samples and for most ethnographic inquiries, non probability sampling would suffice. This can be further categorized into quota, purposive, convenience and snowball - the categorizations are based upon control over the variability of the sample and prior knowledge on the availability of the sample.

In the next section, the authors discuss on how to observe the events and interview the participants. While observing, the researcher can either be an observing participant or participant observer - the difference is that in the case of the latter, the researcher actively participates in the events being observed. The authors suggest treating these two options as extremes and to move between these options based on the research needs. The authors emphasize on planning the observation in terms of "what, where and when". On interviewing, the authors give us nice "rules of thumb" for reference - from the suggested rules, it seems that good interviewing skill comes only through practice and experience. They also emphasize on taking efforts to make connections between the observations and insights from the interview - this, they say would enable us to better understand the context of the events; also, interviews can be strategically placed during an activity and can provide us contextual information, more precisely, on behaviors and events that just happened.

The authors devote the next section towards discussing where ethnographic research might be useful and how insights from the research should be represented to be applicable. In general, they say such an inquiry can be used to evaluate and critique design ideas, act as reference, enable building generative tools and enhance current working models of potential users. Their emphasis, in experience models, opportunity maps and others, seems to be on finding patterns in the results obtained from the observations and interviews, and then representing these patterns in ways such that underlying principles are well brought out and captured depending on the needs.

To summarize, there are no hard rules here, good ethnographic research comes through years of practice and a good understanding of the research problem in question. Understanding the context seems to be the key towards understanding user behavior - and this is not trivial and needs serious intellectual effort. A detailed paper that would serve as an introductory reference on how to go about this.



Siamak Faridani - 10/3/2010 16:33:14

In ‘an ethnographic approach to design’, authors start by pointing out that ethnography has been around for many years but it has been only recently that researchers in the field of HCI has started using it (some only as a qualitative method to justify their work). While ethnography has been used for market research in companies, similar methods could not be adopted for HCI development as they do not provide any insights into how people live their day to day life. In the field of anthropology, ethnography started as a method of studying smaller scale societies and specially not western societies. Among ethnographers, there is a very strong concept of participating in the society that they are studying they used to believe they should become part of that society and partake in day to day activities. Authors then point out four principles of ethnographic research. The most important section of the paper might be the ‘Ethnographic Research’ part. They highlight the most important parts of doing research a) designing the study b)Sampling, etc. To me it is not clear why authors do not want to use a proper statistical method to determine the kind of samples they need to take. They dumb down the whole study by saying anthropologists do not use statistical sampling method. But why not? I am still not sure why they do not highlight online communities as one way of observing people. They talk about harnessing the power of digital media, for example they talk about using online journals but they do not talk about observing people in the World of Warcraft. There are some references to Nardy’s work and as far as I know she is very strong in doing ethnographic research on people in these online sub cultures. Researchers can partake in online activities without influencing the society that they are studying. These kind of remote studies can be also quantified better. I also thought that the section on the ethical issues was more formulated for anthropology researchers and not HCI people. One aspect of ethnography that was missing in the article (in my opinion) was the importance of running pilot studies before the actual study. A pilot study can help the researcher design their study, determine which variables they need to record and in general how they need to run the study. Many of the shortcomings of a designed study surface in the pilot study. In thick description Geertz explains the role of the ethnographer as the one who observers, records and analyzes a society or culture. He is hoping that ethnographers interprets his or her observation by deeply understanding the underlying meanings and signs. He spends majority of the his work defining what a culture is, and most importantly what culture is not. While behaviours mean differently in different contexts. Thick description explains the behavior within its context.


Shaon Barman - 10/3/2010 17:24:16

In Thick Descriptions, the authors argue that the study of culture progresses by observing how people interact. These observations provide a base from which theories and refined and improved upon.

Observing people is a complex tasks, and the final results do not explain all the subtle nuances. The authors provide an example of a winking, which can be interpreted in many different ways depending on the time and situation. So it is essential that the ethnographer, or one who is studying the situation, be able to discern the important details and record them. Understanding the relevant details requires background knowledge about the situation observed plus keen understand of the current situation. Because culture is always changing and evolving and cannot be easily decomposed, one must base their observations with the context. Ethnography differs from most sciences in that it is unable to come up with predictions. Instead theories about culture come about from specific studies, and this work is put in a large framework of society. The hope is then to draw "large conclusions" using these smaller works. As a computer scientist, it is difficult to see the value of "thick descriptions". These descriptions can be effected by personal biases and random events which can drastically alter the conclusions. But it is a start to understand the complex relationships which occur in society.

In An Ethnographic Approach to Design, the authors introduce ethnography, outline the process of conducting ethnographic research and describe how such techniques can be used to analyze human computer interaction.

I enjoyed reading this paper more than the first since it outline ethnographic research in concrete terms: what exactly does an ethnographic researcher do? Because ethnographic research deals with humans in many of the same ways HCI does, it outline several difficulties such as sampling, access, and the affects of observing. The main ways to gain information seem to be: observations, interviews, diaries. With the digital age, observations can be made more easily with less invasiveness than before. I think one important aspect that could have given more attention was the connection between qualitative and quantitative data. Both have their place in providing information and can be used in conjunction with each other. Although the experience models and opportunity maps seem like a good tool, they seem difficult to construct for the general situation. It would have been nice if the authors discussed some of the insights needed to create such models.


Aditi Muralidaran - 10/3/2010 17:41:10

In "Thick Description" Clifford Geertz argues that the idea of "culture", which has carried away so many anthropologists with its promises to explain every facet of human behavior, is properly described as a context in which human behaviors, possessions, creations, and other artefacts can be described in a way that makes all their motivations and intentions clear. In his words, a context in which they can be thickly described - in contrast to a thin description, which only captures the physical actions, and not their symbolic meanings.

As illustrated by the Bloomberg, Burrell, and Guest paper "An Ethnographic Approahch to Design", it is therefore important to choose the correct context in which to interpret the reactions of subjects in a user study or focus group. The authors give a clear, concise description of ethnography and ethnographic methods (videotaping, interviews, diaries) that was a pleasing contrast for me, as a scientist, to the fluffiness of the first paper. They then explain how ethnographic principles can be applied to HCI by designing profiles, scenarios, and mock-ups, resulting in creations that are more engaging and relevant to their users.

These papers do not hold any great revelations to me because all my HCI education has happened here at Berkeley, where we believe in the importance of user-centered design.

As always, the objections to this great emphasis on the users side of things makes it difficult to translate requirements and preferences into the realm of engineering constraints. Another objection is that with emphasis on user-centered design, a project can never deal with a completely new kind of technology or data that has never been seen before, and for which use cases and scenarios don't really exist yet. I have been running into this problem with my research on linked data, which is so different from the model of documents and hyperlinks that I have a hard time finding users who can give me informative feedback about it.


Matthew Chan - 10/3/2010 17:52:09

===An Ethnographic Approach to Design===

When i first took cs160, the concept of ethnography was "okay" for me. Not too interesting. But when i worked with a graduate student and made several trips to do field work for 2 - 3 months, i had a much better understanding of how to approach a different culture, group, and observe in the natural work space/setting.

In this paper, everything written captures my experiences from my first field work experience ("SPRING: Speech Pronunciation Improvement via Games" - by A.Tewari, N.Goyal, M.Chan, T.Yau, J.Canny - ICTD 2010). This paper is definitely important in every field because it details the currently agreed upon process of ethnography and how to observe and participate in the master-apprentice model with ppl if designers seek to improve or create new devices and want to seek a better understanding of how users would use it or what they're currently doing now. Unlike the academic papers we've been reading so far, this was more of an instruction manual in my opinion. it did a very good job covering the principles of ethnography to the beginning and end of research, such as planning, sampling, observation, etc.

This paper related to every piece of technology today because designers must use ethnography in order to better understand their target users and the tasks they are performing now, or how they will perform a task with some new technology. They must observe users in the work space, or the home setting, or wherever needed in the most natural environment. This paper relates to all 3 research work that i've done. For SPRING, my group worked with hispanic high schoolers (some were undocumented) and we conducted demographic interviews, performed the sampling, observed the students using our games with note takers and a video camera, post-questionnaires, and so forth.

In my second work, M.Health (with Kenghao Chang) our goal is to build a model that can accurately predict a person's mental health status; in particular, we were focusing on depression and needed audio recordings of patients with confirmed cases of clinical depression. However, we could not just sit in on a patient-doctor session or have the patients read a script. We needed natural language and to gather it in a way that would not be intrusive either. The decision made was to use audio recordings during the patient-doctor sessions (and informing + giving them some form of compensation for participating in the study).

For now, i do not see any blind spots. Ethnography and the methods/techniques are still being refined when something new emerges and makes us whether something once useful is still pertinent.

Thick Description

This was by far one of the most dense readings i've ever consumed. However, it is still important to this day and age because ethnography is one of the more subjective areas of qualitative research and can be changed any time for any reasons. Clifford Geertz delves deeply into the interpretation of cultures and alludes to Cohen frequently. Just like the other reading, there are no real results, techniques, or methodologies offered. In fact, Geertz believes that ethnography isn't limited to just techniques and methodologies (unlike the previous reading that enumerates them).

This paper related to every piece of technology today because designers must use ethnography in order to better understand their target users and the tasks they are performing now, or how they will perform a task with some new technology. They must observe users in the work space, or the home setting, or wherever needed in the most natural environment. This paper relates to all 3 research work that i've done. For SPRING, my group worked with hispanic high schoolers (some were undocumented) and we conducted demographic interviews, performed the sampling, observed the students using our games with note takers and a video camera, post-questionnaires, and so forth.


Brandon Liu - 10/3/2010 18:02:23

“An ethnographic approach to design”

The article describes how HCI researchers can benefit from an ‘ethnographic’ approach. The author’s main motivation is obvious: since HCI researchers design for people outside of the lab, the techniques social scientists use to observe people should be used. The authors also compare other design methodologies, such as participatory design. These are mostly a subset of the ethnographic approach.

The overview of what ethnography is was useful. Ethnography is a holistic and descriptive investigation where the researcher attempts to gain an individual’s point of view.  Some of the controversy the authors allude to is the role of subjectivity in ethnographic investigation. Ethnographic approaches can help designers of systems prioritize features.

One example given was understanding the experience people have of dealing with financial instruments. By understanding the perception of each kind of instrument as real/play, it helped drive the development of a web application. This study was interesting in that it laid out and systematized user’s perceptions of an idea or concept.

Another valuable idea in the paper was generic profiles for users of a system. These profiles are useful because they limit the scope of concerns, and provide models to predict how people might react. One interesting part of this example was how profiles had information such as “X learned to Y...” and “X experiences ...” which are deeper properties of the profile rather than just their career or possessions.

"Thick Description: Toward an Interpretive Theory of Culture"

The article gives a valuable account of what ‘thick descriptions’ are in ethnographic studies. The example given is winking - a ‘thin description’ would be the mapping of eyelid movements to winks and their meanings, but a thick description is one in which eyelid movements are understood in a top-down manner by what is attempting to be said.

The article made clarifications as to what ethnography is and its goals. An ethnographer isn’t one who attempts to become one of the people they study - instead, their goal is to speak for them. Thus, the ultimate goal of ethnography is to enlarge human discourse.

This article is valuable to HCI in that it helps us understand when studies may be ‘ethnocentric’. One example of ‘thick descriptions’ is in product development studies. Having participated in several focus groups for a major phone manufacturer, many of the conclusions or features proposed fall into the ‘thin description’ fallacy - designers don’t understand the larger context in which a feature will be used. I once had to evaluate a feature which was a gimmick on a phone used to ‘message’ other people by waving the phone around and using persistence of vision. My guess is they came to this feature after observing that people would use phones as toys and party gimmicks, but did not understand how it would work in the larger context of having a portable and discreet device.




Thejo Kote - 10/3/2010 18:03:18

Thick description: Toward an interpretive theory of culture

In this book chapter, Geertz discusses the role of ethnography and the people practicing it. He recommends a capture of what he calls a "thick description" of an event, which includes not just the superficial occurences and behaviour of the agents involved but also the context in which the event occurred. This, he says, provides a much richer understanding of the event to an outsider.

His example of the "wink of an eye" illustrates how important context is in understanding the true meaning of an event or action. Geertz also warns about the risk of trying to formalize the study of culture and to think of it as an algorithm, and the assumption that understading that will allow an outsider to assimilate himself into the culture. He argues that it is not possible to generalize any ethnography. One always has to deal with the specific issue or culture at hand.

That was my understanding of the chapter. It was a pretty difficult read, which I guess was to be expected from a social science paper. The more I read them, the more I'm convinced that there is some undeclared competition going on about who can write the most incomprehensible prose.

An ethnographic approach to design:

Now this was a much more accessible introduction to ethnography. Blomberg, Burrell and Guest provide a history of ethnography, explain how the HCI community came to adopt it and dig deeper into specific techniques and use cases. I enjoyed the practically oriented approach of the authors. They try to answer the question - what is ethnography and why is it relevant to a designer?

They address both the criticisms of the ethnographic approach to user research and defend it with their thoughts on why it is nevertheless a very important tool in building usable systems. After providing a motivation for enthographic study, they provide an overview of the entire process, helpful tools and techniques and gotchas to be kept in mind. They also touch upon other techniques of user centered design like persona generation (they call it profiles).

Maybe it was the focus on HCI and the more relevant aspects of design that made it a better read, but, overall, I thought this was a very good introduction to ethnography.


Drew Fisher - 10/3/2010 18:23:27

Thick Description - Clifford Geertz

Geertz describes the need for understanding the entire context of a situation as a prerequisite for understanding a particular set of actions or interactions between people.

While I find Geertz's discussion to be rather lengthy for the content and written in awkward prose, he has a point. To properly understand people, you have to fully understand their context, norms, and all the things they assume, which Geertz calls "culture."

The value in this is that to understand and evaluate systems that interact with people, we have to understand what the norms are for these people. A performant system is worthless if it solves a problem that people don't have. By better understanding people and their approaches, we can produce better technology.


An Ethnographic Approach to Design - Blomberg, Burrell, and Guest

This paper says, in essence, that to study Human-Computer Interaction, we first must study and try to understand Humans. The paper details some techniques from ethnography and how they might prove applicable and useful to the field of computer science.

In particular, the paper encourages a more holistic, broad approach to looking at problems in HCI. By using case studies, we can better understand why people have the experiences they do, and how we might adapt our systems to be more approachable and useful.

Fieldwork can assist us in this endeavor, but we must do it effectively. The rest of the paper describes techniques for doing effective fieldwork, which can be summed up as observing the users in their natural environment and understanding their points of view. To get this information, we must carefully choose our participants and earn their trust. By taking an ethnographic approach to understanding our users, we can get a more complete picture of what is good or bad in a system.


Kenzan boo - 10/3/2010 18:40:32

Ethnographic Approach to Design - McGrath

In McGrath’s article, Ethnographic Approach to Design, he goes through in detail how to study users and analyze their actions from an ethnographic perspective. Ethnography aims to describe the nature of the people being studied. He gives examples of how to properly interview people, protect their rights, and have them fully understand what they are doing. The article also describes classifying user groups into either specific profiles or people with specific goals. One excellent example we studied in class was the mechanical turk problem, where the user does not have any idea what their work in being used for. Their small contribution could be aiding a goal that they are highly opposed to, like the incrimination of freedom fighters. The researchers should try to fully reveal as much as possible to the people being studied what their goals are. However this provides a huge problem when the research requires that they do not know about what the goal is. One example of this is the experiment of how much voltage one user would apply to another until the person died. The researched just told the person being studied to keep applying voltage with the feedback of someone screaming. They just told them that. If they told them that the person on the otherside was not real and was not having voltage applied they would act much differently. This shows the need to have a balance of how much information is given to the person being studied in order to get a good unbiased study. This reading was very similar to what we learned in CS160.

Thick Descriptions – Geertz

Geertz contrasts the strong difference between thick and thin descriptions. Thin descriptions are purely about the action being done, i.e. someone winked. Thick descriptions show why someone has blinked, e.g. convey an inconspicuous message to one person. The goal of this is to view human behavior as symbolic actions, and analyze what these actions truly mean. The goal of analyzing and describing human behavior as a symbolic action rather than simply a action is very important in user studies. The meaning of the action is in what the user is trying to do. E.g. When they pull up a menu, they might be intending to do something like copy a letter, or they may be looking for help because they are confused as to what they can do. Both are only revealed in a thick description by realizing their need to copy something before or realizing their frustration with the program. These thick descriptions reveal the real problem with the ui design much more than just looking at a sequence of tasks the users are doing.


Linsey Hansen - 10/3/2010 18:47:21

An Ethnographic Approach to Design

In their article, Blomberg, Burrel, and Guest describe what it means to approach a problem from a ethnographic perspective, and how to properly research it. The authors later discuss how ethnography can be used in the design process, because it allows designers to better profile their target users and the scenarios in which the users would use the product.

The article also discusses ethical issues behind ethnography, one of the main ones being the trust between the researcher and the participant. In order to maintain this trust, it requires that the researcher be straight forward with the participant regarding what the experiment is about and what kind of impact it could have on the participant. Based on experiments I have had to participate in for RPP courses, I feel that while researchers generally do disclose the purpose of an experiment with a participant, they hardly ever go into detail about negative impacts it could have on the participant (though they generally do talk about whether or not the participant's identity will need to be disclosed). Since the purpose of some research, such as that relating to mechanical turk, is about eliminating the need for professional-level people to complete some sort of task, I would imagine that using turk would jeopardize the jobs of the professionals. While I am not sure as to how the experiment's are carried out, it has led me to wonder whether or not researchers share the possibility of job loss to professionals they research or if turking is instead presented as some sort of “it will make your life easier” feature. Though some professionals can probably come to this conclusion themselves, some might not.

Thick Description

In his article title Thick Description Geertz contrasts the use of thick vs thin descriptions in evaluating the appearances and behavior of people. He then goes on to discuss the importance of thick descriptions, and how they should be used over thin ones in order to better understand the people being observed.

In summary, a thick description is one that covers the motives behind an action or behavior- such as someone wearing yellow because today is a day of religious significance, and to properly celebrate it, people are supposed to wear yellow, while a thin one would be a person wearing yellow because it was what they put on while getting dressed. If anthropologists just look at the thin descriptions for the people they are observing, they observations will be much more lacking in content vs looking into the thick descriptions. This can also be applied to user interfaces, because when observing a user, many people really do just go for the thin descriptions (ie. “the user clicked the button” vs. “the user is trying to figure out how to start drawing a picture”). By looking for the symbolism behind a user's action, researchers can obtain much more data about the usability of an interface, and better understand the shortcomings of it.


Anand Kulkarni - 10/3/2010 18:52:54

Thick Description: Toward an Interpretive Theory of Culture


The author criticizes ethnographic and cultural theory. In particular, he observes two things: first, its ability to generalize is limited by the fact that much of its analysis is necessarily contextual, and second, that cultural theory is not a predictive science.

This work appears to have limited value from an HCI perspective; ethnographic and cultural theory are valuable when doing field work, particularly in new cultural settings, but the author is describing limitations of work in cultural theory rather than its application to HCI per se. Nonetheless, the author's extended examples are instructional in understanding the limitations of field studies; when we work in new cultural contexts, we should be aware that we cannot observe simply the physical actions users take, but the cultural connotations (the "thick" structure) in which these take place. The primary contribution for HCI here is apparently this fact: cross-cultural analysis, particularly with technology, needs to look for and examine the deeper meaning and reasons behind actions rather than their physical descriptions.

The author is unnecessarily verbose and his reliance on weakly-analyzed examples to make his point obliquely -- his repeated interest in the Berbers' sheep, for instance -- make for a weak argument by the usual standards of HCI literature. The author assumes significant familiarity with the context of his work, which may be appropriate for the venue in which this article was published. It's difficult to judge the validity of the work without being more familiar with existing work in cultural theory. His reasoning that thick descriptions confound the generality of cultural explanations does seem sensible on its face.


Ethnographic Approach to Design

The authors discuss application of techniques from ethnography to HCI research.


The paper's contribution is a comprehensive survey of how the various techniques of ethnographic research can be applied within HCI, including guidance in carrying out observations, interviews and other kinds of user studies, and representing models. This is an excellent and fundamental contribution; the work of ethnographers provides a natural and well-established framework to begin investigations into how users interact with technological systems in HCI. This is doubly valuable when HCI work targets users in cultural settings that are already being studied by ethnographers (residents of developing regions, for instance). I particularly like that the authors provided what seems to be comprehensive coverage of the elements of ethnography and made specific connections to HCI. This is likely to be a paper I consult when carrying out my own user research.

The authors' use of comprehensive case studies at the end to illustrate the principles of ethnographic research in practice is outstanding and appreciated by the reader. Similarly, the early discussion on the variety of views on the nature of ethnographic research in HCI (is it simply any qualitative study?) give the authors some credibility; they don't act as direct defenders of ethnography, but simply explain its principles in several settings. The approach is pragmatic, almost guide-like, and therefore quite compelling. I would strongly endorse this paper for acceptance as it were I a reviewer; the only suggestion I might make is that the authors add some additional explanations early of the value of this kind of research to help illustrate its value to researchers from quantitative, non-HCI backgrounds.


Pablo Paredes - 10/3/2010 18:53:11

Summary for Geertz, C. - The Interpretation of Cultures: Selected Essays - Chapter I: Thick Description: Toward an Interpretive Theory of Culture.

The first element of the essay is to describe what Interpretive means, which is describes as the search of meaning. The notion of finding meaning beyond the description of a behavior, but define this meaning in its context and its relation to society.

The reading of this meaning is done through the construction of descriptions that begin from observations, but observations that are made from an "actor's point of view", meaning, being immersed, but mimicking with the environment. In any case, behavior, and its description of key, as culture depends of the flow of behavior (social action). However, the description of culture seems to be unbounded, unending, coherent, yet not procedural and requires a mix of qualitative analysis that provides descriptors of culture, but not a deterministic recipe.

Ethnography is described as the means to accomplish this interpretive analysis of culture, because ethnographic descriptions have three main characteristics needed for these type of analysis: it is interpretive, it interprets social discourse and provides a basis for some generalization. Among the different settings where ethnography is developed some of the most relevant are: the microcosmos (small societies define large societies where they are embedded) and the natural experiment (defining a setting to analyze its characteristics).

The author clearly closes stating that cultural theory has two components: 1) it is not self-governing, as the variables cannot be modified and it cannot be abstract, i.e. generalization is a compendium of descriptions, and 2) that it is not predictive - it is rather recursive and never-ending.

From my perspective, although it seems then as if cultural research is more a narrative literary field, the advantages of ethnography is that it helps understand humans in their context, helps define common vocabulary and it provides the view of the society as a whole, rather than an personal view (like in the case of a philosopher). It also provides the basic understanding upon which other fields can define ways to improve society.


Summary for Blomberg, J., Burrell, M. and Guest, G. - An Ethnographic Approach to Design.

The most relevant points from this paper is to note the need from technology driven sciences and industries to go back to basics and to make sure there is a clear understanding of the reality of humans before starting to generate theories and deployments on top of conjectures many times biased by the group of people involved in the scientific or product development process.

The paper describes the tools based on ethnography upon which clear observations of society can be made, such as observation, videotaping, interviewing, self-reports, visual stories. Also describes the new challenges faced to understand human culture in the new era of digital domain, where clear boundaries of interaction and behavior have changed dramatically. It catches my attention though that the initial need for immersion needed in the past, which helped account for all the details of the context, is in many cases now being replaced by reading log files. Seems much to me as if ethnography is finding ways to bring society to the lab, to probe social dynamics and generate much more "controllable" experiments (based on highly random variables).

To close the cycle thought, and in the light of HCI, it is fundamental that ethnography does not remain in the usual rather non-assertive and innocuously descriptive plane of study, but that evolves towards a more (at least) assertive translation of the findings expressed in certain tools that can lead to further innovation. Tools such as experience models, opportunity maps, profiles, scenarios, prototypes, are required to make ethnography an active part of the HCI design flow.

To be very honest, it seems to me like computer science, a science created to enhance and support the needs for humans to compute all kinds of information related to every other scientific field, and a science created from humans for humans, finally took a humble step back to remember what its inception meant and what it exists for, i.e. humans. Seems to me almost paradoxical that ethnography is seen in computer science as a modern revolution. It should have always been the origin of any study, even the most complex theoretical analysis of computation should embed at some point a link with a human. However, maybe it was the speed of development, which led to some self-induced conjectures that led to a path with no-end, because the basics of human understanding were not present. I expect that HCI continues to provide insights to all of its forming disciplines, however I can also foresee the possibility for scientists to believe that technology is the new paradigm of evolution. In any case, both cases benefit highly from understanding the underlying human phenomena, and ethnography provides the basis for this task.


Bryan Trinh - 10/3/2010 18:55:38

Ethnographic Approach to Design

Jeanette Blomberg, Mark Burrell, and Greg Guest present an overview of the ethnographic approach to design in the context of human computer interaction. Starting from a philosophical stand point, they describe the relevance of the ethnograhpic approach towards designing for humans. This was followed by a description of different methods and a some explicit examples of the methods in practice.

Today ethnographic methods are one of the core practices of any design firm wishing to create products for human beings. As such, there is a dizzying amount of verbage harking the benefits of anthropological methods in design. They all come with their own flavors of tools and methods, grids and frameworks, yet they miss the most important thing--frame of mind. In my own experiences, interacting with another human being in the hopes of understanding their perspectives, is not as easy as just picking the right method, it is about preparing yourself to accept the information. Getting into the right frame of mind is not an easy task though. It is the type of tacit knowledge that can only be achieved with practice and the associated human feedback.

Without the right frame of mind, an interview might provide little insight. For instance, the ethnographer needs to be careful not to incorrectly trace the genesis of certain actions or statements. It is easy to judge the statements or actions of the interviewee through ones own codes of conduct. This error can completely miss guide the production of any other analysis that might follow. Even among individuals with seemingly identical backgrounds and profiles, logic systems that guide actions are unique. One must not forget this in the process of understanding the data.

Besides this point, the authors did a good job of reviewing the importance of ethnographic methods in the design process.

The Interpretation of Cultures : Selected Essays Clifford Geertz touches upon those areas of ethnogrpahy that have always confounded me--that is, how can we principly understand another human being with just a few chunks of the totallity of their life. Well, we cannot. I read with great pleasure that someone has written about the truth of the matter, yet did so without destroying the scientific significance of the practice. In Geertz view, we do not pretend to be omniscient outsiders, implicating the actions of others in the context of ourselves. Instead we guess, try and contend with the fact that we can only, at the very most, asymptotically reach understanding.

Admitting to this fact does not, however, thrwart the pragmatism of the practice--it extends it. The mapping from observations to understanding is then purposefully left open ended. The information is still put to use, but continues to evolve towards a higher level of understanding. This prevents one from bucketing individuals into preconcieved personas, instead an attempt is made to place the invidual in a unique intelligible frame. The goal is to "not generalize across cases but to generalize within them."

This idea is extended further in the actions that are prescribed after analysis. The author calls the act a diagnosis rather than a prediction. Specifications can be created, yet nothing is concluded and the sureness of the matter is thrown out.


Richard Shin - 10/3/2010 18:57:13

Thick Description

This essay discusses ethnography and its applications in understanding culture. The author begins by stating that anthropology is essentially ethnography (a method that is, unfortunately, never fully described within the essay). By using an example of the distinction between winking and twitching the eye, the author then discusses how culture is symbolic, with meanings attached to seemingly insignificant acts by public agreement. The author states that despite best efforts, ethnography results only in a third-order interpretation, and therefore analysis derived from it would be meaningless without attachment to the actual events that resulted in it.

Overall, the essay seemed to present a warning to the author's fellow anthropologists about what the pitfalls of ethnographic methods, and how its results must be accompanied with context, so that meaning is not lost. Also, the author emphasizes the role of symbolism in culture, in that the meaning of human activity cannot be understood simply by observing it without context, unaware of the meanings. Understanding people is certainly very important to human-computer interaction, so being aware of how pitfalls involved in understanding them seems important as well. When conducting studies about user studies, it would be helpful to keep these ideas in mind.

I have to admit that I had a lot of trouble understanding what this essay was trying to say. The sentences throughout the paper consistently use a very large number of parenthetical statements, which made them perhaps specific and informative, but nevertheless quite difficult to understand. I feel that I would have gotten a lot more out of this essay if it had been written in a clearer style.

An Ethnographic Approach to Design

This paper presents the field of 'ethnographic design', where ethnographic principles are used to guide the process of design. By using ethnographic techniques, researchers attempt to gain an understanding of the people that will use some computer system, and the relationship between them and the system, from users' perspective. The paper presents the principles and methods of ethnography as they apply to interaction design, and then discusses how results of ethnographic research can be analyzed and applied to user interfaces.

The paper argues that ethnography is helpful for the field of HCI and design along several axes. First, people who design and develop computer systems are increasingly different from the users of the systems, which makes intuition and experience unsuitable guides. However, self-reported from users can be unreliable due to bias and limited memory, requiring a more sophisticated method of studying the users. Additionally, designing networked computer systems that are used by many people at once require a deeper understanding of how people collaborate and interact with each other generally. The paper then describes how to gain an accurate, holistic understanding about the users through observation, interviewing, and other forms of data collection. The methods involved seemed quite similar to how I thought user studies are conducted, namely in the emphasis on observation and working with a small number of people, yet different in how ethnographic methods emphasize natural settings while most user research I know of takes place in special labs.

While the paper seemed like a helpful guide on how to apply ethnographic methods to interaction design, I wish that it had further expounded upon the relationship between ethnography and quantitative approaches to design. It seemed to me that, despite the attempts to remove bias as much as possible from the descriptions obtained, the way that people are selected as subjects of study, or how topics of importance are selected, could be subject to much bias from the researchers. In other words, while the data about the selected users might be very accurate, it is hard to tell if the relevant or necessary data has been collected in the first place.


Matthew Can - 10/3/2010 18:59:51

Thick Description

In this essay, Geertz proposes a theory of culture in which the analysis of culture is fundamentally about interpreting the meaning of social events. In Geertz’s framework, the goal is to understand the thick description, the meaning and importance of social discourse. An apparent problem with this theory is that because the analyses are interpretations, anthropology loses its objectivity. Geertz argues that the purpose of an ethnographic account is not to collect information in faraway places and bring it back home but instead to explain what goes on in faraway places. This raises problems for verification, but Geertz states that the better account is the one that is better able to sort out the meaning of events. On the topic of behavior, he argues that it is important to observe because cultural forms draw their meaning from the ways they interact with each other. Geertz characterizes his theory of culture as an effort that will refine the debate in anthropology by providing more answers as opposed to one that will result in a perfect consensus around our deepest questions.

This essay’s relevance to HCI is that Geertz’s theory can inform and guide the way in which we gather information about our users and conduct user studies. Certainly one thing researchers have done well in HCI is to capture user behavior and model it. Examples of this from previous readings include the Model Human Processor and the Keystroke Level Model. This reading stands in contrast to such models because they fail to explain the meaning of the social discourse that takes place between the user and the system.

One question this raises is whether such interpretive meaning is of any benefit in HCI. I think it provides a huge benefit in directing the development of new systems. Let me illustrate this with a hypothetical example. Suppose that every time I log in to Facebook, I go to my best friend’s profile and poke him. A researcher examining only my behavior might design a new system that automatically sends a poke to my friend every time I log in. This will eliminate any errors and reduce the time it takes to execute to the task (essentially to zero), both of which are generally favorable. But, this completely undermines the purpose of poking my friend, an action that I want to engage as a form of digital social interaction. By understanding this, the researcher can develop a better system with a button on the home page that allows me to poke my best friend.


An Ethnographic Approach to Design

This reading describes the ethnographic approach to design in HCI and how that approach benefits the design process. The authors explain why ethnography is relevant to the design of human-computer systems. They discuss general principles of ethnography and the methods by which ethnography is conducted. More importantly for HCI, the paper describes how the insights gained by ethnography can be applied to design and it provides two case studies of the use of ethnography in the design process.

What I found most interesting in this reading was the explanation of how ethnography can inform the design process. Perhaps the greatest benefit is that it provides developers with a rich working model of the user. More specifically, this understanding of the user can be used to build profiles, representations of the characteristics and experiences of a prototypic user (or even a particular individual). Profiles remind developers for whom they are designing a system, and they can be used to understand how a certain kind of user might interact with the system. Profiles can even be viewed as a cost effective (though less informative) alternative to user testing.


Thomas Schluchter - 10/3/2010 18:59:57

    • An ethnographic approach to design

The article outlines the role of ethnographic research methods for design processes and details some of the applications in an industry context.

I think the emphasis on ethnographic methods is extremely important as we design increasingly complex systems that are deployed in environments already saturated with technology. The greatest challenge that designers and developers face in this environment are not technological in nature, they are problems of understanding; understanding what humans do and why they do it. Only if we understand this we have a chance of building things that fit into their world.

It is good to see that the article closes with a reminder to embrace the complexity ethnographic inquiry. The term "ethnography" these days is thrown around wildly for anything and everything that involves 'talking to users'. Ethnography is more demanding than that, it is more rigorous (see Geertz), and it will only produce valid results when undertaken with a solid methodological grounding. Increasingly, companies hire dedicated researchers with methods training to function as a resource to their design and development organizations. This is a step in the right direction, but it leads to the question what the impact is on stakeholders within these organizations -- there is (and will always be?) a bias towards quantitative methods for 'proving' things that businesses base decisions on. The question is how business leaders can be influenced to allow time and money to flow into a deeper understanding of their target customers.

    • Thick description

Geertz uses the example of an ethnographer in Morocco to illustrate how the rich description of a culture can lead to a better, if incomplete, understanding of how human life is embedded in symbolic systems that can be interpreted.

To me, this reading shows that what we loosely call "ethnography" in an industry context is only a shadow of what


Arpad Kovacs - 10/3/2010 19:02:05

The Ethnnographic Approach to Design chapter describes procedures for applying ethnography techniques, developed by anthropologists for studying the everyday behavior and rituals of indigenous peoples, to the design disciples. The main motivation for these techniques is that bringing in outside perspectives via surveys, focus groups, or interviews miss details of tasks that the research subjects may neglect to mention or deem unimportant. Therefore, it is essential for the designer to observe firsthand how everyday tasks are performed in the context of their natural surroundings from a holistic perspective. The ethnography should be descriptive, rather than perscriptive, and the researcher should strive to put themselves in the subject's shoes to gain an accurate "insider's view of the situation." The paper goes into great detail on how to set objectives, plan a research project, find participants and gain access to field sites through sampling, strategies for observation and on-site or remote data collection, as well as performing interviews or gathering self-reported data through diaries or storytelling to gain a more holistic view of the situation.

The second half of the paper deals with how to apply the observations and insights that ethnography yields to improving design. The main contributions of ethnography appear to be providing designers with a more accurate working model of their product's users, as well as providing a sense of direction and guidance for which designs would be worthwhile. In particular, organizing observations into a structured formats such as experience models or opportunity maps help to discover trends and patterns in the data, build user profiles and plausible scenarios, as well as evaluate candidate designs.

The main contribution of this paper is providing a howto guide for performing a scientific ethnographic survey. I was a bit disappointed in that the paper mentions that gaining trust and maintaining a relationship based on reciprocity play a central role in running an effective ethnography, however there is no disussion on how to establish the initial relationship and build rapport with subjects (other than the advice to use a recruiting agency or some other intermediary and ask for advice). Another potential problem with the ethnographic approach is that it could be quite time-consuming, and requires a long-term approach and could take months, if not years. Therefore I wonder how applicable it is to the modern product design development cycle, which must often go from idea to launch in only a few weeks or months. In this case, it seems that hiring a market research firm (with expertise in the field of interest) to perform several thousand ethnographies and summarize the results would be a more efficient use of the designer's time rather than single-handedly observing a half-dozen on-site use cases. Otherwise, the chapter was very comprehensive and detailed, and convinced me that ethnography for design is a worthwhile endeavor.


The Interpretation of Culture: Thick Description reading was quite difficult, since the author kept digressing off into wild tangents, resulting in a low signal-to-noise ratio. As I understood it, throughout the paper, the author attempts to define culture as a web of influences, basically a public system of symbolism and meaning. However, he is unable to define it precisely, therefore he gives numerous examples and anecdotes to convey the difficulty of his task. The author also discusses the role of the anthropologist: to inscribe social discourse, in order to create a written record, and thus expand the boundaries of human perception and knowledge.

The main contribution of the paper to HCI is the snippet of the story of robbery and recompensation in the French colony, what the author believes is an excellent example thick description, and his analysis of why it is effective. The passage is very descriptive and accurately conveys the mood and context of the period, which allows the reader to put themselves in the perspective of its characters. The author also has a long debate about why ethnography is interpretive; ultimately he concludes that interpretation is interwoven into any written account, and it is the role of the anthropologist to extract insight and provide his/her own interpretation of the ordinary, everyday social discourse in order to maintain a record for future analysis.

With the development of accurate recording media such as photographs, tape recorders, and video cameras, it seems that the role of the anthropologist as a mere recorder of social discourse has been toned down, and instead the interpreter role has come to the forefront. Likewise, I think that modern research in the field of HCI has the capability to gather terabytes of data through ethnographic field surveys and usability studies, so the real challenge for the designer is not recording this data, but rather making sense of it and discovering interesting patterns and observations.


Aaron Hong - 10/3/2010 19:48:44

In "An Ethnographic Approach to Design" by Blomberg et al. they talks about "ethnography" which is defined as "the study of the characteristics of various peoples and the differences and relationships between them." The relevant definition to HCI is "the methods and techniques that enable the development of a _descriptive_ and _holistic_ view of activities as they occur in their _everyday setting_ form the _point of view of study participants_." This is useful for making informed design specifications, although the results of ethnographic study might not directly translate.

No doubt, I agree that they are extremely useful because we can't just imagine what people need and do on a day-to-day basis--we need to be in the thick of it. However, the methodology as outlined in the article is tricky: how do you insert yourself in there without changing what the participates would do; also how do you extract information that the participates might not even know about themselves. This is extremely relevant to research when we deal with human factors.

In the first chapter of "The Interpretation of Cultures: Selected Essays" by Geertz, he talks about an "Thick Description" or ethnography. He goes through some people's description of culture and ways of understanding it, but ultimately concluding that it is an difficult task--people looking for "turtles on turtles." One solution is to look at the surface before we look at "all-too-deep-lying turtles." "To look at the symbolic dimensions of social action--art, religion, ideology, science, law, morality, common sense--is not to turn away from the existential dilemmas of life for some empyrean realm of deemotionalized forms; it is to plunge into the midst of them." I agree that this is the case, that we must not dig too deep into this fictional world all the while thinking we're doing anthropology.