In the past I’ve mentioned the observer’s paradox, as introduced by William Labov. The gist of it is that awareness of being observed affects people’s behavior, as explained by him in his book ‘Sociolinguistic Patterns’. This is typically the case when someone is physically present to observe you. It can, however, also occur when some feels like they are observed. The might be security cameras in rooms or devices that seem like security cameras but aren’t security cameras. In any case, what matters is that people are bound to alter their behavior in such cases.
Observer’s paradox
Why was and still is an issue in sociolinguistics? Well, it is an issue for the simple reason that research focuses on people and their behavior is marked by this paradox, as explained by Labov (209):
“[T]he aim of linguistic research in the community must be to find out how people talk when they are not being systematically observed; yet we can only obtain these data by systematic observation.”
Labov was not as pessimistic about overcoming this paradox as I am, but the thing is that he was solely interested in how people speak and not what they speak. They must be speak something, but that is beside the point. His (209-210) solution was to engage in casual conversation about some topic close to the people involved, often in groups, so that the people engage with one another and not think of the situation.
Now this applies to any research involving people. It can be an interview situation, a panel discussion or the like, but it can also be observing their behavior, whether they, for example, spent their time loitering in the coffee room or by the water cooler at work.
What about landscape researchers? Well, I would say that most landscape researchers do not have a clue who Labov is or, rather, was, as he died not that long ago, and therefore it is highly unlikely that they are familiar with the observer’s paradox. We can, however, edit Labov’s (209) original formulation of the paradox to fit the context of landscape research:
“[T]he aim of [landscape] research in the community must be to find out [what] people [say about the landscape] when they are not being systematically observed; yet we can only obtain these data by systematic observation.”
Without getting lost in the details, many landscape researchers are interested in what people have to say about this and/or that landscape. Unlike linguists, they could not care less about how they say it, inasmuch as it does not change how what they say is to be understood (e.g. sarcastic remarks).
It is also unlikely that landscape researchers have heard of the experimenter effect or the Hawthorne effect. I will explain them in this order. I will not provide an extensive list of all the effects though. I may do that in the future, if I have enough time for that.
Experimenter effect
Robert Rosenthal accounts for this paradox from the perspective of a behavioral scientist in his book ‘Experimenter Effects in Behavioral Research’. Therefore he uses certain terms that benefit from clarification.
- experimenter = researcher
- observer = researcher
- research subject = participant
It is also worth noting that this is all applicable outside research, as noted by him in the Preface (viii):
“In this sense it is not at all crucial that the experimenter happens to be the collector of scientific data. [One] might as well be a teacher interacting with [one’s] student, an employer interacting with [one’s] employee, a healer interacting with [one’s] patient, or any person interacting with another.”
Therefore, this can be simplified even further, if necessary:
- researcher = one person
- participant = another person
This can, of course, also be reversed. This is, indeed, minimally about one person interacting with another person and how that interaction affects their behavior.
To be clear, this is not to say that the roles do not matter. They do. Researchers, teachers, employers and healers are the superordinates, whereas the people they interact with, the research participants, students, employees and patients are the subordinates. That means that there are certain things that they can do or say and, conversely, cannot do or say. There are also certain expectations, what they should or shouldn’t say or do.
It’s worth keeping in mind that researchers are the ones in charge. They are the ones who are named and get credit for the research, whereas the participants are typically reduced to data points. They are also the ones who frame the interaction between them and the participants. Plus, they ask the questions and expect answers to them from the participants.
In summary, Rosenthal (vii) reckons that understanding human behavior is difficult, not only because human behavior is complex, but also because human behavior is affected by the behavior of other humans. This also applies to research, as he (vii) points out:
“Some of the complexity of … the research subject, resides not in the subject … but rather in the particular experimenter and in the interaction between subject and experimenter.”
In other words, each person who takes part in research, i.e., participant, is different other participants, in this and that way, and to this or that extent and that explains their behavior to a certain degree. However, the researcher must also take themselves into account when they attempt to explain behavior of their research participants. It’s also not just that their own presence changes the participants’ behavior. It’s not a binary, yes, it yes it does, or no, no it does not. Instead, it’s the interaction between them and the participants, or lack thereof, that the researchers must take into account.
It just is
To connect this to landscape research, many landscape researchers are interested in what people see and what they look at in the landscape. They’d also like to know what they think about the landscape and its features. However, as Peirce Lewis (11) once put it in his essay ‘Axioms for Reading the Landscape Some Guides to the American Scene’, most people don’t really think that much about their surroundings. It is not that they just look at their feet instead of the landscape or that they don’t look at certain landscape features. It’s rather that they like or dislike what they see and that’s it, as Lewis (11) points out. We could also go beyond that, to account for other senses, not only vision, but that’s beside the point here. What matters is that people are uncritical of what they see and pay attention to.
So, there’s this passive disposition that interests the landscape researcher. If the researcher approaches people, tells them about the research and asks them to tell them what they see, what they pay attention to and what they think about it all, the researcher interacts with them in a way that has an effect on them. In other words, this interaction prompts the participants to change their disposition from passive to active. While this can be productive, for everyone involved, it is counterproductive inasmuch the researcher wants to know what people say and/or do in relation to the surrounding landscape in the absence of the researcher.
To be clear, the problem has little to do with interviewing people. Instead, it has to do with research ethics concerning interviews. Contemporary research ethics define the rights of research participants, as specified by Finnish Advisory Board on Research Integrity (9-10) in ‘The ethical principles of research with human participants and ethical review in the human sciences in Finland’:
- right to voluntary participation
- right to consider participation
- right to discontinue participation
- right to withdraw from participation
- right to be informed
- right to transparency
To elaborate the first four of these rights, no one has to participate and no one shall be compelled to participate. They must be given enough to time to consider whether they want to participate. Plus, if they agree to take part, they retain the right to quit at any time, no questions asked, either temporarily or permanently.
To expand on the last two of these rights, people must also be made aware of what the research pertains to, how it is conducted, what their role is and how the data involving them is handled. They must be informed about the positive and negative mental and/or physical effects it may have on them. This must all be done in a clear and understandable manner, in writing, and any questions people have need to be answered.
It is worth emphasizing that this must all be taken care of before anyone participates in research and it must be done in a way that is transparent. It should be clear what the research is for and what kind of impact participation may have on people.
The main point here is informed consent, as noted (9) in ‘The ethical principles of research with human participants and ethical review in the human sciences in Finland’. This is all well and good in principle, but this makes critical research more or less impossible. For example, James and Nancy Duncan (404) state in their article, ‘Theory in the Field’, that:
“As critical social scientists we must decide whether our obligation to reveal the consequences of the conservation and historical preservation movements that animate the citizens of Bedford takes precedence over our commitment to informants.”
To give you a bit of context, the Duncans researched this specific town for some three decades and as they (1) point out in their 2004 book, ‘Landscapes of Privilege: The Politics of the Aesthetic in an American Suburb’, many residents of Bedford did not like them. To be more specific, they were unhappy with how they were depicted in the Duncans’ published work. To be clear, the Duncans did not write hit pieces or dupe anyone into saying anything stupid, or the like. They wrote down what people said about the landscape and then reported on that. What upset people is that they said things that made them look bad.
Oh, and I agree with the Duncans. If someone says something stupid, on record, it’s on them. It’s just that times have changed. It is no longer up to the researcher to strike any kind of balance between one’s readers and one’s informants.
I don’t know exactly what the Duncans said to people when they interviewed them, but I highly doubt that they carefully explained, in a clear and concise manner, that whatever the informants say may end up harming their reputation. I think they are part of a generation that took it for granted that taking part in an interview comes with the risk of coming across certain way, which may then be good and/or bad for you, depending on the readership.
They do, however, mention a caveat in ‘Theory in the Field’. It’s pretty funny actually, considering how accurate it is. Anyway, they (404) point out that they opted to do what they did, the way they did, because it really made no difference:
“Given that most academic writing has minimal or no social impact, either positive or negative, we have chosen to go ahead[.]”
I mean, that’s only funny because it’s true. Many of peer-reviewed articles are read like a couple of hundred of times and some those times might be by the same people clicking the same article to read it again. The chances that someone local reads it, yeah, the social impact is so minimal that it’s basically zero.
So, the only way to get this kind of information these days is to use existing accounts, like what people have said in television or newspaper interviews etc., but the problem here is that this kind of data does not grow on trees. As a researcher, you’re just hoping that journalists happen to have covered whatever you are interested in and that you have access to it. It would be better to have more views, but it is what it is.
Hawthorne effect
Much like observer’s paradox and experimenter effect, Hawthorne effect is a catch-all term. It has utility inasmuch you know what it pertains to, like what the deal is, but you do have to explain what it is all about if you intend to use it to explain something related to human behavior. It’s simply not enough to say that this or that is attributable to it. You have to be more specific.
Gustav Wickström and Tom Bendix comment on the matter in ‘The “Hawthorne effect” — what did the Original Hawthorne studies actually show?’. In summary, there was this Western Electric Company factory. It was known as the Hawthorne Works. It started its operations in the early 1900s and it was situated in Hawthorne, a small town located near Chicago, Illinois, hence the name of the factory and the effect. It was a massive factory even by today’s standards. The owners of the factory were interested in the efficiency and productive of their workforce. Therefore, a number of studies conducted in the factory.
These studies were not all the same, as noted by Wickström and Bendix (363-365). They first altered the lighting to see what effect that has on productivity and then moved on to do tests on how pauses and work time effect productivity. The thing is, however, that they couldn’t really figure out anything noteworthy. In the end, the only things seemed to make a difference were the experiments. When people took part in them, they knew they were participating in an experiment and therefore changed their behavior accordingly.
According to Wickström and Bendix (366), the results of the studies were, at best, a mixed bag. There was no clear increase in productivity that applied to everyone during the experiments, nor subsequent decrease in productivity following the experiments, once the employees returned to their usual work.
In this regard, I’d say that experimenter effect is a better term for this. I mean, it does have to do with the effect the experimenters have on the people who take part in an experiment, whereas Hawthorne effect refers to that factory that closed decades ago. It should probably be experimenter effects, in plural though. It’s actually in this regard that I prefer observer’s paradox, because it’s not about the effects, as such, but about the paradox of having to do deal with those effects that appear only in those experimental conditions.
Jim McCambridge, John Witton and Diana R. Elbourne also address this in ‘Systematic review of the Hawthorne effect: New concepts are needed to study research participation effects’. They (267) state that:
“Increases in productivity were observed among a selected group of workers who were supervised intensively by managers under the auspices of a research program.”
The gist of this is that people behave in a different way when they are being monitored. The trio (268) also point out that there is no clear cut definition for the effect:
“[It] has usually taken on the meaning of alteration in behavior as a consequence of its observation or other study.”
They (268) add that this is commonly attributed to the ways in which people acknowledge one another and, more specifically, their expectations:
“Awareness of being observed or having behavior assessed engenders beliefs about researcher expectations. Conformity and social desirability considerations then lead behavior to change in line with these expectations.”
What is particularly interesting here is that people not only tend to alter their behavior if they are observed or if they think they are being observed, but they also tend to alter it to match the expectations or, rather, the imagined expectations of others. In practice, this means that people may answer questions or provide explanations that they think the researcher expects them to say or wants them to say. The obvious problem is that the researcher does not want people to say things just because that’s what they think they should be saying or what researcher would like to hear.
Note how this can not only be conscious, but also unconscious. As a researcher, you might not be influencing the people involved willingly, knowingly, but it doesn’t mean that you aren’t influencing them.
References
- Duncan, J. S., and N. G. Duncan (2001). Theory in the Field. Geographical Review, 91 (1/2): 399–406.
- Duncan, J. S., and N. G. Duncan (2004). Landscapes of Privilege: The Politics of the Aesthetic in an American Suburb. New York, NY: Routledge.
- Finnish Advisory Board on Research Integrity (2019). The ethical principles of research with human participants and ethical review in the human sciences in Finland. Helsinki, Finland.
- Labov, W. (1972). Sociolinguistic Patterns. Philadelphia, PA: University of Pennsylvania Press.
- Lewis, P. F. (1979). Axioms for Reading the Landscape: Some Guides to the American Scene. In D. W. Meinig (Ed.), The Interpretation of Ordinary Landscapes: Geographical Essays (pp. 11–32). New York, NY: Oxford University Press.
- McCambridge, J., J. Whitton, and D. R. Elbourne (2014). Systematic review of the Hawthorne effect: New concepts are needed to study research participation effects. Journal of Clinical Epidemology, 67 (3): 267–277.
- Rosenthal, R. ([1966] 1976). Experimenter Effects in Behavioral Research (Enlarged Ed.). New York, NY: Irvington Publishers.
- Wickström, G., and T. Bendix (2000). The “Hawthorne effect” — what did the original Hawthorne studies actually show?. Scandinavian Journal of Work, Environment & Health, 26 (4): 363–367.