I went out to search for such studies, and you are right, there are few where people actually identify the norms - most often they measure a norm that is already defined (what percentage of people in a group drinks, smokes, litters or speeds). Here are some that I found:

Cialdini (the guy who came up with the descriptive/injunctive distinction) was doing it experimentally - he would provide his participants with an opportunity to litter and then record their behavior http://media.cbsm.com/uploads/1/AFocusTheoryofNormativeConduct.pdf - "From a hidden vantage point, an experimenter noted ... whether the driver littered the bill. Littering was denned as depositing the handbill in the environment outside of the vehicle." - so there you have a measure of what is the normative behavior. 

In economic psychology, the social norm is often operationalized as a solution to an economic game (http://www.kabbalah.info/forums/arosa/engmaterials/Altruism/Ernst_Fehr__Urs_Fischbacher__Social_Norms_and_Human_Cooperation.pdf for a review)

To identify norms for cooperation, Heide and John (1992) measured three aspects of such norms: flexibility, solidarity, and negotiation, with a survey http://www.jstor.org/stable/1252040

These guys talk about identifying norms in organizations, using quantitative and qualitative methods, and refer to some literature on identifying norms in organizations: http://www.emeraldinsight.com/doi/pdfplus/10.1108/EUM0000000001202 . Overall, there seems to be a separate tradition in organizational psychology. 

These guys were trying to identify the norm of prior judicial experience for being elected to the US Supreme Court: http://www.jstor.org/stable/pdf/3481406.pdf?acceptTC=true ; They argued that the norm got introduced under Eisenhower and did the following to test their idea: 1) checked the percentage of sitting judges in the short-lists for USSC; 2) checked the difference between those elected before Eisenhower and after; and 3) checked if the number of sitting judges who get elected increases over time. 

This piece deals specifically with norm identification and measurement, and makes some intuitive sense: http://www.unicef.org/protection/files/4_09_30_Whole_What_are_Social_Norms.pdf

There are other examples, but generalizing from the above, people seem to get creative. What one want's to know when trying to identify a norm is what is the group (social norms still only exist in groups), and what are the beliefs of the members on how prevalent (descriptive) and desirable (injunctive) are behavioral options in a specific situation.

This can be done with behavioral observation, with interviews, games, or surveys. And it also seems to depend a lot on the context of the study. 

I found the question a little hard to understand. 

Your examples seem to be asking what is the normative behavior. That is not measuring, that is identifying normative behaviors. You could do that with these questions, with interviews, field observations, and so on, but that will not provide you with a good measure for quantitative analysis. 

In asking what is the appropriate behavior in a context, you also might be confusing descriptive ("what do people actually do") and injunctive ("what people should do") or personal ("what should I do") norms. Descriptive norms by definition do not deal with appropriate behavior - they deal with actual behavior, and they almost certainly need a reference group. You could ask your respondents "What are people in general doing in these contexts", but they might reply with a "depends on what people".

Injunctive norms are often measured as perceived peer/parents/stranger's approval for a specific behavior (e.g. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2605291/), or by asking people to imagine people's reaction about norm transgression. Injunctive norms are also part of Ajzen's theory of planned behavior, and there are hundreds of studies on that - you can find many different measures of injunctive norms there.

Personal norms can be measured with statements like "it is okay for someone like me to do XYZ" (http://jod.sagepub.com/content/36/1/147.full.pdf)

Hi Vladimir,

thank you very much for your extensive answer! You are right, the wording of my question was a bit unclear. I am not referring to injunctive norms, even though the word "appropriate" might have suggested that. I want to ask people which behavior is normal in a certain context as in how would most people behave.

I find your notion of "identifying normative behaviors" very interesting. That might actually be quite close to what I had in mind. I did not find any paper that focuses on "identifying descriptive norms". Do you, by any chance now such studies?

Thanks,

Arne

I went out to search for such studies, and you are right, there are few where people actually identify the norms - most often they measure a norm that is already defined (what percentage of people in a group drinks, smokes, litters or speeds). Here are some that I found:

Cialdini (the guy who came up with the descriptive/injunctive distinction) was doing it experimentally - he would provide his participants with an opportunity to litter and then record their behavior http://media.cbsm.com/uploads/1/AFocusTheoryofNormativeConduct.pdf - "From a hidden vantage point, an experimenter noted ... whether the driver littered the bill. Littering was denned as depositing the handbill in the environment outside of the vehicle." - so there you have a measure of what is the normative behavior. 

In economic psychology, the social norm is often operationalized as a solution to an economic game (http://www.kabbalah.info/forums/arosa/engmaterials/Altruism/Ernst_Fehr__Urs_Fischbacher__Social_Norms_and_Human_Cooperation.pdf for a review)

To identify norms for cooperation, Heide and John (1992) measured three aspects of such norms: flexibility, solidarity, and negotiation, with a survey http://www.jstor.org/stable/1252040

These guys talk about identifying norms in organizations, using quantitative and qualitative methods, and refer to some literature on identifying norms in organizations: http://www.emeraldinsight.com/doi/pdfplus/10.1108/EUM0000000001202 . Overall, there seems to be a separate tradition in organizational psychology. 

These guys were trying to identify the norm of prior judicial experience for being elected to the US Supreme Court: http://www.jstor.org/stable/pdf/3481406.pdf?acceptTC=true ; They argued that the norm got introduced under Eisenhower and did the following to test their idea: 1) checked the percentage of sitting judges in the short-lists for USSC; 2) checked the difference between those elected before Eisenhower and after; and 3) checked if the number of sitting judges who get elected increases over time. 

This piece deals specifically with norm identification and measurement, and makes some intuitive sense: http://www.unicef.org/protection/files/4_09_30_Whole_What_are_Social_Norms.pdf

There are other examples, but generalizing from the above, people seem to get creative. What one want's to know when trying to identify a norm is what is the group (social norms still only exist in groups), and what are the beliefs of the members on how prevalent (descriptive) and desirable (injunctive) are behavioral options in a specific situation.

This can be done with behavioral observation, with interviews, games, or surveys. And it also seems to depend a lot on the context of the study. 

Wow, thank you so much, Vladimir. The Heide and John (1992) paper comes very close to what I had in mind. And the UNICEF Report yields some great ideas, too. Thank you for taking the time to list all those articles, I greatly appreciate your help.

Cheers,

Arne

Glad some of that was of use!

I think you may be asking about their PERCEPTION of descriptive norms rather than identifying the ACTUAL descriptive norm. Of course, their perception is what is what influences their behavior, However their perception and the actual are usually not congruent.

We often ask a sample to identify their perception of  a descriptive social norm in one of two ways:

 1) "Does a TYPICAL student at your school wear a seat belt?" or "How many drinks does a TYPICAL student drink at a football game?"  and 2) "What percentage of students at your school wear a seatbelt?" or  "What percent of students drink 5 or more drinks at a football game?"

We also ask, of the same respondent,  whether they wear a seatbelt  or how many drinks they drink at a football game.

We can then construct a perceived norm and compare it to an actual descriptive norm. Almost always the preception of descriptive norms is not accurate. Usually pro social and health positive behaviors are perceived as being less normative than they are.