At the end of the day, you have to decide where your ethical boundaries are personally, and ask a question that will get you a response that tests that boundary for this company.
Hypothetical questions are very useful because you can cater them to whatever your boundary is. If you are concerned about how the company handles differences between the client's request and the result of the company's research, you could ask:
Hypothetical question. A lucrative client is getting sued and wants us to defend them by proving X (which they insist is true). We perform research and it comes to conclusion Y. We check again with further research and come to conclusion Y again. How will our company bridge the gap between the client's conclusions and our own conclusions?
This is a very macro approach that discusses how the firm handles customer relations.
If your concern is less how the firm handles customers, and more how the firm handles researchers, you could try the above situation with a modified question:
What are my responsibilities as a researcher if my conclusion does not match that of the client?
Taking an Ethical Stance
I like hypothetical questions because they are open-ended and don't suggest which answer you want. However, they aren't for everyone.
If you are deadset on a certain ethical standard for the company or your work, you can address that specifically with a direct question during the interview.
If the conclusions of my research don't match what the client was expecting I don't feel comfortable adjusting the findings to match the client's request. Will that be a problem?
If that is too aggressive for you, you can tone it down slightly.
I believe strongly in separating the business-side of requests from the results of research. As a researcher, will I be expected to reach a conclusion independently of the customer request?
At the end of the day, ethics are relative to how much a person can see. Take a simple case:
Researcher Angela comes to conclusion X.
Analyst Bert receives conclusions X, Y, and Z from the researchers, and creates analysis A.
Salesman Chris receives analysis A, and makes recommendations B and C to the customer.
If Angela reads Bert's analysis, she may find it unethical (especially if it puts a higher emphasis on Y and Z than her research in A). For Bert, he may find it perfectly ethical because he feels his analysis was appropriate in the situation.
On the other hand, Bert may feel that his Analysis A was misrepresented to the customer with recommendations B, and C. Chris on the other hand may think it was no problem, because given the customer relationship, he felt non-false B and C would go over better than going exactly with analysis A.
You have to decide how many "stages" of ethics are required to make you feel comfortable in the office. If the entire process needs to be ethical for you to be content, I strongly advise against picking any job which works for potentially guilty corporations.