Paper 2: Free Speech / Advertising


Choose one of the following. Note that the first one is a conventional paper and the second is more of a research project / experiment.

Speech Policy

You and two friends, Alice and Bob, are starting a new website in which user comments figure prominently. Users comment on various products and also on the reviews and comments provided by other users. You anticipate that the majority of users will use their real names on the site, though pseudonyms are available.

Right now you're trying to work out a policy for dealing with defamatory and otherwise negative comments. Alice has argued

We don't need to do anything; section 230 of the CDA clearly means we have no liability for what our users post, and thus no obligation to remove libelous content. Many other sites, such as youtube.com and aol.com, don't seem to remove such content. Besides, how do we determine if an allegedly defamatory post is in fact true? Would we have to post the other side's position?

Bob is not so sure; he's argued

We simply should not let libelous content remain, in particular if it's just plain insulting. At the very least, it will annoy our other users. We need to have a clear standard of behavior; this is about "takedown" and not aboutarguing the points of libel.

Your job is to propose a policy and then argue in support of it. If you leave something out of the policy, such as a way for users to complain that they have been unfairly depicted, be sure you explain why you don't think the feature is necessary. You should also make clear whether features of your policy are their to address legal risk to the site or are there to make users feel more comfortable.

You can take a legalistic approach, an ethical approach, or a combination. As I said on the midterm, when making ethical arguments in a business context it is sometimes helpful to recognize that ethical behavior can be closely tied to a business's own long-term self-interest.

Length: 3-5 pages

Advertising

We've discussed the idea that advertisers track you on the internet. Perform some experiments to verify this. Use a browser with a clean slate to visit some sites and then see to what extent the advertisements on other sites reflect things you searched for or expressed interest in, or other ads you clicked on. Seeing ads "follow you around" should be relatively easy to document. Do you see other things? Do ads for competing products appear?

Be sure to use a "clean-slate" browser: either create a new computer account and use that, or delete all cookies. A new account is simpler.

The Wall Street Journal did a series of articles about web tracking; one can be found at http://online.wsj.com/article/SB10001424052748703940904575395073512989404.html. See the class notes from Week 6 for some other links.

Some things you can search for to get the advertisers include automobiles, vacation travel, sports equipment, electronics/computers, upscale clothing, or holiday shopping items. Be aware that lots of sites have car or travel ads for everyone; not every ad is targeted. To search, start with google and go to a few main sites (dodge.com, orbitz.com, etc). If you want to be daring, try searching for some medical conditions. I recommend trying at least two unrelated product categories.

Some advertising-supported websites to look at might include dictionary.com, slate.com, cnn.com, msn.com, aol.com, chicagotribune.com, thedailybeast.com, yahoo.com and of course vidtomp3.com. I'm sure you can find many others. (Facebook has its own advertising operation, and so is probably not applicable here.)

When you write up your results, spell out what product you searched for and where you searched for it, what you expected in terms of advertiser response (the "hypothesis"), and what you actually found. While you don't have to do any kind of statistical analysis, you should make an effort to repeat the experiment a few times, with cookie deletion in between if appropriate, to avoid being influenced by one-time quirks.

It is not necessary to identify the sources of the ads (eg google.com, doubleclick.net, lotame.com), though mouseover often does identify the source.

You will be graded on the nature and detail of the hypotheses you test, the nature of your experiments, and the clarity with which you write up your results.

Length: probably the same as the above, not counting pasted-in results (which you might not choose to do)