What they reveal is interesting but what they conceal is vital.

Neither the Slate article nor the MediaCurves article ever make it clear how many doctors responded. Slate indicates that 560 doctors were asked to watch the video, but never says how many responded. The bar charts are all in percentages for your confusion...I mean convenience. The MediaCurves article is more vague in saying a study of 560 doctors. Does that mean they asked 560 or that 560 doctors agreed to watch or that 560 doctors responded?

Doctors were invited to participate by direct and email. Doctors are notorious for having time management problems with a few exceptions like dermatologists. If the doctors weren't busy, I wonder why?

Given the level of uncertainty of the above, one logically now must ask the question, what kind of doctors were these? Were they dentists, medical examiners, pediatricians or what? HCD presented an edited version of an interview with Kevorkian, not balanced presentation. Response was determined during the video as to whether they agreed or disagreed with what was being said but it is unclear if this is a continuous rating, a periodic rating or a designed irregular interval. Let's create a little scenario in which you the readers can participate by rating the following possible material as to whether you agree or disagree:

Today we have a special guest. This is Dr. Kevorkian.

<Make your rating now>

Dr. K, you have been in prison for 8 years.

<Make your rating now>

Do you find dying tragic, Dr. K?

<Make your rating now>

Wouldn't you like to find a way to prevent people's suffering?

<Make your rating now>

You know how interviews go. This was an absurdity. Statistically was there a difference between before and after? It is impossible to tell without the raw data. Even then without knowing how the questions were worded, it is impossible to tell what biases may have been present beyond the fact that the interview was not a debate. For example, the following question is obvioiusly biased:

Would you rather your patients die a slow agonizing death or would it be better for them to have a sweet immediate release?

Just two last comments.

Beware of studies concealing data and methods; they tend to conceal something vital.

Was this really an exercise in the effectiveness of advertising?

But I have to say, this was the most fascinating of these MediaCurves segments that I've seen yet. Most of the past ones have shown (surprise!) that Democrats and Republicans largely support their own platform and oppose one another's platforms, with Independents somewhere in between.

But this one - it actually picked an unusual fault-line (religious faith) which signified a notable discrepancy on a theoretically unrelated issue (medical ethics).

I wouldn't argue that it says anything. But it certainly suggests a wide range of possible conclusions.

Even if the methodology's flawed, I'd argue that the tragedy is more that this video snippet will likely stop at the threshold of an intriguing observation, rather than follow it up and investigate how much significance there is to its results.