# 2017 Self-evaluation

It's been a couple of years since we did an automated self-evaluation and it seemed like a good time to look critically at this site again. (I hadn't thought of it until I gathered the questions, but the recent eclipse must have brought in more visitors in the last few months.) We aren't doing automated evaluations anymore, but we still think introspection is a good idea.

Below, I've listed 5 questions chosen at random from the last quarter that have received at least some anonymous feedback. I also included two scores*:

1. How helpful the page seemed to regular users.
2. How helpful the page seemed to anonymous visitors.

The range is from 0 to 100% where bigger numbers mean more helpful than small numbers. The final number is the number of total views the question had received at the time of sampling.

If you'd like to participate, please copy the question list into an answer for evaluation. Exactly how you evaluate the questions and answers is up to you, but it would be useful to:

• Find better answers on the internet (if they exist).

Exactly what "better" means is up to you, but the criteria might include:

• correctness,
• understandability,
• authoritative (e.g. citing official sources),
• clean formatting, or
• easier to find with a search.
• Figure out why regular users and anonymous visitors have different opinions of the usefulness of the page.

The most common reason for low visitor scores is low views, which limits the number of anonymous users who provide feedback. It's not a problem if some questions are so niche that nobody outside of the community reads them. But chance visitors who find answers via search are the best source of new contributors.

More interesting: sometimes visitors disagree with the site's community about the usefulness of a question and its answers. Obviously we can't know why people who can't leave comments might differ from current users, but we can make an educated guess.

• Fix any problems you notice and describe your changes.

After doing the above analysis, you will be in good shape to edit titles and tags to help Google direct searchers to the question, clean up formatting and grammar problems, and link to authoritative sources. There's no need to wait; go edit.

The goal is to use the sample to reflect on how the site is progressing in terms of "making the internet a better place for people to get answers to their astronomy questions". If this site already does a good job, that's wonderful. If there are some things that need fixing, please talk about that. In either case, this evaluation mostly stands to improve this community's understanding of itself.

Key: link (registered voters; anonymous voters; views)

1. Is the sun any brighter during a solar eclipse (76%; 33%; 74,180)
2. What exactly is the "paradox" in Olber's Paradox? (91%; 73%; 3,064)
3. What does a solar eclipse at 99% totality look like? (72%; 15%; 8,449)
4. Why are nearby stars like Proxima Centauri and Barnard's star not visible to the naked eye? (86%; 65%; 2,985)
5. Why are three year old eclipse glasses not recommended? (57%; 18%; 868)

Footnote:

* The scores are actually the lower bound of Wilson score confidence interval for a Bernoulli parameter, which is a measure of how likely a given set of upvotes and downvotes is overall positive. For the first score, I used the sum of all votes on the question and all answers. For the second score, I used anonymous feedback. Please see the query I used for implementation details.

I've started in on this, albeit two months late, and I've finished going through one or two questions. I'll add to this list when I have more to say.

## What does a solar eclipse at 99% totality look like? (72%; 15%; 8,449)

This is one of a number of questions about eclipses we received around the time of the total solar eclipse earlier this year. Three of the five questions on this evaluation are eclipse-related, and two have very high views. This is almost certainly because of the accessibility of the phenomenon. People can go out and see a major astronomical event with their own eyes, and it takes no technical knowledge to understand what's going on. That said, there is a lot of information on the Internet (some good, some bad) about solar eclipses, so it might be hard for this question to stand out.

The first thing I noticed is that none of the answers included images, say from previous eclipses. This would have been immensely helpful. Words don't do justice when someone's asking what something looks like. A simulation linked in one of the answers was helpful, but there's no actual image information contained in the post body. We're still directing people elsewhere to get a better idea of what the eclipse would look like at 99% totality.

There are certainly plenty of places online that show either comparisons of an eclipse at different locations, or a series of photographs taken at one spot during different times of an eclipse (including totality and the periods leading up to it).

• The Washington Post compared simulated views from a number of cities including St. Louis (observing about 99% coverage) and Charleston (experiencing a total solar eclipse). The comparisons in particular help. Many news outlets have similar articles.
• Many educational institutions (e.g. here, as well as some of NASA's pages) show images at different points in an eclipse, which give possibly the best idea of what the shape of coverage looks like.

So, I do think we fall short in actually showing people what a 99% eclipse would look like, which is arguably the most important thing to cover. That could account for the 15% score. Maybe an answer with images from the 2017 eclipse would come in handy, especially from someone who was there to take them.