Wednesday, March 25, 2020

SearchResearch Challenge (3/25/20): What research questions are YOU searching out these days? (The COVID-19 edition of SRS)

We live in an information rich world... 

... but with the rise of COVID-19 / Coronavirus, we've been living in an information universe that is overwhelmed by a tsunami of content--some of it good, some it misleading, some of it questionable, and some of it outright wrong and dangerous.  

P/C CDC.gov

I can't ignore this outsize opportunity to help people with their medical searches.  So... 

Rather than pose yet-another-Challenge for you, I'd like for you to share what YOUR personal COVID-related research Challenges are.  We'll talk about them, and discuss in the comments.  From this discussion thread, I'll summarize some of the SRS issues that arise, and what strategies and tactics we should be adopting as a way to cut through the noise and get to the useful and correct information.  

So, let me pose this week's Challenge to you like this: 

1.  During the past 2 weeks, what SearchResearch Challenges have you confronted about COVID-19 / Coronavirus?  

2.  What did you do to try and answer those Research Questions?  

Naturally, I'm curious about what you're asking, and how you're answering those questions.  

Frankly, I find myself being a bit obsessive about COVID data from multiple sources.  I have one display that shows nothing but data feeds from different monitoring agencies (e.g. the CDC, Wikipedia data visualization, Johns Hopkins University, Google's data, etc.).  I'm trying to triangulate my sources to see what's going on.

One of my RQs is "what's the most accurate data source?"  This is a huge problem, especially when there's so much variance between the data sources, and when there's so much political investment at stake.  

So... What are your questions?  And, what are you doing to answer your questions?  

Let's chat.  See you in the discussion thread.  (Since I'm now working-from-home, I'll be responding more often than usual.) 

Note:  Google does have a pretty good COVID-19 website up and running.  Check there.  

Another note...  I'm also ramping up production of a few short videos to become better searchers.  You'll start seeing these coming out tomorrow, with roughly one or two a week for the duration.  This is a restart of the 1MM series ("One-Minute Morceaux" like this one on Word Order Matters) that I ran a while back.  Each is a short video between 1 and 4 minutes long with a brief SRS strategy or tactic for you.  Let me know how you like them!  

Search on!  

Wednesday, March 18, 2020

Answer: When do you need to provide context?

Context is...everything!  

As you know by now, the most common answer to a SRS Challenge is "it depends."  What does it depend on?  A:  Context.  

Gee, thanks!  

As you'll see below, context is all of the information surrounding the question and the answer.  You can't define it ahead of time because the context varies from situation to situation.  It's the other information you need to make sense of the question and answer.  

While these all seem like fairly straightforward Research Questions, I want to you think a bit about the context surrounding each of these.  What do you need to do to contextualize the RQ and the answer you give?  

Let's talk about context in the case of these three Challenges.   

1.  How many countries are there in the world?  

So, the obvious query here is: 

     [ how many countries are there ] 

The Google response is: 

You might be tempted to say the answer is 195, but the diligent and careful SearchResearcher would notice something here--something in the additional information on the page that provides a bit of context to the answer.  

Look under the "People also ask" section and notice the top two questions, "How many countries are there in 2019?" and "Are there 197 countries?"  

Already we see that there's some clarification and nuances that we need to take into account.  

   a. what's the definition of "country"? 

   b. looks like the number of countries
        varies by time... 

If you click and open those two "also ask" questions, you see a bit more context information: 

The first question points out the difference between "Sovereign states" and "State with partial recognition."  

The second question claims 197 countries + 2 UN observers (Vatican and Palestine) + Taiwan + Kosovo.  Importantly, it makes the point that this number changes by time.  What's true in late 2019 might be different a year later... or whenever it is that you ask this question.  

So... what defines a country?  (Answering this question will be important for later questions as well.)  

Let's look at the first article link (an article by Strafor Worldview, a geopolitical analysis company).  In this article they put out this analysis of what defines a country, carefully noting that there are least 6 different (and equally valid) ways to define a country: 

..."Country" and "nation" are casual words for what political scientists call a "sovereign state," meaning a place with its own borders and completely independent government. The question of which places count as sovereign states can be controversial, but for starters we normally count all the member and observer countries of the United Nations (U.N.):

U.N. members: 193
U.N. observer states: 2
Total: 195

These countries mostly all accept each other as sovereign states, and they're the ones you'll see on most world maps and lists of the world's countries. Almost every country you've ever heard of is probably a member of the U.N., and the two U.N. observer states are Vatican City (represented by the Holy See) and Palestine. If you want to know the names of all 195, Wikipedia has a complete list.

The last addition to the list was in 2012, when Palestine became a UN Observer State, and the last time the number of full U.N. members changed was when South Sudan joined in 2011...

It's interesting to notice that the Wikipedia page lists 206 total nation-states with 193 UN member states, two observer states and 11 other states.

Clearly, the definition of nation is a political topic, subject to many points of view.  In particular, the Other category on the Wikipedia page includes lots of special cases.  There are many places we think of as countries (e.g., Bonaire), but are actually part of another country (in this case, the Netherlands).  Likewise, American Samoa, Guam, and the US Virgin Islands are all territories of the US and not independent countries.  

Then there are the complex and messy politics of places like North and South Korea.  North Korea claims that South Korea is part of its territory (and vice-versa, South Korea claims North Korea as part of its territory).  North Korea is recognized by the UN, but not recognized by France and Japan.  Or are Palestine and Israel both countries?  Palestine has no official borders, while Israel isn't recognized by some 32 other UN member countries.  It's complicated... 

The important point for SRS is that the context of this question includes (a) the definition of a country, (b) what counts as a recognizable country, and (c) who's doing the recognition?  Is it the UN (195), is it countries that have partial recognition from other UN members (201), is it self-declared countries such as Artsakh (Nagorno-Karabakh), Transnistria,  and Somaliland (204), or what?

Bottom line:  There are many answers depending on your definitions and who is doing the counting.  In other words, the context here is necessary to understand what's being counted and why. 

So while I could say "there are 197 countries," the better answer is to say something like "According to the UN there are 195 countries + 2 observer states (Vatican and Palestine)."  And, of course, give links/references for where you found your data.  Oh, and of course you'll say when this count was taken, right?  

2. What's the smallest country?

You can see where this is going--after we've defined "country" now we need the context around what "smallest" means.  Do we mean population?  Area?  GDP?  

Let's just use area as the determiner of "smallest."  

     [ smallest country in the world by area ]  

directs us to the Wikipedia list of countries and dependencies by area, listing the Vatican at 0.44 square km.  But is it really a country, or merely an "observer"?  The second smallest on this list is Monaco at 2.02 square km.  It's recognized by the UN as a legitimate country and votes regularly.  

Interesting side-note:  If you look at the area of "self-declared sovereign states" you have to also include Sealand, a World War 2 platform in the North Sea off the east coast of England that has declared itself a separate country. It has an area of only 0.004 km2, but isn't recognized by any other country.  

Here, the context is fairly straightforward--it's just that you need to pull in the context around "what defines a country" in order to have it make is understandable.  

3. How many languages are spoken in India?   

Just as we had to provide some context for the definition of country, we're going to have to do the same for language.  What's a language after all? 

Let's start with the obvious query: 

     [ how many languages spoken in India ] 

Alas, the current answer given by Google in reply to this is clearly incorrect.  It says two, which is clearly the number of official languages, not the number of languages spoken in the marketplace.  (Trust me, when you're in India, you're going to hear a LOT more than 2 languages.)    

But even this is somewhat confusing. Article 343 of the Indian constitution says that the official language of the Union should be Hindi in Devanagari script (instead of the very commonly used English at the time of enactment), connecting both a language with its written form. Later, a constitutional amendment, The Official Languages Act of 1963, allowed for the continuation of English alongside Hindi in the Indian government until the legislature decides to change it.

On the other hand, if you ask Bing, you get a very different answer.  They take a much more "what's commonly spoken" approach:   

This makes more sense to the common meaning of "languages spoken in India."  The Eighth Schedule of the Indian Constitution lists 22 "scheduled" languages.  That is, these are the languages given recognition, status, and official encouragement. In addition, India recognizes several "classical" languages (Kannada, Malayalam, Odia, Sanskrit, Tamil and Telugu) that are recognized as culturally important.  see: Wikipedia article

And if you check Wolfram Alpha, we see yet another, very different answer: 

I got this page of results by clicking on the "More" button several times they list a bunch more languages, with a total of 391!  Interestingly, they list even rather small languages--the last in the list shown here is Brokskat (AKA Brokpa, Brokpa of Dah-Hanu, Dokskat, Kyango, Dardu, Hanu), which has only around 3000 speakers (according to  

According to the Census of India of 2001, India has 122 major languages and 1599 other languages. But,  due to differences in definition of the terms "language" and "dialect," these numbers are subject to some interpretation.  That census recorded 30 languages spoken by more than a million native speakers and another 122 which are spoken by more than 10,000 people. (See the India language census of 2001.)  

And of course both Persian and English are very important languages (both historically and in use today to varying degrees).  

So... let's get to the bottom line:  How many languages are spoken in India?  Officially, 2; although in reality, probably around 1700, but many of those are "smaller languages" with less than 5,000 speakers.  

According to the most recent census of 2011, after thorough linguistic scrutiny, edit and rationalization on 19,569 raw linguistic affiliation, the census recognizes 1369 rationalized mother tongues and 1474 names which were treated as ‘unclassified’ and relegated to ‘other’ mother tongue category.[42] Among, the 1369 rationalized mother tongues which are spoken by 10,000 or more speakers, are further grouped into appropriate set that resulted into total 121 languages. In these 121 languages, 22 are already part of the Eighth Schedule to the Constitution of India and other 99 are termed as "Total of other languages" which is one short as of the other languages recognized in 2001 census.

Is that enough context for you? 

Search Lessons 

Context is all the information that you need to interpret what you've found.  Clearly, this varies from person to person and RQ to RQ.  But in general I've found that searchers are far too willing to give just a basic, bald answer, and not enough context to understand the nuances.  

One big lesson: 

1.  Find and provide the context!  It's not always the case that more context is better, but you, as a discerning SearchResearcher should be able to explain (A) the definitions of the terms used in your questions and the results you've found, and (B) how the data was gathered and analyzed.  

Whenever you search for some kind of answer to a question, think to yourself--What kind of context do I need to have in order to really understand what I find?  And if you're in the business of giving answers to Research Questions from other people (e.g., reference librarians or intelligence analysts do), your answer is incomplete if it does NOT HAVE the context around it.  
Search on, contextually!  

Tuesday, March 17, 2020

The 6 things you need to know about videoconferences from home...

In these days of lockdown and quarantine, 

I thought I'd write a few posts about things we should know to make it through this time.  This is the first of my special quarantine posts!  

As work changes during this time of pandemic change, one thing is for sure:  More of us will be doing videoconferences from "informal workspaces" (that is, mostly from home--maybe your office, maybe your kitchen).  

Here are a few tips about how to be a better participant.  Do these things and your co-workers will appreciate you!  

1.  Think about what your video image looks like.  

It's pretty easy to do... just look at what your own video idea is--that's at least half of the battle!  Try to keep the background relatively uncluttered.  (Although you might want to use some of that background space to express your personal style--just be sure it communicates what you want. See the India print sheet backdrop below as an example of a bit of style...)   

For instance, here are a couple of different shots of how my home office looks with different backdrops and lighting.  (Here's a quick tutorial on lighting.)  I put up these drapes which are just sheets that I thumbtacked to the ceiling to hide all of the distracting books, boxes, and kid art that's behind me.  It's better to have a blank sheet than an office that makes you look like a hoarder.  

Plain backdrop; daylight. Note highlight on forehead.

Plain backdrop; fill light. Better skin color, more even lighting.

India print backdrop; daylight from the upper right.

India print; fill light + daylight.  

Bottom line:  Check your appearance to make sure that you look like what you want to project.  Add a light if you need one.. be sure you don't look like a ghost hanging out in a seedy part of town.  (A common lighting mistake is to have a bright backlight that makes you look like an animated silhouette.)  

2.  Mute yourself when you're not speaking.  

Too often people who aren't speaking will have accidental noises (think cats, dogs, kids, ringing phones), and when that happens, the video software will switch to focus on that person... which might not be want you want, especially if your dog is whining!  

4.  Present only the tab with your presentation in it.  

When you want to share your screen, most videoconferencing systems let you choose your whole screen, or an app/browser window.  But some (Hangouts!) let you choose a tab within the browser window.  Spend the couple of seconds it takes to figure out which tab your want to share.  (And avoid the trauma of accidentally sharing your private email window!)  

Practice this BEFORE going live.  

5.  Make sure your headphones and microphone work.  

Some videoconferencing apps (e.g. BlueJeans) have a way to test your microphone.  Use it, or conference with a friend and have them (honestly) tell you how you sound.  Be sure to check both volume AND clarity.  You don't want to sound like a quiet mouse that's underwater.  

6. Learn how to set up your video conferencing software beforehand.  

In particular, TEST things BEFORE your videoconference.  Don't be that guy who can't figure out how to get the software going.  

Take 10 minutes and practice making your software work.  

If you use Google Hangouts, there are two ways to set it up.  

A.  Use a Google Calendar invitation.  

Here's how:  First, open your Google Calendar.  Then set up your meeting: 

Once you've made the meeting, click on "Add location or conferencing."  This will automatically create a Hangout at that time.  (And remind you of it 10 minutes before the event happens.)  It will look like this when you're setting it up.  

After you set it up, click Save.  Then it will look like this in your Calendar.  

When the right time comes, just click on the "Join Hangouts" link, and you'll start up a hangout.  

B.  For spontaneous meetings, go to and make a Hangout on the fly.  


Now you can just click "Video Call" and then you'll get a popup that lets you invited anyone you'd like.  (Note that YOU need to be logged into Google, but you can invite people who aren't logged in.)  

To invite non-Googler-users, click on "COPY LINK TO SHARE"--which you can then text / Slack / IM to anyone else.  Easy peasy. 


There are lots of articles on how to do a good job of video conferences.  Here're a few:   
Google's about 6 tips for better video conferences 
Another from (a video conferencing company) 
And another from Seth Godin (who probably spends half of his life in a VC).  

And remember to practice!  A few minutes spent the day before your conference will be well worth it when you actually get to the moment of the conference. 

Search on! 

Wednesday, March 11, 2020

SearchResearch Challenge (3/11/20) When do you need to provide context?

When we answer a search Challenge... 

... we usually just give the answer, and since it's SearchResearch, we give the method by which we found the answer.  

But sometimes, an answer needs to be a bit more than just an answer.  Sometimes you need to explain your answer by providing some additional context.  

Since many of you are staying home for the next little bit, here are a couple of Challenges that will exercise your ability to give context.  Here they are... Can you not just answer the question, but provide the context needed UNDERSTAND the answer? 

Here we go! Challenges about global information!   

1.  How many countries are there in the world?  

2. What's the smallest country?

3. How many languages are spoken in India?   

While these all seem like fairly straightforward Research Questions, I want to you think a bit about the context surrounding each of these.  What do you need to do to contextualize the RQ and the answer you give?  

Let us know HOW you found the answers, and what your thought process is WRT contextualization!  

Search on!  

Monday, March 9, 2020

Answer: Does banning plastic bags actually help? How can you find out?

Sorry about the delay.  You can guess what I'm doing!  (Yes, I'm off SCUBA diving one more time.  Expect more exciting challenges about the natural history of islands.) 

However, my trip to <an unnamed location> is relevant to last week's SearchResearch Challenge.  

Plastic trash in the sea is a huge problem, albeit plastic nets are a bigger concern than plastic straws.  

But plastic bags are a bigger problem than you might think--they look like jellyfish to sea turtles, and as they break up into pieces, the plastic bag fragments look like other kinds of food, leading to fish and seabirds eating lots of plastic, which is both not nutritious and difficult to excrete.

As I mentioned, I was listening to one popular podcast about this topic on NPR (here's the transcript and the audio) pointing out that passing a law banning thin-film plastic bags in supermarkets seems to have actually caused an INCREASE in the number of plastic bags sold. 

That seems counterintuitive. And THAT's the kind of thing that makes me start doing a bit of research.  

After hearing this story about plastic bags, I naturally asked the following Research Questions: 

1. The story is clearly talking about a paper that Rebecca Taylor wrote.  Can you find that paper?  (What's the title?  Where was it published?)  
2. Once you find that paper, can you tell us where the data was collected from?  How representative is this data?  
3.  What do you call the counterintuitive effect when a partial regulation of consumer products results in the increased consumption of these products?  Is there a technical term we can use in future searches on this topic?  
4. How well has banning plastic bags worked in other places?  Can you find another study of a place where plastic bags have been banned?  How well did that work out?

1.  Finding that Rebecca Taylor paper wasn't too hard.  My query was: 

     [ Rebecca Taylor plastic bags ] 

which took me quickly to her web site,  On her site you'll find on the effects of particular economic choices, including the paper we're looking for, "Bag Leakage: The Effect of Disposable Carryout Bag Regulations on Unregulated Bags"  
which appeared in the Journal of Environmental Economics and Management in January, 2019.  

2. Where is the data from?  The obvious place to look for the sources of the data is the original paper.  However, the original paper isn't available at the above landing page.  (Why not?  The paper is published at a journal that is owned and operated by the publishing house of Elsevier.  They run many journals.  That publisher sees it in their interest to not make the original paper available as an open access article. But fear not... read on!)  

So now what?  

Many authors (thankfully!) can't publish the final form publication of their paper, but CAN make a pre-final version available.  These are often called preprints and can be found without much trouble. The idea of a preprint is to get early feedback from colleagues.  It's often somewhat different from the final version.  However... in this case, the preprint is easy to find and tells us what we want to know--the origin of the data! 

I just did this search: 

     [ filetype:pdf  "Bag Leakage: The Effect of
        Disposable Carryout Bag Regulations
        on Unregulated Bags" ] 

That is, I used the filetype: operator to search for PDFs of the title of the paper, which I surrounded in double quotes to be sure I found ONLY this paper.  

This leads us to the paper at the SSRN (Social Science Research Network, which is, oddly enough, owned by Elsevier--the publisher that didn't share the open-access paper on their primary website).  

Reading through her paper, it's clear that she collected her data from California.  Quoting Taylor's paper: 

California provides an exceptional quasi-experiment for analyzing the effects of  disposable carryout bags (DCB) policies... [thus] policies have varied greatly in both their implementation dates and locations. It is important to note that local jurisdictions decide when DCB policies will be operative; the stores within a jurisdiction do not make this decision.

This is what makes California a great location for a "natural experiment."  By collecting data from those places that implemented a change AND those places that did not, it was possible to have an experimental group and a side-by-side natural control group that did not have any new DCB policies.  

In particular, the data comes from... 

... the Retail Scanner Database collected by AC Nielsen and made available through the Kilts Center at The University of Chicago Booth School of Business. 
... I use a subset of retail scanner data from participating stores in California between January 2009 and December 2015. ... I focus my analysis on food stores (i.e., supermarkets, grocery stores, and specialty food stores), mass merchandising stores (e.g., supercenters and big-box stores), and drug stores because these stores formats regularly sell non-food grocery items, such as trash bags. ... I include stores in [cities or counties] that meet the following criteria: (1) ... is no more than 50 miles from the coast, and (2) the jurisdiction is either an entire county or can be uniquely identified based on its 3-digit zip code. ... I limit the sample to the stores in 11 counties and 8 cities uniquely identified by their 3-digit zip code. This gives me a total of 546 stores. 

Looking farther down in the paper, there are charts and measures of the representativeness of the stores.  (See Table 1) It includes cities like Santa Monica in southern California, and counties like Alameda in northern California.  Income ranges from low to high, and population varies from 92K up to 1.6M.  

It's fairly diverse.  But it IS only in California. On the other hand, the data samples 546 stores, so this is a pretty good sample size.  

3.  What do you call the counterintuitive effect when a partial regulation of consumer products results in the increased consumption of these products?  Is there a technical term we can use in future searches on this topic? 

In this paper, Taylor gives us a term in the very first line:  "Leakage occurs when partial regulation of consumer products results in increased consumption of these products in unregulated domains."  

You might know what "leakage" means, but clearly, she's using it in a specialized sense.  THAT's the sign of a technical term.  (Much like the word boot in Computer Science meaning "to start up." One boots a computer at the start.)  

So, in this paper, leakage is used as a technical term meaning the effect of partly regulation something in order to decrease its use, but instead has the effect of increasing its use.  

Obviously, the title of the paper is a play on words:  bags leak when they're not supposed to, and "bag leakage" is the counterintuitive effect of increasing use when the goal is decrease.  

I did a search for [ leakage economics ] in Google Scholar to find other uses of this term, and found many.  I didn't know this, but it's a fairly common term that basically refers to capital or income that exits an economy or system rather than remaining within it. In economics, the term refers to the outflow from a circular flow of income model. 

Taylor is extending this definition slightly to talk about plastic bags leaking out of the recycling cycle.  As Taylor says, "If unregulated consumption [ of other plastic bags ] is easily substituted for regulated consumption [of banned bags ], basing the success of a regulation solely on reduced consumption in the regulated market overstates the regulation’s welfare gains..."  

In other words, it won't help to ban a product if another nearly-the-same product is easy to get.  

4. How well has banning plastic bags worked in other places?  Can you find another study of a place where plastic bags have been banned?  How well did that work out?

I started with the straightforward search: 

     [ how well do plastic bag bans work ] 

hoping to find articles about the effectiveness of plastic bag bans in other locations.  

It was pretty easy to find articles about banning plastic bags in Kenya (National Geographic, mixed success),  Massachusetts (Wired, mixed success), Maine and Vermont (Conservation Law Foundation, uncertain), Ireland (a levy, not a ban, Smithsonian, and the Ireland Institute for Environmental Policy, works well), and Denmark (EcoWatch, works well).  And so on.  It really is, pardon my pun, a mixed bag of results.  

After reading many of these articles, it's pretty clear that banning (or imposing a tax) DOES reduce the use of lightweight plastic bags, but it's not completely a clear win because people shift to other sources of bags.  That's Leakage in the sense that Taylor defines it, and so getting to the bottom of this requires a large-scale systems analysis.  Which is what Rebecca Taylor is trying to do in this paper.  

The key point of Taylor's paper might be in the last sentence of the abstract:  "policy evaluations that ignore leakage effects overstate the regulation’s welfare gains."  

It's easy to believe that a ban will solve the problem, but it also opens up the prospect of unanticipated consequences... which is where we started this discussion.  

My bottom line:  It's complicated, but the only way to really answer our Research Question is to do a full, sophisticated, extensive analysis of what the effects of a ban are... and the leakage that inevitably comes with it.   But it's an approach that has some promise!  

Search Research Lessons 

It's tempting to try to completely resolve the "do plastic bag bans work" question here, but I'll leave that to some enterprising graduate student in environmental studies to do that kind of work.  

For us, there are several points to make vis-a-vis SearchResearching for these kinds of things... 

1.   Finding the original paper was tricky, requiring a  search for a preprint.  Keep in mind that when you hit a paywall (where you'd have to pay for the paper), it's often possible to look around and find a copy of the paper available as a preprint.  It might not be 100% the same as the final version, but as you saw, in our case, it worked really well.  In our case, using filetype:PDF as a filter got us to that preprint. 

2. Finding the provenance of the original data is easy.. once you have the paper.  After you've got that, it's a simple matter of reading a bit to figure out from whence the data came.  (That's called its provenance.)  

3.  Watch out for technical terms.  Words that are used in a special way (or specialized context) can be useful for searching for other articles that help amplify / clarify what you're searching for... Check out "leakage" in this context.  

4. Looking for other contrasting studies is a great way to get context on what you're trying to understand.  In this post, we just used a simple question ("how well do plastic bag bans work")  as a query and found lots of other sources of information from other places.  

Hope you enjoyed this romp through online search methods!  

Search on!