One of the most controversial measures in the pension reform package was suspending annual raises ((Not a cost-of-living adjustment, or COLA, as some call them)) for current retirees. I have noticed two main critiques of this element. The first criticism was that ending this practice constitutes a decrease in benefits to existing retirees who did not consent to these changes, constituting a breach of contract and assault on property rights. This critique is outside of the scope of this post. What I would like to address is the second criticism, that annual raises are critical to retirement security due to inflation, especially for the most vulnerable pensioners who earn near-poverty level wages from their pensions.

While I am broadly supportive of the changes made to the pension system in Rhode Island, I also believe that it is important to recognize the differential impact suspending annual raises has on a retired statehouse janitor who currently earns $22,000 a year from their pension and a former state department director earning $70,000 a year from their pension. Protecting the income of those most vulnerable to inflation is a worthy goal¹.

I have a simple recommendation that I think can have a substantial, meaningful impact on the most vulnerable retirees at substantially less cost than annual raises. This recommendation will be attractive to liberals and conservatives, as well as the “business elite” that have long called for increasing Rhode Island’s competitiveness with neighboring states. It is time that Rhode Island leaves the company of just three other states– Minnesota, Nebraska, and Vermont– that have no tax exemptions for retirement income². Rhode Island should exempt all income from pensions and social security up to 200% of the federal poverty level from state income taxes. This would go a long way to ensuring retirement security for those who are the most in need. It would also bring greater parity between our tax code and popular retirement destination states, potentially decreasing the impulse to move to New Hampshire, North Carolina, and Florida.

It’s a progressive win. It’s a decrease in taxes that conservatives should like. It shouldn’t have a serious impact on revenues, especially if it goes a long way toward quelling the union and progressive rancor about the recent reforms. And it’s far from unprecedented– in fact, some form of retirement income tax exemption exists in virtually every other state.

We should not be proud of taking away our most vulnerable pensioners’ annual raises, even if it was necessary. Instead of ignoring the clear impact of this provision, my hope for 2013 is that we address it, while keeping an overall pretty good change to Rhode Island’s state retirement system.

1. Interesting, increases in food prices has largely slowed and the main driver of inflation are healthcare costs. I wonder to what extent Medicare/Medicaid and Obamacare shield retirees from rising healthcare costs
2. http://www.ncsl.org/documents/fiscal/TaxonPensions2011.pdf

I have grown increasingly unhappy with WordPress lately. My blog is simple. My design tastes are simple. My needs are simple. I like control. I am a geek. And I really need an excuse to learn Python, which seems to be rapidly growing into one of the most important programming languages for a data analyst.

I have decided to migrate this blog over to Pelican, a static site generator written in Python. Static sites are the “classic” way to do a webpage– just upload a bunch of HTML and CSS files, maybe some Javascript. But no databases and no constructing the page a user sees in the browser as they request it. This puts substantially less strain on a web server and makes it far easier to export and move a webpage since all you need to do is duplicate files. What makes static sites a real pain is that there is a lot of repetition. Folks adopted dynamic sites that use content management system so that they can write a page called “post.php” one time, and for each unique post just query a database for the unique content. The frame around the post, layout, components, etc are all just written once. Static site generators allow you to build a webpage using a similar, but far more stripped down, layout system. However, rather than generate each page on the web server, you generate each page by running a script locally that transforms plain text documents into well-formed HTML/CSS. Then you can just upload a directory and the whole site is ready to go.

Pelican comes with a pretty good script that will take WordPress XML that’s available via the built-in export tools and transform each post into a reStructuredText files, a format similar to Markdown. I prefer Markdown so I used pandoc to convert all my *.rst posts into *.md files.

So far, so good.

But one of the really big problems I had with WordPress was a growing dependency on plugins that added non-standard, text-based markup in my posts that would be rendered a particular way. For example, text surrounded by two parenthesis, ‘((‘  ’))’, became a footnote. For code syntax highlighting, I use a “short code”, which puts “sourcecode language=’r'”, for example, between brackets []. All of these plugins have been great, but now when you try to export a post you get the non-standard markup in-line as part of your posts. It makes it very difficult to recreate a post the way it looks today.

This presents a great opportunity to learn a little Python. So I have begun to scrounge together some basic Python knowledge to write some scripts to clean up my Markdown files and convert the syntax of the short codes that I have used to properly formatted Markdown so that when I run the pelican script it will accurately reproduce each post.

Unfortunately, I’ve hit a snag with my very first attempt. Footnotes are a big deal to me and have standard Markdown interpretation. In Markdown, footnotes are inserted in the text where “[^#]” appears in the text, where # = the footnote identifier/key. Then, at the end of the document, surrounded by new lines, the footnote text is found with “[^#]: footnote text” where # is the same identifier. So I needed to write a script that found each instance of text surrounded by two parentheses, insert the [^#] part in place of the footnote, and then add the footnote at the bottom of the post in the right format.

I created a test text file (ignore the \ in the first set of parenthesis, it’s the only way I could escape the WP-Footnote plugin):

This is a test (\(test footnote)).
And here is another test (\(footnote2)). Why not add a third? (\(Three Three)). 

The goal was to end up with a file like this:

This is a test [^1]. And here is another test [^2]. Why not add a third? [^3].

[^1]: test footnote

[^2]: footnote2

[^3]: Three Three 

Unfortunately, the output isn’t quite right. My best attempt resulted in a file like this:

This is a test [^1] And here is another te[^2])). Why not add a t[^3]ree)).

[^1]: (\(test footnote))

[^2]: (\(footnote2))

[^3]: (\(Three Three))

Ugh.

So I am turning to the tiny slice of my readership that might actually know Python or just code in general to help me out. Where did I screw up? The source to my Python script is below so feel free to comment here or on this Gist. I am particularly frustrate that the regex appears to be capturing the parenthesis, because that’s not how the same code behaves on PythonRegex.com.

If anyone can help me with the next step, which will be creating arguments that will understand an input like *.rst and set the output to creating a file that’s *.md, that would be appreciated as well.

import re

p = re.compile("\(\(([^\(\(\)\)]+)\)\)")
file_path = str(raw_input('File Name >'))
text = open(file_path).read()

footnoteMatches = p.finditer(text)

coordinates = []
footnotes = []

# Print span of matches
for match in footnoteMatches:
    coordinates.append(match.span())
    footnotes.append(match.group())

for i in range(0,len(coordinates)):
    text = (text[0:coordinates[i][0]] + '[^' + str(i+1)+ ']' +
            text[coordinates[i][1]+1:])
    shift = coordinates[i][1] - coordinates[i][0]
    j = i + 1
    while j < len(coordinates):
        coordinates[j] = (coordinates[j][0] - shift, coordinates[j][1] - shift)
        j += 1

referenceLinkList = [text 1="'
'" language=","][/text]
for i in range(0, len(footnotes)):
    insertList = ''.join(['\n', '[^', str(i+1), ']: ', footnotes[i], '\n'])
    referenceLinkList.append(insertList)

text = ''.join(referenceLinkList)

newFile = open(file_path, 'w')
newFile.truncate()
newFile.write(text)
newFile.close()

Update with solutions:

I am happy to report I now have two working solutions. The first one comes courtesy of James Blanding who was kind enough to fork the gist I put up. While I was hoping to take a look tonight at his fork tonight, Github was experiencing some downtime.  So I ended up fixing the script myself a slightly different way (seen below). I think James's approach is superior for a few reasons, not the least of which was avoiding the ugly if/elif/else found in my code by using a global counter. He also used .format() a lot better than I did, which I didn't know existed until I found it tonight.

I made two other changes before coming to my solution. First, I realized my regex was completely wrong. I didn't want to capture anything within the two parenthesis when no parenthesis were contained, as the original regex did. Instead, I wanted to make sure to preserve any parenthetical comments contained within my footnotes. So the resulting regex looks a bit different. I also switched from using user input to taking in the filepath as an argument.

My next step will be to learn a bit more about the os module which seems to contain what I need so that this Python script can behave like a good Unix script and know what to do with one file or a list of files as a parameter (and of course, most importantly, a list generated from a wild card like *.rst). I will also be incorporating the bits of James's code that I feel confident I understand and that I like better.

Without further ado, my solution (I updated the gist as well):

from sys import argv
import re

name, file_path = argv

p = re.compile(r"[\s]\(\((.*?[)]{0,1})\)\)[\s]{0,1}")
# The tricky part here is to match all text between "((""))", including as many as
# one set of (), which may even terminate ))). The {0,1} captures as many as
# one ). The trailing space is there because I often surrounded the "((""))" with
# a space to make it clear in the WordPress editor.

# file_path = str(raw_input('File Name >'))
text = open(file_path).read()

footnoteMatches = p.finditer(text)

coordinates = []
footnotes = []

# Print span of matches
for match in footnoteMatches:
    coordinates.append(match.span())
# Capture only group(1) so you get the content of the footnote, not the whole
# pattern which includes the parenthesis delimiter.
    footnotes.append(match.group(1))

newText = []
for i in range(0, len(coordinates)):
    if i == 0:
        newText.append(''.join(text[:coordinates[i][0]] +
                               ' [^{}]').format(i + 1))
    elif i < len(coordinates) - 1 :
        newText.append(''.join(text[coordinates[i-1][1]:coordinates[i][0]] +
                          ' [^{}]').format(i + 1))
    else:
        newText.append(''.join(text[coordinates[i-1][1]:coordinates[i][0]] +
                          ' [^{}]').format(i + 1))
        # Accounts for text after the last footnote which only runs once.
        newText.append(text[coordinates[i][1]:]+'\n')

endNotes = []
for j in range(0, len(footnotes)):
    insertList = ''.join(['\n','[^{}]: ', footnotes[j], '\n']).format(j + 1)
    endNotes.append(insertList)

newText = ''.join(newText) + '\n' + ''.join(endNotes)

newFile = open(file_path, 'w')
newFile.truncate()
newFile.write(newText)
newFile.close()

Educators may have the best intentions when trying to infuse their practice with evidence, but they all too often are not equipped as critical consumers of research. Worse, the education profession has historically been wrapped in “thoughtworld”¹, where schools of education have taught same ideas about effective teaching and learning for decades without a basis in empirical research. These same ideas are taught to principals, district administrators, and teachers, so nary a critical voice can stop the myths from being repeated and mutually reinforcing each other.

Effectively conducting empirical research, translating research for policymakers, and implementing research-based program design is my job. I came to education purely from a research and policy perspective, and I am equipped to understand some of the empirical research done on effective schooling².

I have to confront an awful history of “outsiders” like myself who have brought round after round of poorly supported, poorly evaluated reforms. I have to confront the history of districts and schools discarding some very effective programs because of leadership changes, lack of resources, and most of all a lack good, systematic evaluation of programs. And I have to be damn good at what I do, because even a small misstep could paint me just like every other “expert” that has rolled through with the newest great idea.

I think this is why I tend to favor interventions that are very small. Simple, small, hard to “mess up” interventions,  based in research, implemented just a few at a time have tremendous potential. I love the oft-cited work on filling out the FAFSA along with tax filing at H&R Block. It is simple. There is no fear of “dosage” or implementation fidelity. There are both sound theoretical reasons and empirical results from other domains that suggest a high likelihood of success. It has the potential to make a huge impact on students without adding any load to teachers who are, say, implementing a brand new and complicated curriculum this year. This is how you earn trust through building success.

I am also a fan of some really big, dramatic changes, but how I get there will have to be the subject of a future post.

1. E.D. Hirsch’s term
2. In the area of neuroscience and cognitive science, I am probably only marginally better off than most teachers. My Sc.B. is in chemistry. So a background in empirical physical sciences and my knowledge of social science may help me to access some of the research on how people learn, but I would probably be just as susceptible to overconfidence in my ability to process this research and repeat untruths as many very intelligent educators

Time and time again, following descriptions of well designed and built hardware¹, reviewers were forced to add some form of, “But the ecosystem cannot compete with Apple’s 275,000 tablet-optimized application.” I think this understates the power of Apple’s amazing developer advantage.

I use three distinct computing platforms every day: my phone, my tablet, and my traditional PC (laptop and desktop). There are times where I use an application which is specific to one platform or the other. Dark Sky, for example, is incredibly useful on my iPhone but would be pretty pointless on my Mac Mini or Macbook Air. This kind of platform-specific, quality application is what most would consider the App Store advantage. Not me.

Apple’s true advantage is when applications are available across all three platforms, offering simultaneously a device-optimized and consistent experience no matter what I am using.

They offer a frictionless experience.

There is a good reason people were so excited for Tweetbot for OSX and love to use Reeder on iPhone, iPad, and OSX. The features, feel, gestures, and even notification sounds having consistency across environments makes it easier to use computers. The so-called “halo effect” of the iPod was widely discussed in the early 2000s. iTunes on every Windows machine represented the tail end of a long play that pushed the promise of frictionless computing with Apple products. iOS delivers on this promise in spades.

Google knows a big selling point of Android is offering the best mobile experience with their web products. As an early and voracious user of Gmail, Google Contacts, and Google Calendar, I do find this enticing. But Android apps are never going to be able to offer the frictionless experience offered by Apple across the mobile and desktop space. ChromeOS is Google’s best effort to push a frictionless platform, but it’s entirely limited to non-native applications so anything but Google products require major modifications and just won’t be the same.

Microsoft sees the Apple advantage clearly, and they understand Google’s inability to fully compete. That’s why they are launching Windows 8, in many ways attempting to even further integrate the tablet and desktop than Apple. The Surface, and Windows 8 writ large, is a bet that Apple made a mistake grouping tablets with cell phones. The tablet, according to Microsoft, is about replacing laptops and should be grouped with the desktop.

I think this is a smart play, regardless of some of the rough reviews of both the Surface RT and Windows 8. Version 1 has some awkward transitions on both devices, but that may be worth the cost to take advantage of a near-future where the power available on a large tablet will be comparable to that of a laptop or even desktop computer. Just as the Macbook Air is every bit as good a consumer computer as “the fatter” laptop market, soon tablets will be every bit as good a consumer computer that exists. Microsoft’s bet is that with that power will come more sophisticated and complex uses, better suited to applications at home on the desktop. They are betting the future is the past– a full multitasking enabled, file-system revealing environment. If that’s what users will eventually want from their tablets, Windows 8 will have these capabilities baked in from the start while iOS struggles to pump out new features and APIs to mimic (or create) these capabilities.

The future is frictionless. Apple’s true advantage is they can already offer one version of that future. If Microsoft plays its cards right, and if it is not too late², they can offer an equally compelling alternative. It won’t win over the real, dyed-in-the-wool Apple fans, but it may stem the tide carrying the consumer market swiftly away.

1. Hardly a given in the past from either Google (LG/ASUS) and Microsoft partners. Although Microsoft’s actual hardware, until now primarily keyboards and mice (do you pluralize a computer mouse? It seems strange.)
2. I really think it might be. Windows Phone 7 was brilliant, but released 2 years too late behind at least 1 year of development.

In many ways, the elementary school model works better for middle school students and vice versa.

Anyway, some research showing K-8 schools have a built-in advantage against the traditional middle school:

The Middle School Plunge

Stuck in the Middle ((A more “popular” version on Education Next here.))

A small rant.

]]> http://blog.jasonpbecker.com/blog/2012/10/17/some-reasons-redeveloping-i-195-land-wont-save-downcity/feed/ 0 http://blog.jasonpbecker.com/blog/2012/10/09/where-i-share/ http://blog.jasonpbecker.com/blog/2012/10/09/where-i-share/#comments Tue, 09 Oct 2012 17:44:36 +0000 jason http://blog.jasonpbecker.com/?p=407 I have been meaning to write this post for the past couple of weeks. Like most other people, I am constantly experimenting with different ways to publish and share my thoughts and engage with social networking. Lately, I have settled into what feels like an “end state” workflow¹. I will devote a future post to the details of how I manage my online reading, writing, and sharing workflow but for now I just wanted to let folks know where they can find me.

For random thoughts throughout the day I mostly turn to my Twitter account or increasingly my App.net account. I am a retweet abuser, so if you follow me there be warned. I often just retweet things I find funny or interesting, write some random complaint about coding, policy, or education when I’m frustrated and don’t understand the world, and try syndicate some of the other sources I’ll list here. I also like to talk to people on Twitter, so if you’re looking for conversation that’s the place to go. I almost use it like it’s the new IRC/AIM Chatroom. My Twitter account is a bit more Providence/Rhode Island heavy than most other ways to follow me.

Some of you may know that I also have a Tumblr that has fallen in and out of favor. I used to blog over there before creating this WordPress site². Recently, I have used my Tumblr account much more. Since Google Reader removed its social features I have tried to find the best way to share the best stuff I read each day with a few thoughts. I toyed with Google Plus, but it really is dead. I don’t find good content there and engagement with my sharing has been very low. Also, the lack of a write API makes it very challenging to incorporate in a non-disruptive way.³ So right now, head over and follow my Tumblr (natively or RSS) if you want to get 5-10 link posts each day of things I’ve collected across the web. Some of my favorite online friends found me through my Google Reader sharing and I suspect that they would enjoy my Tumblr most of all. If I start getting more engagement around what type of links folks are enjoying I can begin to shift the topics I post on. I collect many more links than what end up in Tumblr in Google Reader and Pinboard. I have a very specific path to end up in Tumblr that leans more towards long reads and shares from friends and not what I am watching on RSS.

A few months ago I ditched my original Facebook account from 2005 and reopened a fresh one. I did this for two reasons: 1) I had collected many friends that I was not truly in contact with. Because of the layers and layers of privacy changes that Facebook went through, it became very difficult to maintain settings I was comfortable with. I wanted to start fresh with friends and fresh with how I manage privacy. 2) Related to 1, I never used Facebook as a networking tool. To me, it was always supposed to be a way to interact and keep in touch with friends from “real life”. Ultimately, I didn’t find that aspect of Facebook to be all that valuable. So I’m trying to be a believer and use Facebook more like I use other social media–  a way to tap into my “interest graph” and meet new people and read new things and have new conversations. You can follow me there with a few caveats. I hate using Facebook, so it is probably going to have the least content. There will still be some personal stuff as most of my friends still see Facebook as an intimate space, shocking though that may seem. Finally, I may not friend you back. Yes, the point of this account is to be more open, but Facebook still creeps me out and on any given day I may feel more or less incline to be open on there.

This blog will remain where I write longer pieces that are primarily “original” analysis/thoughts and less news/broadcast-like. I hope to share a lot more code and thoughts on current research in the near future now that I’m changing jobs.

1. Subject to change, but I’m betting it’s more tweaks at this point than dramatic shifts
2. and I really want to leave WordPress, but that is going to be a big project
3. Definitely more on this in my future workflow post

Building great mapping data is no easy task. It takes years. It takes human intervention. It takes users. Short of a massive acquisition of an existing player, like Garmin, there was little hope of Apple developing a great map application for day one of release. Hell, in my experience, most stand alone GPS data is pretty awful in all the ways the Apple data is awful. That’s why I primarily used my iPhone as my GPS the last few years. The experience was consistently better and less frustrating. Perhaps even more critically, Apple is just not a data company. Google is the king of data. The skills required to build great geographic data simply doesn’t map well against previous Apple competencies. None of this means that the Apple Map situation is good or even “excusable”. I just think the map situation is “understandable” and would not be with different guidance.

But then I reevaluated and realized that there is a major way that management could have improved Apple Maps for iOS. Managers should set the bar for quality, make sure that bar is met, and adjust both resources and expectations when a project is not meeting user expectations. It must have been obvious to Apple management that the quality expectations were not going to be met.

What could Forstall have done? Some have suggested thrown substantially more money at the project. Others say he should have “winked” at Apple users and clearly signaled that Maps were in their infancy. And of course there were those who said he should have waited another year for the Google Maps contract to expire. John Gruber is rather convincing that simply waiting another year was not an option. Apple really couldn’t swap maps out of iOS in the middle of the OS cycle. It would be jarring and far more frustrating than the current situation.

I would have recommended a third option.

Apple should have released iOS6 Maps as US only.

@jdalrymple what if Apple execs realized it wasn’t going well & made maps US only & world in 6-12mo. Still had Google contract time for that

— Jason Becker (@jasonpbecker) September 29, 2012

One of the major themes of the iPhone 5 release was that this was a global phone. Global LTE, with day one launches in more countries and reaches far more countries, faster, than ever before after that. In fact, the Verizon CDMA iPhone comes with an unlocked GSM radio. But mapping is hard, and that problem becomes orders of magnitude more difficult with each inch of the planet that needs to be covered. When it became clear that Apple had a beautiful application, but awful data, Forstall and the rest of Apple management should have adjusted expectations and promised a US-only release that met the quality that consumers have come to expect. This would serve to increase resources, while winking at users, and utilizing the remainder of the Google contract for international mapping. With six additional months Apple could make great strides improving international data and possibly signing some additional, high-profile maps data deals with local sources/competitors that would love to be associated with Apple, even if it is just in a footnote. US users would rave about the great vector mapping, the turn by turn directions that are brilliantly integrated into the lock screen and always provide just enough information, and the cool integration into Open Table and Yelp. US maps would get better because they would have constant users. The rest of the world would lap up iPhone 5s and wait anxiously for their chance to taste the Great Apple Maps.

In this scenario, it is possible that Apple could have had the best of both worlds: a far worse data set in an application that cost just as much, but by limiting the scope to their key market, a reputation for excellence that would lead to excitement for the end of a competitor’s product.

I am sure there were other challenges with producing a US-only² that I am not considering. I think this is at least one typical techniques in IT management that Apple could have employed for a smoother, better release of their first efforts into a complicated and competitive space.

1. Of iOS skeumorphism fame.
2. Or North America only. There are barely any roads in Canada, right?

The thrust of Tough’s argument, if I were to be so bold, is that there is a definable set of non-cognitive skills, called “grit”, that are at least as important as academic achievement in determining long-term positive outcomes for kids. Great schools, therefore, would do well to focus on developing these habits as much and as intentionally as they do developing content knowledge and academic prowess. This, according to Tough, is a big part of the “magic sauce”¹ of “No Excuses” schools like KIPP. They teach “grit” as a part of their intense behavioral management and culture efforts.

I think Tough is an engaging writer and has a great knack for finding some of the most interesting research not often read in education policy circles, but which is clearly relevant. While listening to the EconLog podcast I found myself often disagreeing with his interpretations/conclusions. But more often, I found myself desperately wishing for a different, slower format because so much of this work begged deeper questioning and conversation. What better reason could there be to buy and read a book-length treatment of these ideas?

Anyway, I thought I’d share just a few of my thoughts on “grit” based on this interview and the earlier New York Times Magazine piece.

Teaching conscientiousness in a society that has been so unconscientious

It seems fairly obvious that people who don’t “play by the rules” and aren’t easily motivated to conform to certain habits are less likely to be successful. It is unsurprising that Tough finds research that suggests that there is a “grit” gap between rich and poor. I want to know more about why, and I have, what I hope, is one interesting idea of what contributes to the “grit gap”.

I believe that deterioration of the built environment, especially among the urban and truly rural poor, is a major contributor to low grit. Some parts of this country with high concentrations of poverty look– bombed out. Roads are littered with deep potholes and scars. The houses have chipped paint, rotting wood exterior elements, and unkept yards. Storefronts were built decades ago on the cheap, aged poorly, and were never updated. Their schools lack good lighting, decent HVAC systems, and functioning toilets. There is no pride found in any of these spaces.

Children growing up in poverty do not see neighbors obsessing over their lawn. They do not watch one house after another repaint and reface their exteriors to ensure they weren’t the ugliest house on the block. They do not see brand new cars, fresh asphalt roads, and schools that resemble palaces. I don’t think virtually any of this has to do with the people who live in these neighborhoods. I do think it reflects the pathetic state that society has deemed acceptable, so long as it remains sight unseen by those with resources.

Growing up in poverty often means being surrounded by spaces that society has left to rot. How can these children learn conscientiousness when the privileged have been so unconscientious?

The M&M Study

Tough mentions a study where students first take an IQ test under normal conditions. These same students are then given an IQ test but are rewarded with an M&M each time they get a question right. This tiny immediate incentive resulted in a massive, 1.8SD improvement in mean IQ.² The implications are fascinating. It demonstrates the importance of motivation even while taking a test that is supposedly measuring an intractable, permanent attribute people have. This seems obvious and is fairly well known, but forgotten in many policy circles. I have often lamented that the New England Common Assessment Program has a sizable downward bias when measuring achievement because the exam is low stakes for students. The dramatic decrease in performance observed on the 11th grade NECAP math exam is almost certainly in part due to lower intrinsic motivation amongst high school students compared to their 8th grade and younger selves.

There are some students that have no measurable response to the M&M incentive. These students are exhibiting qualities of Tough’s “grit”, conscientiousness that leads one to do well simply because they are being measured, or perhaps because there is no reason to do something if it is not going to be done well. I believe that there is also a bias against schools with concentrated poverty because of an uneven distribution of “grit”– suburban middle to upper class students with college ambitions will likely be the students who will sit down and try hard on a test just because they are being measured whereas urban students living in poverty are far less likely to exert that same effort for an exercise with no immediate or clear long-term consequences.

All of this would be pretty blasé were it not for the more distal outcomes observed. The group of students that did not respond to the M&M incentives had significantly and practically better outcomes than those that responded to the incentive. I can’t recall exactly which outcomes were a part of this study, but Tough cites several independent studies that measure a similar set of qualities and find far better outcomes with GPA, graduation from high school, post-secondary degree attainment, juvenile delinquency or adult criminal activity, and wages.

Tough’s interpretation of these results seems to mirror my feelings on grading. Low stakes testing (or in this case, no-incentive testing) has omitted variable bias which leads to observing students who lack “grit” as lower achieving than they are. The test results are still excellent predictors of later success but lack validity as a pure measure of academic achievement. My complaint about grades that use behavior, attendance, and participation³ does not stem from their lack of validity at predicting later outcomes. These grades are excellent predictors of later outcomes. Rather, it stems from these grade conflating two very different qualities into a single measure, making it far more difficult to design appropriate interventions and supports that target individual needs.

Tough seems thinks this means that high stakes placed on test scores over emphasizes one quality over the other when both are very important. I disagree. I feel that high stakes test scores recreate the M&M incentive and leads to a better measure of academic ability. That is not to say that we don’t need to cultivate and measure non-cognitive skills. It just means that trying to measure both at once⁴ results in less clear and actionable interpretations.

Is the “grit” problem properly described as a failure to recognize long-term benefits?

Repeatedly both Tough and host Russ Roberts point to the need to provide students who lack grit more information on the long-term benefits of “doing well”. For example, Tough cites KIPP’s posting of the economic benefits of a bachelor’s degree on walls in the halls of their schools as a way to build grit. Somewhat left unsaid is the idea that grit-like behaviors may not describe some kind of “intrinsic” motivation, but instead represent an understanding of the long-term extrinsic benefits of certain actions. Grit really means understanding that, “If I behave appropriately, I will gain the respect of this authority and earn greater autonomy/responsibility,” or perhaps, “Doing my homework each night will teach me good habits of work and help me to learn this academic material so I can succeed in college and get a better job.”

Can grit really be just a heuristic developed to better respond to long-term incentives?

I am not sure. I am equally unsure that the activities of a “No Excuses” school actually generate the long-term benefits of “grit”. If grit is a powerful heuristic to optimize long-term outcomes, how do we know that many short-term incentives that build behaviors toward academic success mean that students better respond to a broad set of long-term outcomes? Should we believe that behavior bucks/demerit systems, constant small corrections, repeatedly stating the goals of education and its benefits, and other KIPP-like culture-building strategies build a bend toward acting in ways that maximize long-term outcomes? Do students aspire to college because they have internalized its importance, or do the stack of short-term incentives build a desire for sprokets, wignuts, and widgets that just happened to be called a “bachelor’s degree” in this case?

1. My term, not his. Probably stolen from one of my colleagues who uses this term a lot.
2. From 79 to 97 according to EconTalk
3. among other non-cognitive, non-academic skills and activities
4. Or inadvertently measuring both at once, as many low-stakes standardized tests do

But isn’t mastery of content essentially, “Do well on your assignments and tests”? And while a massive, standards-based report card may be hard for a parent to read, is it any more confusing than seeing awful results on standardized tests and a student who clearly doesn’t read on grade-level receive good grades because of participation, attendance, and behavior? As a parent, how do you know to intercede on your child’s behalf when you see a “B” which actually represents a C- on content knowledge and skills and an A+ for effort, behavior, and completion?

Ultimately, I am against including behavior, attendance, and effort as a part of the same grade as academics. I think there needs to be a clear place to present evidence of academic ability and growth independent of behavioral growth. Both are important, and while linked, are certainly not moving in lockstep for the typical child. Accurate information in both domains is far better than falsely presenting a singular, mixed-up “truth” about a child’s success in school.

For the same reason I am not a fan of school report cards with a single letter grade rating, I am not for just a single letter grade for students. Ultimately, they both represent poor combinations of data that obscure more than they reveal.

Developing report cards or “grading” systems, both for program evaluation and for students, always conjures one of the few concepts I recall from linear algebra. It seems to me that any good grading system should provide a basis, that is, a minimal set of linearly independent vectors which, via linear combination, can describe an entire vector space. Remove the jargon and you’re left with:

Measure the least amount of unrelated things possible that, taken together, describe all there is to know about what you are measuring.

A single grade that combines all the effort, behavior, attendance, and various unrelated academic standards I might get an overall description that says “round”. But by separating out the data at some other level, the picture might describe a golf ball and its dimples, a baseball and its stitches, or a soccer ball with its hexagon-pentagon pattern.

I think we need to find a way to let people know what kind of ball they have.