Thursday, April 21, 2011

Weekly Tech Tip: Making and Using Infographics

I am sorry it has been so long since my last Weekly Tech Tip but we have been doing state testing this month. This eats up district bandwidth so most use is prohibited for both students and teachers. Use of the Internet interferes with the ability for Pearson Education to deliver online assessments for both Math and Science and since we can only test a handful of kids at time (due to how much bandwidth it uses) it takes nearly the whole month to get all these students tested. As a result, most of this week's tech tip had to be done without the use of the Internet. I have also been assigned to proctor these tests so I couldn't talk while making this week's tech tip. I apologize for the lack of narration. Anyway, this week's tech tip is on infographics and how they can be used to help students make sense of numbers while analyzing complex problems using data and statistics. Since I have been knee deep in state testing I thought data regarding testing would make for an appropriate data set for this tech tip. Enjoy.

Weekly Tech Tip:


related links:
  • Infographic Generators:
Video Festival:
Link Stew:
Blog Carnival:

Retweetable Tweets:

Where there's smoke, there's fire. Where there are standards, there's standardized testing.less than a minute ago via Twitter for Mac Favorite Retweet Reply



@DianeRavitch Principal of my child's NYC charter school has asked the Board to change her title to Chief Education Officer - yes, CEO.less than a minute ago via Twitter for iPhone Favorite Retweet Reply



Here are some amazing videos related to mathematics. http://bit.ly/gc9muG #edchat #mathchatless than a minute ago via TweetDeck Favorite Retweet Reply



@markgarrison "'tchr wanted to know if they learn schoology today, are they going to have to learn something new in five years.' I hope so"less than a minute ago via Power Twitter Favorite Retweet Reply



Dewey: what the best and wisest parent wants for his child is what we should want for all children...anything less..destroys our democracyless than a minute ago via web Favorite Retweet Reply


On CNN just now, @m_rhee just said that the impact of a great teacher is "immeasurable." She said it herself.less than a minute ago via TweetDeck Favorite Retweet Reply


. @techlearning The "pro-teacher" language of individualism & personlization is code for drill & a steady diet of standardized testing.less than a minute ago via web Favorite Retweet Reply


@garystager it's staggering to think abt the cost of testing in US- supposed to be cheap, efficient- NOT.less than a minute ago via TweetDeck Favorite Retweet Reply


@DianeRavitch Maybe it's time to inform #parents that they can opt their kids out of #standardizedtests -- especially unreasonable onesless than a minute ago via web Favorite Retweet Reply

Tuesday, April 19, 2011

Make Your Own Infographics - How State Testing Impacts Technology Access

Yesterday I crunched some numbers and posted some infographics showing a conservative estimate of how much money standardized testing is costing schools. In reality the amount the state spends is probably much larger since this figure doesn't account for the salaries of employees at the Minnesota Department of Education whose jobs are either explicitly devoted to or have come to be consumed with test-driven data. It also doesn't account for the benefits packages of many of the employee salaries included in my figures. It doesn't account for capital expenses or utilities spent in the service of testing or test prep including computer labs, electricity, buildings, heating costs, maintenance staff, and all the other things that schools spend money on to keep their buildings operational. One could, I suppose, figure out what percentage of the school year is monopolized by these high stakes tests and their subsequent culture and paraphernalia. If all those were factored in, the actual cost of standardized testing becomes astronomical.

Today, however, I would like to focus on how these high stakes tests impact student access to school technology resources, namely computer technology.

Now school districts vary slightly their student contact days but most schools in Minnesota are in session around 172 days per year. Of those days:
  • 20 monopolize district technology resources for the MCAs. This includes not only student-use computers but also the internet and intranet throughout the district. So, even if a teacher's class is not being tested their classroom use of technology (either student computer stations or teacher computer including whatever machine they might use to display content on an interactive whiteboard).
  • 60 monopolize district technology resources for the NWEA. The same issues regarding Internet and intranet also apply.
  • 60 monopolize district technology resources for online test prep.
This leaves 32 instructional days where both the Internet and student computer labs are accessible to students for something other than testing or practice testing. This is shown in this infographic I created the other day using ManyEyes:

School District Technology Use Many Eyes
In those 32 days where district technology is not monopolized by testing or test prep scarce resources must be shared. This means that if district technology resources are only available for 18.6% of the year for higher-order learning experiences and those resources have to be shared that in a school with a 4:1 ratio of students to computers that they are really only available 6.2% of the time.

So, if testing and test prep monopolize 81.4% of the school's technology resources then we can clearly attribute that percentage of the district technology costs to the impact of testing.

But, lets take just a conservative estimate and only take a small percentage of these expenditures into our definitive figure for our running tally of how much state testing costs Minnesota. To do this we will assume that two computer labs in each building are utilized for testing and/or test prep. If these labs are on a 7 year replacement cycle then the equation to figure this out would look something like the following:

f(technology cost of testing)=2[(Cost of a computer lab X .814)/7] + (annual cost of internet * .814)

So, lets say the average computer lab costs $30,000 and the annual cost for internet in each school is around $30,000 (taken from my own school district data) then:

f(technology cost of testing)=2[($30,000 X .814)/7] + ($30,000 * .814)
f(technology cost of testing)=2[($24,420)/7] + ($24,420)
f(technology cost of testing)=2($3,488.57) + ($24,420)
f(technology cost of testing)=$6,977.14 + $24,420
f(technology cost of testing)=$31,397.14

So, according to the state website Minnesota has 2,637 schools so multiply f(technology cost of testing) by this number and we should have a conservative estimate of how much expense is spent on technology alone. That ends up being: $82,794,258

Add that to the running total from my last post we are now at:

$939 Million Dollars



Of course, if it were not for these tests most of this money would still be spent but it would be spent differently. Still the only discernible extra spending I can find so far is the $72 million dollars I discussed yesterday.

Monday, April 18, 2011

Make Your Own Infographics - What State Testing Costs Minnesota Schools

Ok, so far I have used Wordle, StatsPlanet, ManyEyes, and Gapminder to create infographics. But, when it comes to visualizing data, often the best infographics are ones you create yourself using a plain old image editor. For the following infographics I am going to use a combination of Microsoft Excel to calculate the data and Paint.net to spruce things up.

So, lets start with a baseline. $23 million is what the state of Minnesota pays Pearson Education each year to manage the Minnesota Comprehensive Assessment Tests in Math, Science, Reading, and Writing.


So, what could the state of Minnesota buy with that $23 million dollars?

Lets assume that each of these six teachers = 10 teachers:

And lets assume that this school bus = 1 field trip for every classroom of 25 children in the State.


And, lets assume that every book in this stack represents one book averaging $25/ea given to each classroom in the state of Minnesota:
And, lets assume that this laptop represents 1 laptop for each classroom in the state of Minnesota:

And, lets assume that this iPod Touch represents 1 hand-held multi-touch mobile computing device per classroom in the state of Minnesota:

So, what would $23 Million Dollars buy?


Now, the actual cost of testing goes beyond what we pay Pearson Education. These tests alone cost the district quite a bit in time, money, and resources to administer. Pearson does not provide proctors or testing equipment to schools so schools are left footing the bill for this. So, how much does it cost school districts each year?

Now, in my experience a school can only realistically test 25 students a time for a computer-based assessment. The state of Minnesota has 823,826.00 students in public school. But, tests are only given to 4 grade levels. If we divide the total number of students by 13 we will get an average number per grade level, then if we multiply that number by 4 we will get a good idea of how many students will be taking these tests. If we divide that number by 25 we get a conservative estimate of how many proctoring sessions for each computer-based test (Math & Science). Now, most schools I have worked in take all students taking a paper and pencil test and herd them into a large auditorium or cafeteria and get that test done all at once so for that we can take the number of schools in the state and multiply it by two to get the number of days it takes to proctor (assuming one session takes 1/2 a day).

Number of students......................................823,826.00
divide by 13 multiplied by 4
Number of students taking tests.................253,484.92
divide by 25 then multiply by 2
Number of proctoring sessions for
Math & Science.................................................20,287.79
Number of schools in Minnesota............................2,637
Multiply by 4
Number of Reading & Writing Test Sessions.....10,548
Add number of proctoring sessions for all tests
Total number of proctoring sessions..............30,826.79
# of proctor sessions per day per school.......................2
# of teachers needed to proctor tests............................3
Average Daily Salary + Benefits for a teacher..$327.87
Divide the total number of session by 2, multiply by 3, then multiply by cost of teachers
Cost for test proctors................................$15,160,718.28


So, it costs the state of Minnesota a total of $15,160,718.28 just for test proctors.
Now that $23 Million is more like $38 Million. Now, these funds are not ones that can be recouped but the time these teachers spend proctoring tests could be reallocated for other purposes if it was not for testing.

Of course, this is not the end of it. Since test results are somewhat difficult and cumbersome to analyze, and since these tests have such high stakes attached to them, most districts have hired someone to be their data analyst and testing coordinator. Sometimes this is a teacher position but usually it is an administrator which means it costs districts more to hire them. Lets just take a conservative estimate and say the way districts do this is 50/50. If the average teacher salary for someone near the top of the salary ladder is $60,000 (assuming that a district is not going to hire a new and inexperienced teacher for this position) and the average school administrator makes around $100,000 then half-way between is somewhere around $80,000. For sake of being conservative in our estimate lets leave out benefits for this one, which would ultimately drive the cost even higher. If there are 543 school districts in the state and each one needs a data specialist/test coordinator then that cost will be:

$80,000 X 543 = $43,440,000 annually for data specialists/testing coordinators

It is highly likely that most districts would not have this expense if it were not for the high-stakes testing required by NCLB. This means that we can add this number to the the $23 Million as extra spending.

Now both because Pearson does not release test results until well into the summer when it is hard to use them to differentiate instruction and because expensive data specialists need to justify the expense of their jobs, most school districts also use NWEA tests to get actionable data throughout the year on student achievement. The NWEA tests give immediate feedback and the subscription cost is fairly low but the tests are online, given for every grade level, and done quarterly. Which means, the actual cost to districts is much higher than the sticker price.

Assuming that every district decided to use this NWEA test the subscription alone would cost:

543 districts X $5,000 = $2,715,000 in NWEA subscriptions

Now this number is also additional spending and can be added to our running tally which is now:

$69,155,000

So, what is the cost for proctoring these tests? Lets find out:

Number of students in the state..................823,826.00
divide by 25 then multiply by 3
Total number of proctoring sessions..............98,859.12
# of proctor sessions per day per school.......................2
# of teachers needed to proctor tests............................3
Average Daily Salary + Benefits for a teacher..$327.87
Divide the total number of session by 2, multiply by 3, then multiply by cost of teachers
Cost for test proctors................................$194,476,957.38


Of course, since these tests have such high stakes attached to them most districts end up spending money on test-prep curriculum like Study Island. It is highly doubtful teachers or districts would use such a tool if it were not for the high-stakes of these tests. But, schools need students to learn how to take tests and be used to taking them online because in the NCLB world test taking is an important life skill. Lets assume that between Study Island and other test-prep curriculum districts spend money on that they average $5,000 a year for curriculum and subscription fees. Seems like a small price to pay for better test scores.

543 school districts X $5,000 = $2,715,000 in Test-Prep Curriculum

But, buying curriculum also means time devoted to implementing it. In the districts I have worked in, teachers have felt pressure to use this drill and practice curriculum that emphasizes low-level cognitive skills such as data recall and decoding. So how much of our tax dollars go to paying for this implementation?

Number of students......................................823,826.00
divide by 25
Total number of classrooms...........................32,953.04
# of days doing test prep for NWEA = 20 x 3 = 60 / 7 class periods * 4 tested subjects
# of days doing test prep for MCA = 20 / 7 class periods * 4 tested subjects
Average Daily Salary + Benefits for a teacher..$327.87
Cost to deliver test prep curriculum...................$493,909,733.02


So, total cost of ownership for these high-stakes tests is: $856 Million Dollars

and the total amount that would not be immediately reallocated toward other purposes:


What could the state do with $72 Million Dollars?



So, what kind of data-driven decision is this?

Tomorrow I will demonstrate what this means for how often kids get to use computer labs for anything but testing or practice testing.

Gapminder - Education Statistics

I spent a few minutes today playing around with Gapminder World to see if the data they had could show anything interesting regarding education. Perhaps such a powerful infographic/data visualization tool could show us clearly what factors have definitive correlations between math and literacy rates. I wish they had statistics that measured how happy people are, or data that showed the breadth of study (i.e. how many of them had access to art, music, theater, culinary arts, philosophy, or other classes). But, I'll have to do with what has been measured, which by-the-way is an interesting set of data in itself. What gets measured is a direct indicator of the values of those collecting data. This act in itself contains an important bias. It means that the values of those who do not put much faith in data and statistics are always under-represented in data and statistics. Therefore, naturally data and statistics will more likely support the values of the "data-driven" rather than those who are not.

That said, here are some of the more interesting data sets I found:


What was interesting about this graph was when you animate it there doesn't appear to be any correlation between GDP/capita and math achievement in 1995 but as time goes on over the 12 years of the data set all data sets come to convergence upon an emerging diagonal line indicating that the world is trending toward income being an indicator of math achievement level but perhaps it didn't used to be.

Another place where GDP/capita appears to have a direct correlation is with adult literacy rates as shown here:


But, the correlation is not nearly as strong when it comes to those only between the ages of 15 and 24. Odd. I wonder why this is.

While the GDP/capita data seems to show a strong correlation on math achievement and literacy rates, what seems to be a stronger indicator is fertility rates. Gapminder clearly shows an inverse relationship between math proficiency & adult literacy and the number of children per family.



When it comes to wealth distribution the results are less clear. Gapminder has many different data sets that could be used to draw this comparison but none of them are absolutely conclusive. By looking at who the outliers are in these it seems how nations acquire the majority of their wealth matters as much as how it is distributed. It seems like we could derive from these statistics that the more stable and sustainable the nation's wealth combined with how evenly distributed it is determines the level of math achievement and adult literacy.