Minimum information for a table

When thinking about a table of information, it might be helpful to consider the following elements as a minimum.
  • Link/s - How will this information be linked to other information?
  • Time - When does this action take place?
  • Type - Are there different types or categories for the action?
  • How much - Are there are differences in how much something is applied?
  • What quality - Are there differences in how well this this action is applied?

Linking fields

We should never create a table of information that cannot be linked to other information. Typical linking fields include;
  • Cert Number
  • Student ID
  • Social Security Number
  • Federal Institution Code
  • WA District Code
Machines are literal.  For instance, they will even see " Math" and "Math" as two different objects.  Misspellings, capital letters, odd characters, and even white spaces are challenges for maintaining machine readable information.  This is one of the reasons WA uses "123456A" rather than "John Smith" to identify a specific individual.  We should  reuse established and reliable codes whenever possible.

Time fields

People tend to consume information as a snapshot in time, i.e. "How many teachers graduated last year?" or "What are their latest test scores?"  Information systems report on this information by tracking the time element within the table, which might include;
  • Academic Year
  • Calendar Year
  • Semester/Quarter
  • Date
We like to consume information in reports like this;

Completions in 2010

John   Elementary Education


Completions in 2011

Dave   Biology


However, the underlying table would collect the information like this;
 YearName  Program
 2010John   Elementary Education
 2011Dave Biology

Reports generated from the table might look like this;
"Select the name and program for all people completing in 2010"
 YearName  Program
 2010John   Elementary Education


Type

People tend to think of information at the person level, such as;
"John received 80 hours of pre-service field experience and 150 hours of student teaching."  
Person Pre-service Student Teaching 
John 80 150 

Information systems don't care about the person, instead they might say;
"There were two actions taken, 80 hours of pre-service attributed to John and 150 hours of student teaching attributed to John."
 PersonType 
John Pre-Service  80
John Student Teaching  150

A report from the data base might be;
"SELECT John's field experiences and PIVOT by type", which would yield the following extract;
Person Pre-service Student Teaching 
John80 150 

You could also ask;
"Select John and the SUM of his field experiences"
Person SUM(N) 
John 230 

The type field allows for more complex information to be collected using a simple table structure.

"How Much"

Use "How Much" for the less important units of measurement, such as;
  • How much time was spent doing this?
  • How many credits were received?
  • How long did it take?
If there are differences between practices (or dosages) among individuals within a program, then this information should be collected.  "How Much" is not intended to paint a full picture, it is only collected to add important context to an action.  The "How Much" elements are the measurable elements that are often deemed too simplistic to be of much equative use.

"What Quality"

"How Much" and "What Quality" are two sides of the same coin, the difference is how people tend to react to an element. While "How Much" is usually considered too simple to be of much consequence, "What Quality" is surmised to be too complicated to be reduced to a quantifiable measurement.  Here are some typical statements that let me know when we've moved from "How Much" to "What Quality";
  • "This area of interest is made of complex interactions that can't ever be truly known."
  • "This item will only explain a portion of what is known, which will yield incomplete and incorrect answers."
  • "The item is ensconced within something unique, which means that even if it can be measured, it can not ever really be contextualized."
"What Quality" is the brass tacks of a information system, but it is also the most challenging to navigate.  For instance, in the WA Teacher Evaluation system districts are encouraged to use any system of measuring quality they wish, as long as it has X number of levels.  This compromise is in place because teachers and principals believe quality teaching is a complex and inherently personal element that is difficult, if not impossible to measure or track.  The achilles heel of this argument is that it usually makes less sense when the perspective changes.  For instance, while it might be amiable for each district to have its own method of measuring teaching quality, it is more difficult to make the case that quality teaching is fundamentally different between teachers and children in Washington versus North Carolina.