Features of the Scoring Service API

The calls provided by the Scoring Service API are accessed via a standard XML-RPC interface.

Cloud Based Access

The Scoring Service API is a MetaMetrics-managed web service; clients using the service are not required to install or manage anything. MetaMetrics provides highly available and scalable services using Amazon Web Services (AWS) in geographically-diverse data centers.

Asynchronous Operation Recommended

The Scoring Service API has been designed to take advantage of modern computing hardware; multi-threaded operation benefits from multiple CPU cores. To enjoy these performance benefits, it is necessary that clients implement an API interface that operates in parallel. If an implementation operates serially, performance will be limited to the power of a single CPU core. To that end, MetaMetrics has provided an asynchronous operation capability. Each API call takes an optional boolean argument which when set to “True” makes the call run asynchronously. When called this way, the Scoring Service API will return a key that can then be passed to the “api_call_status” endpoint to fetch results as they are completed. Asynchronous operation is recommended in all cases where non-real-time operation is appropriate or high volume simultaneous requests will be made.

Multiple Call Types Supported

There are two different types of XML-RPC calls supported by the Scoring Service. The first type provides a student measure for an assessment when there is no prior information available about the student (i.e. measures, uncertainties). The second group of calls provides a student measure that accounts for a student’s previous performance using Bayesian scoring.

Call Information

When a student is administered an assessment that can report Lexile reading or Quantile mathematics measures, estimates of the student’s ability (Lexile reading/Quantile mathematics measure) must be calculated. An estimate of a student’s ability is comprised of an ability measure (i.e. Lexile reading or Quantile mathematics measure) and an associated uncertainty (also known as sigma). If prior ability information about the student is not available, the student’s performance on the singular assessment can be used to estimate a measure and associated uncertainty.

When a student's prior ability measure and associated uncertainty are available and used with a Bayesian scoring function call, they represent an estimate of ability given all of the student’s previous testing experiences and are referred to as priors. When this prior student data is available, the Scoring Service API can provide a more-accurate estimate of a student’s current ability by using Bayesian scoring (Yen and Fitzpatrick, 2006). The Scoring Service API will need the student’s last ability information, or priors, as well as how long ago the priors were calculated. Bayesian scoring uses the time elapsed since the last test or age of the estimate to determine how much additional uncertainty should be associated with the current estimates. The Scoring Service API will return an updated estimate of the student’s ability (Lexile reading/Quantile mathematics measure) and its associated uncertainty.

The Scoring Service API supports seven different function calls.