ARM scientists collect and provide data that can be used to increase the accuracy of models describing atmospheric clouds, radiative transfer, and air quality. The ARM Data Center distributes this information while offering tools for visualizing and analyzing it.
The ARM Climate Modeling Data Center (MCDDC) offers scientific computing services, including access to high-performance computers and software for data processing and visualization as well as expertise in model development. MCDDC serves researchers from around the US and internationally.
To provide researchers with data sets of sufficient dimensionality and quality for research purposes, the MCDDC has devised various methodologies to reduce data size. One such technique, known as resampling, converts raw data into formats compatible with high-performance computing systems – making data archived at MCDDC available even after projects that initially acquired it have concluded.
Student Growth Percentiles (SGPs) are measurements of relative progress that compare students’ current assessment results with those of academic peers who share similar MCAS scaled score histories. SGPs can be determined using up to two years’ worth of historical MCAS test scores as input. Academic peers of any given student can be identified from all students statewide who have taken at least one previous MCAS assessment in the same subject and grade. This student population encompasses all demographic groups and educational programs such as sheltered English immersion and special education. A statistical procedure called quantile regression is then employed to place academic peers’ performances onto a normative scale and ascertain their percentile rank within current test scores.
Each year, SGPs are calculated by comparing student assessment results against those of their academic peers and repeating this process for every statewide test administered during that year. As a result, differences between students should be interpreted with care as they may be misleading.
Student A surpassed 75% of his or her academic peers in one subject area in 2024, whereas Student B scored within 55% – possibly due to having more recent MCAS scaled score history in 2024 than Student A.
The SGPdata package included with the data sgp function provides four sample data sets that can be utilized in analyses of student growth. Of these sets, sgpData defines data in WIDE format which is utilized by lower level SGP functions studentGrowthPercentiles and studentGrowthProjections; two other datasets (sgptData_LONG and sgpdData_LONG) use LONG formats that are employed by higher-level wrapper functions like abcSGP prepareSGP or analyzeSGP; finally there is also one last data set named sgpdData_INSTRUCTOR_NUMBER which serves to produce teacher level aggregates.