Technical and statistical information

Technical reports

The ICSEA and SEA quarters calculation model is subject to a process of continuous refinement and enhancement. The latest model and procedures are explained in full details in the ICSEA 2013 technical report.

 

ICSEA 2017 technical report (348 kb)

ICSEA 2016 technical report (345 kb)

ICSEA 2015 technical report (312 kb)

ICSEA 2014 technical report (429 kb)

ICSEA 2013 technical report (877 kb)

My School student gain technical report (282 kb)

Indicative standard errors

The NAPLAN average score (or mean score for a school) and the NAPLAN median score are estimates of a school’s performance based on the students who were tested in the school. Those estimates are more accurate when the number of students tested is large and when the students perform at a similar level. Standard errors reflect the accuracy of the estimates and the degree of confidence one can have in the estimates.

The PDF documents below provide indicative standard errors of NAPLAN school performance scores for both the average and the median with different numbers of tested students. They provide an indication of the standard error for a school rather than the actual value for each school. These are used to derive the margin of error that is shown on the graphical depiction of NAPLAN results.

An indicative error is an estimated error associated with a number of students in a school's test group, where the estimate comes from a regression analysis.

In the case of the average, the prediction is derived by regressing standard errors on a number of students in a test group.

In the case of the median, a standard error is first calculated using a process called 'bootstrapping', which recalculates the median for each school 1,000 times using 'resampled data with replacement', and takes the distribution of the resulting medians to calculate a standard error.

'Resampled data with replacement' means that the same number of student records is sampled each time, except that the same student record may be sampled more than once. With more than 1,000 replications, this produces a highly accurate estimate of the standard error of the median.

That bootstrapped standard error is then modelled, as was the standard error of the average, and the predicted values from that modelling are the indicative errors.

Standard errors for medians tend to be slightly larger than standard errors for averages.

Caveats to accompany data collections

In some instances, data fields may vary in their treatment across states and territories. Where appropriate, caveats are collected to accompany data submissions so that how jurisdictions treat data fields can be captured and reported. Please follow links below to where these are.

 

Caveats for senior secondary outcomes

Caveats for VET in schools

Caveats for Finance (Financial limitations)

Financial data reporting on My School

When collecting financial information from jurisdictions, data fields can vary in their treatment across states, territories and school sectors. Therefore financial limitations are also collected to accompany financial data submissions so that these differences can be captured and reported. See financial limitations for all jurisdictions from 2009 - 2016.

 

At the end of each year's finance data collection ACARA release a Financial data reporting methodology report, and Deloitte release a letter of accounting advice and an independent assurance practitioners compliance report. See financial reports and letters for 2009 - 2016.

 

Lastly, historical independent school finance data should be read in conjunction with the financial reporting for non-government multi-campus schools.