FE Success Rates Methodology 2004/05
1 This definition covers the production of learner success, achievement and retention data from 2002/03 to 2004/05 that became operational in November 2005.
Purpose
2 Learner outcome data used by the LSC as part of Performance Review and the data used by OFSTED during inspections focuses on success, achievement and retention rates extracted from the individualised student record (ISR) and individualised learner record (ILR).
3 This methodology shows how to derive the qualification level variables relating to learning aims success, achievement and retention used in the inspection and summary spreadsheets and in the institution level benchmarking data produced by the LSC. It also shows how to derive the learner characteristic variables known as the demographic data.
4 The data can be aggregated in different ways to produce the following:
For all institutions –
- Benchmarking data primary tables, which show national averages of success, achievement and retention rates by institution type and notional level
For FE colleges –
Inspectorate spreadsheets which form a part of the college performance report (CPR) produced by the Office for Standards in Education (OFSTED)
Supporting data which shows results by length, qualification type, level, sector subject area and age
National benchmarking data for individual qualifications by sector subject area
Demographic data showing national averages of success, achievement and retention by learner characteristics such as age, gender, ethnicity and disability
Relevant Collections
5 The method is run on data collected in the final data collection (F05) for each of the following years –
ILR (FE) 2004/05
ILR (FE) 2003/04
ILR (FE) 2002/03
ISR25
ISR22
Source Data
6 The method uses a number of datasets for source data. These are historically known as 'tri files' and consist of data from the collections above pre-processed.
Output Datasets and Derived Variables
7 A number of 'measures' are derived including the number of learning aims started, completed and achieved, along with those achieved at high grades. These are stored in the following datasets
Detailed definition
8 The format and content of the data collected on learners has changed from year to year. The pre-processing of the ILR and ISR collections to form the 'tri-files' is partly to ensure that they are of comparable format to that of the latest year. The detail for this is contained within the tri-file definition and not repeated here.
9 There is a certain amount of functionality within the code for the FE Success Rates Methodology that is concerned with handling exceptions, inconsistencies and peculiarities within the data. The details of this are not covered here but may be seen within the sample code.
Initial processing of tri-files
10 For each tri-file, learning aims information is matched in from the hierarchy table. The hierarchy file is an output from the Learning Aim Database (LAD) that includes a grouping variable map code and other data fields from the LAD for each learning aim.
11 For each 'tri-file' the actual start year, expected end year and actual end year is calculated for each learning aim.
Merge all five ISR/ILR tri-files together.
12 The separate files are merged such that any entries for each year where the entry appears in more than one file for the same learner at the same institution with the same aim, then the latest year's entry is the only one kept. This is to ensure that where a learning aim spans several years then it is counted only once.
Drop cases that are not required for analysis.
13 Entries are dropped where not required for analysis. There are a number of reasons for this, however the predominant ones are
where the expected end year is not one of the last three years of the sequence ie 2004/05, 2003/04 or 2002/03
where a learner withdraws from the aim before 1 November of the first year
Produce a series of categorical derived variables
14 A number of derived variables are calculated and stored in the output datasets. These variables have values that categorised or banded.
a The learner's age at 31 August of the year in which the started is used to categorise into various age groups.
b The expected duration of the learning programme is used to categorise into various duration bands.
c The qualification type is used to categorize the aim in a number of different ways based on both level and duration.
Create a series of scaled derived variables
15 There are a number of measures that are required to calculate success rates.
a The number of learning aims started
b The number transferred to another learning aim
c The number that have achieved the learning aim
d The number that have completed the learning aim
e The number that are continuing the learning aim
f The number where there is a known outcome for the learning aim
g The number of high grade results
h The number of full level 2 qualifications in their own right
i The number of full level 3 qualifications in their own right
Merge in Skills for Life and Basic Skills markers
16 Skills for Life and Basic Skills markers are merged in from the Basic Skills Lookup Table.
Remove non LSC funded learner cohorts
17 Success Rates are calculated for aims where there is at least one learning aim in a cohort that is funded by the LSC. A cohort is generally a group of learners with the same aim, the same duration category, expected to end in the same year and provided by the same institution. The reason for this inclusion of some non-LSC funded learners is that OFSTED inspections cover all aims where there are any LSC funded learners.
18 Records are removed for learning aims where the whole cohort is not LSC funded. This means that some non-LSC funded learners are included in success rates.
Aggregate to create aim master file.
19 The data is aggregated to produce a source file for further analysis by learning aim characteristics at institution level.
20 The institution and LSC office details are then matched in from the institution lookup file and saved as the Success rate Aim Master File.
Aggregate to create demographic master file.
21 Ethnicity is re-coded to the categories used for the Individualised Student Record (ISR) to enable comparison between years.
22 The data is aggregated to produce a source file for further analysis by demographics institution level.
23 The institution and LSC office details are then matched in from the institution lookup file and saved as the Success Rate Demographics Master File.
Sample Code
24 The following sample code is available
Last Modified: 23 Dec 05
