1
Non Progression Among Higher Education New Entrants:
A Multivariate Analysis
Dr Selina McCoy, ESRI
Dr Delma Byrne, NUIM
28 October 2010
2
Moving to a Value-Added Approach
● Crude/overall patterns of non progression● Doesn’t take account of differences in student
intake across institutions and sectors● Need for like-for-like comparison● Reduces risk of creating incentives for greater
student selection● Does not negatively label institutions with more
diverse student intakes● Focus is on institutional effectiveness, taking
account of student intake
3
Two main questions:
Which students are most likely to progress?
Taking account of these characteristics, does the average chance of progression vary across institutions?
4
Student Characteristics● Gender ● Social Class Background (Father’s Position)● Educational Attainment at Second Level
● Leaving Certificate Points ● Attainment in Mathematics, English and Irish
● Nationality ● Grant Recipient ● Field of Study ● Course Level (Level 6,7 & 8)● Institution
5
Information we don’t have
Motivation for enrolling in HE Financial well-being Participation in part-time employment Academic engagement Views on teaching staff, educational experience Attendance, participation in extra-curricular activities Institutional supports for students
6
Approach STATA Takes account of clustering – students in the same
institution share common influences, may be more like each other
Overall (unadjusted) differences in progression chances across institutions
Net differences – do the chances of progression vary across institutions
Results presented in odds ratios: chances of group not progressing relative to reference group
7
Findings 1: Characteristics of Students Who Do Not Progress/Not Present
– Model 1: Gender, Age, Nationality, Social Class
– Model 2: Leaving Certificate Points and Receipt of Grant
– Model 3: HE Sector
– Model 4: Field of Study, NFQ Level
8
Non Progression and Gender
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Model 1: Age, Nationality,Class
Model 2: LC Performance &Grant
Model 3: Sector Model 4: FOS, Level
Male
9
Non Progression and Social Class
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Professional Manager Non Manual Skilled Manual UnskilledManual
Class Unknown
Model 1: Gender, Age, Nationality
Ref: semi-skilled manual
10
Non Progression and Social Class
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Professional Manager Non Manual Skilled Manual UnskilledManual
Class Unknown
Model 1: Gender, Age, Nationality Model 2: LC Performance & Grant
Ref: semi-skilled manual
11
Non Progression and Social Class
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Professional Manager Non Manual Skilled Manual UnskilledManual
Class Unknown
Model 1: Gender, Age, Nationality Model 2: LC Performance & Grant
Model 3: Sector Model 4: FOS, Level
Ref: semi-skilled manual
12
Non Progression and LC Performance
0
0.5
1
1.5
2
2.5
3
0-150 155-200 205-250 255-300 305-350 355-400 405-450 455-500 505-550 555-600 Unknown
Model 2: Gender, Age, Nationality, Class, Grant
13
Non Progression and LC Performance
0
0.5
1
1.5
2
2.5
3
0-150 155-200 205-250 255-300 305-350 355-400 405-450 455-500 505-550 555-600 Unknown
Model 2: Gender, Age, Nationality, Class, Grant Model 3: Sector Model 4: FOS, Level
14
Non Progression and Field of Study
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Education Science, Agri,Vet
ComputerScience
Engineering Construction Healthcare Services Combined &Other
Model 4: Gender, Age, Nationality, Class, LC Performance, Grant, Course Level
Ref: Social Science, Arts, Law
15
Non Progression and Sector
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Institute of Technology Other HE College
Ref: University Sector
16
2. Non Progression Across Institutions
Model 1 – All Individual (named) HE Institutions (ref: UCC)
Model 2 – Gender, Age, Nationality, Class Model 3 – LC Performance, Grant Model 4 – Field of Study, Course Level
17
Non Progression Across Institutions
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Model 1
Ref: UCC
18
Non Progression Across Institutions
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Model 1 Model 4: Gender, Age, Nationality, Class, LC Performance, Grant, FOS, Course Level
Ref: UCC
19
3. Non Progression Within Institute of Technology Sector
Model 1: All Individual (named) IoTs (ref: Blanchardstown)
Model 2 – Gender, Age, Nationality, Class Model 3 – LC Performance, Grant Model 4 – Field of Study, Course Level
Level 6 & 7
20
Non Progression Within the IoT Sector
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Model 1
Ref: Blanchardstown IT
21
Non Progression Within the IoT Sector
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Model 1 Model 2: Gender, Age, Nationality, Class
Model 3: LC Performance, Grant Model 4: FOS & Course Level
Ref: Blanchardstown IT
22
Non Progression Within the IoT Sector
No real gender differences Some age differences Irish students higher non progression Grant impact Some class difference (skilled manual) LC Performance – impact of low performance, high
performers do better Clear Field of Study patterns
– Computer Science higher, Healthcare lower than Science, Agriculture & Veterinary students
23
4. Non Progression Within the University Sector
Model 1: All Individual (named) Universities (ref: UCC)
Model 2 – Gender, Age, Nationality, Class Model 3 – LC Performance, Grant Model 4 – Field of Study
All Level 8
24
Non Progression Within the University Sector
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
DCU UCD NUIG UL NUIM TCD
Model 1 Model 2: Gender, Age, Nationality, Class Model 3: LC Performance, Grant Model 4: FOS
Ref: UCC
25
Non Progression Within University Sector
No gender differences No age differences No nationality difference No grant difference Some class difference (managerial v non manual ) Clear LC Performance patterns Clear Field of Study patterns
– Computer Science higher, Education & Healthcare lower than social science, law & arts students
26
SummaryWhich students fare well?
– LC Performance strong predictor of progression– Maths performance particularly strong influence– Importance of grant support, particularly for IOT
students– Strong disparities across subject areas and fields
– No gender differences– Delayed entry not significant– Class differences are small, and largely operate
through LC performance
27
Summary
Do the average chances of progression vary across institutions?– Wide raw differences shrink dramatically when
taking account of student intake– Dangers of crude league tables– Main differences between sectors – Colleges of
Education and NCAD doing very well
28
Summary
Need to unpack the processes underlying institutional effectiveness: what can institutions do?– Pre-entry guidance, informed choices– Academic supports in 1st year– Attention to particular FOS– Financial wellbeing– At risk students– Broader student engagement