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