Student retention at university The Social Market Foundation 1 Why - PowerPoint PPT Presentation

Student retention at university The Social Market Foundation 1

Why worry about student retention?  Participation rates at university have grown remarkably in past decades  There has been significant attention on widening the pool of applicants to and enrolments in university, but less focus on whether those that attend stay on  Teaching Excellence Framework may raise this up the agenda.  Those from specific demographic groups may be more likely to drop out. This is an important social mobility agenda. 2

Regional dropout rates over time Proportion of UK-domiciled students that drop out of higher education by end of first year 10% 9% EAST MIDLANDS EASTERN 8% NORTH EAST NORTH WEST SOUTH EAST 7% SOUTH WEST WEST MIDLANDS 6% YORKSHIRE AND HUMBERSIDE GREATER LONDON 5% TOTAL 4% 11/12 12/13 13/14 14/15 • The difference in dropout rates between London and the rest of the country is much less pronounced than two years ago. • However, the region is still performing poorly. 3

Significant differences between younger and older students Proportion of UK-domiciled students that drop out of higher education by end of first year 13.3% 13.1% 14% 11.9% 11.8% 11.3% 11.1% 12% 10.3% 10.1% 9.9% 10% 7.8% 7.7% 8% 6.7% 6.5% 6.2% 5.9% Young Dropout 5.1% 6% 5.0% 4.6% Mature Dropout 4% 2% 0% All cross- sections from 2014/15 • Mature first-degree entrants have significantly worse retention rates than their young counterparts in all regions of England. Mature is defined by HESA as aged 21 or over on September 30 th on the year of entrance. • There is imperfect correlation between the retention rates of the two groups regionally. The factors influencing their dropout rates are likely to be different. • Our analysis focuses premodinantly on young students. 4

Principal retention factors A sense of belonging  Building relationships with other students and academic staff; feeling that the institution is suited to your needs Engagement  Participating in lectures, tutorials and extra-curricular groups Financial Constraints  Being able to afford the costs of living, the costs of study material and of socialising and participating in university life Examples of how such factors may be manifested  Those from poorer backgrounds, or ethnic minorities, may find it more difficult to engage with their studies (differences in culture, attitudes).  Poorer students may also find it difficult to afford the necessaries.  Those living further from university, and especially not in halls or on campus, may find it more difficult to engage in student life generally. 5

Socio-economic background Relationship between socio-economic score and drop-out rates 70 • NS-SEC 4-7 is a measure of whether a 60 student’s parents are / were in specific NS-SEC 4-7 (%) 50 occupations (including routine jobs, lower 40 supervisory roles and small employers) 30 indicative of a lower socioeconomic class. 20 There is high correlation between the 10 proportion of student’s who fall into this 0 group at a university and dropout rates. 0 5 10 15 20 Dropout Rates (%) Socioeconomic background by region (NS-SEC 4-7) • London’s rate of students 45% 40% whose parents are classed as 35% NS-SEC 4-7 is around the 30% median for universities. The 25% capital’s figure (35%) is much 20% 15% lower than the West Midlands’ 10% 41%. 5% 0% 6

Prior attainment of students Relationship between UCAS score and drop-out rates • Evidence shows a negative correlation between prior attainment and dropout rates. Average UCAS score by region YORKSHIRE AND HUMBERSIDE • London has one of the highest average WEST MIDLANDS UCAS tariffs SOUTH WEST SOUTH EAST • But, data on north east suggests that NORTH WEST prior attainment is unlikely to be NORTH EAST attributing to the regional differences GREATER LONDON EASTERN EAST MIDLANDS 310 320 330 340 350 360 370 380 390 Average UCAS points 7

The importance of ethnicity Relationship between ethnicity and drop-out rates  Certain ethnic groups have higher 35% drop out rates than others. Given 30% the differences in regional ethnicity, Black Students (%) 25% this could explain regional differences. 20% 15%  Black students are more likely to 10% drop out than other ethnic groups. 5% 0% 0.0% 5.0% 10.0% 15.0% 20.0% Dropout (%) Regional Ethnicity Breakdown 16%  London has by far the highest proportion 14% 12% of Black students. Other regions with a Ethnicity (%) 10% high proportion also have high dropout 8% rates: Eastern England and the West 6% Midlands come third and fourth bottom 4% 2% for retention. Black 0% Asian  HESA provides only two minority divisions when reporting institutional-level dropout rates: Asian and Black. We know that retention levels among different Asian groups varies significantly. 8

Does being on a campus make a difference? University non-continuation rates: campus versus non-campus Non-campus Campus Universities MEAN 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% Dropout rates  We find that, overall, students at campus universities have a non-continuation rate of 5.9%, considerably lower than the average of 6.8% for non-campus students.  In regions outside of the North and London, the population of campus universities have at least a one percentage point advantage over non-campus universities.  In London and the North this picture is reversed – drop out rates for campus universities are higher than for non-campus universities.  When looking at campus retention rates by university tariff type the pattern remains for low and high tariff institutions.  Medium tariff campus universities have higher dropout rates than their non campus counterparts 9

Student satisfaction Relationship between student satisfaction and drop out rates: 100 95 Proportion 'satisfied' NSS 90 LONDON ELSEWHERE 85 Linear (LONDON) Linear (ELSEWHERE) 80 75 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% Average proportion satisfied by region Institutional dropout rate 89 88 87 86  On average London has the lowest 85 84 proportion of satisfied students based on 83 figures from the NSS 82  However, the trend line is flatter within 81 London suggesting less of an influential effect 10

Student Living – Living at home Relationship between living at home and drop out rates 35% 9% 8% 30% 7% 25% 6% 20% 5% Proportion living at 4% 15% home 3% 10% 2% Young Dropout 5% 1% 0% 0%  An important metric of student living is the proportion of students living at home. Our hypothesis was that students living at home would have higher dropout rates, given the increased difficulty of engaging socially and academically in the student experience. It is possible that this is also be a product of cultural factors and / or socio-economic factors.  The data looks to support this hypothesis, given the obvious correlation over regions. 11

Student Living – commuting distances to study Relationship between commuting distance to university and drop out rates 9 9% 8 8% 7 7% 6 6% Commute to university 5 5% (miles) 4 4% Young Dropout 3 3% 2 2% 1 1% 0 0%  Commuting distance, may capture the effects of those living at home being less engaged, and tending to live further away.  However, it may also represent an isolated effect of living far away limiting the amount of time that may be spent in university, whether studying or otherwise engaging in the experience. 12

Institutional cost of living Average cheapest accomodation cost by region • London universities have the YORKSHIRE AND HUMBERSIDE highest average cheapest room WEST MIDLANDS costs available to students. SOUTH WEST SOUTH EAST • There is difficulties in NORTH WEST representing the regional value NORTH EAST as institutions will have varying GREATER LONDON accommodation policies. EASTERN EAST MIDLANDS 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 • At an institutional level there appears to be no correlation between the cheapest accommodation costs and the dropout rate. 13

Regression analysis To understand better whether there is a ‘London effect’ present we conducted a regression analysis that controlled for a range of factors that we believe could influence student retention. Variable Coefficient P-value (*sgf) The first thing to note when looking at these results is that whilst we have a London % of Black students 0.139* 0.007 effect present , this result does not hold when we do not control for POLAR3. This is due to % of Asian students -0.116* 0.036 the low level of POLAR3 scores amongst % Female -0.407* 0.061 London institutions. The London effect is % NS-SEC 4-7 0.763* 0.000 unlikely to be robust. % POLAR3 0.299* 0.000 However, the results add substantial evidence Campus dummy -0.005 0.938 to the debate around the importance of student characteristics. A significant amount UCAS -0.001* 0.003 of variation in institutional student dropout % Satisfied (NSS) -1.184* 0.057 rates is attributed to the characteristics of their student population. University population 5.71e-07 0.877 Some of the variables that are insignificant at Distance moved by 0.002 0.317 an institutional level may be of importance to region certain subgroups of students. London dummy 0.262* 0.008 * Statistically significant 14