Meritocracy or reputation? The role of rankings in the sorting of international students across universities

Figures & data

Figure 1. Average marginal effects predicting standardised international students.

Table A1. Descriptive statistics.

Interval variables	Mean	Standard deviation	Min	Max
International Students	2732.43	2168.02	60	14560
Total Students	18869.63	7810.23	3845	41180
% Change in Ranking (2010-2018)	11.19	52.20	−86.60	246.15
% EU Students	5.97	3.82	.66	19.63
% International Research Funds	24.19	12.70	0	70.00
% Expenditure on Facilities	5.61	4.77	0	28.38
% Tuition Fees Income	52.66	19.25	9.81	86.10
Binary Variables	%
Reputation	18.14

Notes: Each interval indicator has been standardised by subtracting its average value from the absolute value and dividing the result by its standard deviation. Values are rounded at the second decimal.

Table A2. Random effects model with lagged independent variables predicting the number of international students.

Variables	Model 1	Model 2
Reputation	.605*** (.075)	.246*** (.050)
% Change in Ranking	-.023 (.048)	-.047^+ (.025)
% EU Students (T-1)	.169*** (.044)	-.081* (.039)
% International Research Funds (T-1)	.007 (.019)	-.010 (.022)
% Expenditure on Facilities (T-1)	.013 (.027)	-.023 (.037)
% Tuition Fees Income (T-1)	.096*(.038)	-.032 (.044)
Total Students (T-1)	.243*** (.062)	.054 (.035)
International Students (T-1)		.589*** (.063)
Wald Χ^2	123.82***	435.27***
N	258	258
R^2 (within universities)	.09	.15
R^2 (between universities)	.67	.88

Notes: ⁺p<.10; *p<.05, **p<.01, ***p<.001.

Robust Standard Errors in the Parentheses.

Coefficients and Robust Standard Errors Rounded at the Third Decimal.

The model controls for dummy variables for years.