Skip to main content

Table 2 Results of logistic regression modelling

From: Factors associated with junior doctors’ decisions to apply for general practice training programmes in the UK: secondary analysis of data from the UKMED project

   Outcome 1 Outcome 2
   Applied to GP training Applied solely to GP training
   Model 1
Sample A
All medical school entrants
Model 2
Sample B
Non-graduate entrants only
Model 3
Sample A
All medical school entrants
Model 4
Sample B
Non-graduate entrants only
Predictor df χ2 P value χ2 P value χ2 P value χ2 P value
Sex 1 76.37 0.0000 66.71 0.0000 43.87 0.0000 37.04 0.0000
BME 1 31.81 0.0000 22.92 0.0000 n/s n/s 9.05 0.0026
NS-SEC 4 n/s n/s    13.27 0.0100   
IMD 4    10.95 0.0272     
POLAR2 2         
Secondary school type 1        3.66 0.0558
UK secondary educated 1 16.43 0.0000 n/s n/s 8.10 0.0044   
Total UKCAT z-score 1    16.87 0.0000    5.44 0.0197
Total HESA tariff z-score 1         
Medical school entry status 2 9.73 0.0077    6.25 0.0440   
Medical school 32 91.56 0.0000 83.18 0.0000 79.78 0.0000 76.17 0.0000
Intercalated 1 36.03 0.0000 40.89 0.0000 17.63 0.0000 22.61 0.0000
EPM top two quartiles 1 6.85 0.0088 4.50 0.0339 n/s n/s   
Foundation school 26 n/s n/s 44.43 0.0136 47.88 0.0056 45.07 0.0116
ARCP satisfactory progression 1 5.07 0.0243 n/s n/s n/s n/s   
Model statistics
 Minimum required sample size 234 232 392 303
 Actual sample size 6177 4327 6177 4375
 Mean probability (95% CI) 0.427 (0.416 to 0.440) 0.431 (0.417 to 0.445) 0.255 (0.245 to 0.266) 0.264 (0.256 to 0.281)
 Hosmer–Lemeshow test χ2(8) = 10.03
P = 0.263
χ2(8) = 7.43
P = 0.491
χ2(8) = 6.09
P = 0.637
χ2(8) = 10.45
P = 0.235
 Area under ROC curve 0.6421 0.6698 0.6394 0.6603
  1. Blank cells indicate non-significant predictor in preliminary model and excluded from final model
  2. ARCP Annual Review of Clinical Progression, BME Black and minority ethnic, CI confidence interval, EPM Educational Performance Measure, HESA Higher Education Statistics Agency, IMD Index of Multiple Deprivation, n/s non-significant predictor in reported final model, NS-SEC National Statistics Socio-economic Classification, POLAR2 Participation Of Local Areas classification, ROC receiver operating characteristic, UKCAT UK Clinical Aptitude Test