Table 4

Comparison of the four outcome groups
Fail (a) Repeat 1st Passed after Passed all ANOVA (r) Linear ANOVA (r) Levene test Homogenous
year (b) re-sits (c) first time (d) trend F(1,n) Nonlinear F(2,n) subsets (s)
Overall Mark (p, q); All cases −2.644 −1.924 −1.110 .235 2843.8 36.9 P ≤.001 a,b,c,d
(1.28, 96) (.99, 94) (.79, 565) (.80, 4,056) P <.001 P <.001
Theory Mark (p) −1.258 −1.322 -.654 .250 619.9 23.1 NS ab, c, d
(.44, 29) (.82, 29) (.76, 294) (.68, 1,723) P <.001 P <.001
Skills Mark (p) −1.079 -.891 -.616 .214 602.8 32.0 P ≤.001 ab, bc, d
(.84,40) (.87, 438) (.90, 438) (.71, 2,655) P <.001 P <.001
Totoal UKCAT score (UKCATtotal) 2492 2457 2486 2544 43.7 6.8 NS abc, ad
(192, 96) (230,94) (205, 565) (205,4,056) P <.001 P = .001
zUKCAT (p) (zUKCATtotal) -.121 -.312 -.186 .036 25.3 5.3 NS abc, ad
(.95,96) (1.01, 94) (.99, 565) (.99, 4,056) P <.001 P = .005
UKCAT abstract reasoning (p) (zUKCATabstractReasoning) -.163 -.224 -.096 .022 13.1 0.75 NS abcd
(.94, 96) (1.02, 94) (.97, 565) (1.00, 4,056) P <.001 NS
UKCAT decision analysis (p) (zUKCATdecisionAnalysis) -.064 -.302 -.129 .026 14.01 3.70 NS abc, ad
(1.08, 96) (.99, 94) (.98, 565) (.995, 4,056) P <.001 P = .025
UKCAT quantitative reasoning (p) (zUKCATquantitativeReasoning) .045 -.110 -.116 .018 3.38 3.38 NS abcd
(.97, 96) (1.03, 94) (1.06, 565) (.99, 4,056) NS P = .034
UKCAT verbal reasoning (p) (zUKCATverbalReasoning) -.087 -.136 -.127 .023 9.56 2.13 NS abcd
(.97, 96) (1.16, 94) (.96, 565) (1.00, 4,056) P = .002 NS
Educational attainment (p) (zEducationalAttainment) -.441 -.653 -.563 .104 156.1 29.5 P ≤.001 abc, d
(.942, 65) (1.06, 60) (1.14, 414) (.94, 2,893) P <.001 P <.001
Three best A-levels (Alevels_TotalBest) 28.89 28.28 29.06 29.39 44.8 6.47 P ≤.001 a,bc,d
(2.82, 56) (1.58, 49) (1.36, 333) (1.22, 2,326) P <.001 P = .002
Five best highers (SQAhighers_TotalBest) 48.71 47.50 48.70 48.99 3.8 1.66 NS abcd
(1.57, 17) (2.58, 16) (2.46, 88) (2.40, 648) NS NS

Notes:

P-values are standardized within schools (Note: this is not the case for the raw, UKCAT total mark)

q. In a small proportion of cases, as described in the text, the overall mark is based on a normal score derived from the four-point categorical scale. For comparability with other analyses, the first row includes these cases. However, the second row analyses only cases where an overall mark was explicitly provided.

r. The denominator df, n, can be calculated as N-4, N is the total number of cases (provided in individual cells).

s. If values are together then they are not significantly different from one another and form a homogenous subset with P >.05 using the Ryan-Einot-Gabriel-Welsch range test. As an example, “abc, ad” means that groups a, b and c (Fail, repeat first year and passed after re-sits) do not differ from one another; likewise groups a and d (Fail, Passed all first time) do not differ from one another. Group d (Passed all first time) is significantly different from Repeat first year and Passed after re-sits. “a, b, c, d” indicates each differs from each of the other three groups, and “abcd” indicates no significant post hoc differences.

McManus et al.

McManus et al. BMC Medicine 2013 11:244   doi:10.1186/1741-7015-11-244

Open Data