Skip to main content

Table 3 Summary of observations from the two-treatment case

From: Investigating the relationship between predictability and imbalance in minimisation: a simulation study

 

Predictability at P= 1.0, %

Na

Recommended P- value

Predictability at this P- value, %

Reduction in predictability, %

P-value at which imbalance occurs

One variable

Two categories

50.0

<90

0.8b

45.0 to 48.8

1.2 to 5.0

0.7

  

90 to <200

0.7b

49.0

1.0

0.6

  

≥200

0.7b

49.5 to 49.9

<1.0

0.5

Two categories - unequal prevalence

 

< 80

0.9b

45.0 to 48.8

1.4 to 5.0

0.8

  

80 to <200

0.8b

49.0

<1.0

0.7

  

≥200

0.7b

49.0 to 49.9

<1.0

0.5

Three categories

 

<90

0.9b

45.0 to 48.8

1.2 to 5.0

0.8

  

90 to <200

0.8b

49.0

1.0 to 1.1

0.7

  

200 and <300

0.7b

49.0

1.0

0.6

  

≥300

0.7b

49.3 to 49.6

<1.0

0.5

Four categories

 

≤100

0.9b

40.0 to 49.0

1.0 to 5.0

0.8

  

200 and <400

0.7b

49.0 to 49.3

0.7 to 1.0

0.6

  

≥400

0.7b

49.5 to 49.6

<1.0

0.5

Two variables

Both with 2 categories

72.0

100

0.8

63.0

9.0

0.7

  

200

0.7

60.0

12.0

0.6

  

≥300

0.7

60.0

12.0

0.5

Both with 2 categories - unequal prevalence

 

<100

1.0

-

-

0.9

  

100

0.9

68.0

4.0

0.8

  

200

0.7

60.0

12.0

0.6

  

≥300

0.7

60.0

12.0

0.5

One with 2 categories, one with 3 categories

 

100

0.9

68.0

4.0

0.8

  

200 to 300

0.7

60.0

12.0

0.6

  

≥400

0.7

60.0

12.0

0.5

Both with 3 categories

 

≤100

1.0

-

-

0.9

  

200 to 300

0.7

60.0

12.0

0.6

  

≥400

0.7

60.0

12.0

0.5

Both with 4 categories

70.0

≤100

1.0

-

-

0.9

  

200

0.8

64.0

6.0

0.7

  

300 to 400

0.7

60.0

10.0

0.6

  

500

0.7

60.0

10.0

0.5

Three variables

All with 2 categories - equal prevalence

80.0

100

0.8

68.0

12.0

0.7

  

200 to 300

0.7

63.0

17.0

0.6

  

≥400

0.7

63.0

17.0

0.5

All with 2 categories - unequal prevalence

 

100

0.9

74.0

6.0

0.8

  

200

0.8

69.0

11.0

0.7

  

300

0.7

63.0

17.0

0.6

  

≥400

0.7

63.0

17.0

0.5

Four variables

All with 2 categories- equal prevalence

85.0

100

0.9

78.0

7.0

0.8

  

200

0.8

72.0

13.0

0.7

  

300 to 400

0.7

65.0

20.0

0.6

  

≥500

0.7

65.0

20.0

0.5

All with 2 categories- unequal prevalence

 

100

1.0

-

-

0.9

  

200

0.8

72.0

13.0

0.7

  

300 to 400

0.7

65.0

20.0

0.6

  

≥500

0.7

65.0

20.0

0.5

  1. aThe categories of N are dependent upon the point at which imbalance is observed (the value of probability of assignment P). bFor one prognostic variable the reduction in predictability is so small as the probability of assignment P is reduced that the recommended P- value is 1.0.