Unmatched Logic classification

[Meshal AlQhtani]

Dear Stata Member,



I am facing a problem in classifying BMI measurement as obese, overweight, normal, and underweight. and after running these logic below, there is still obs that did not match

this is a sample of my do file

//classifiying BMI as 1 "Obese" 2 "Overweight" 3 "Healthy Weight" 4 "Underweight" //

//Bmi Categ for obese children for GILRS//

gen bmicateg_female= .

replace bmicateg_female=1 if gender==2 & age ==2 & bmi >=19.5

replace bmicateg_female=1 if gender==2 & age ==3 & bmi >=18.5

replace bmicateg_female=1 if gender==2 & age ==4 & bmi >=18

replace bmicateg_female=1 if gender==2 & age ==5 & bmi >=18

replace bmicateg_female=1 if gender==2 & age ==6 & bmi >=18.5

replace bmicateg_female=1 if gender==2 & age ==7 & bmi >=19

replace bmicateg_female=1 if gender==2 & age ==8 & bmi >=20

replace bmicateg_female=1 if gender==2 & age ==9 & bmi >=21.5

replace bmicateg_female=1 if gender==2 & age ==10 & bmi >=23.0

replace bmicateg_female=1 if gender==2 & age ==11 & bmi >=25

replace bmicateg_female=1 if gender==2 & age ==12 & bmi >=26.5

replace bmicateg_female=1 if gender==2 & age ==13 & bmi >=28

replace bmicateg_female=1 if gender==2 & age ==14 & bmi >=29.5

replace bmicateg_female=1 if gender==2 & age ==15 & bmi >=30.5

replace bmicateg_female=1 if gender==2 & age ==16 & bmi >=31

replace bmicateg_female=1 if gender==2 & age ==17 & bmi >=32

replace bmicateg_female=1 if gender==2 & age ==18 & bmi >=32

replace bmicateg_female=1 if gender==2 & age ==19 & bmi >=32

//Bmi Categ for OVERIEGHT children for GILRS//

replace bmicateg_female=2 if gender==2 & age ==2 & bmi >=18 & bmi <=19.4


replace bmicateg_female=2 if gender==2 & age ==3 & bmi >=17 & bmi<=18.4

replace bmicateg_female=2 if gender==2 & age ==4 & bmi >=17 & bmi<=17.9

replace bmicateg_female=2 if gender==2 & age ==5 & bmi >=16.5 & bmi<=17.9

replace bmicateg_female=2 if gender==2 & age ==6 & bmi >=17 & bmi<=18.4

replace bmicateg_female=2 if gender==2 & age ==7 & bmi >=17 & bmi <=18.9

replace bmicateg_female=2 if gender==2 & age ==8 & bmi >=18 & bmi<=19.99

replace bmicateg_female=2 if gender==2 & age ==9 & bmi >=19 & bmi<=21.4

replace bmicateg_female=2 if gender==2 & age ==10 & bmi >=20 & bmi<=22.9

replace bmicateg_female=2 if gender==2 & age ==11 & bmi >=21 & bmi<=24.9

replace bmicateg_female=2 if gender==2 & age ==12 & bmi >=22.5 &bmi <=26.4

replace bmicateg_female=2 if gender==2 & age ==13 & bmi >=24 & bmi<=27.9

replace bmicateg_female=2 if gender==2 & age ==14 & bmi >=25 & bmi<=29.4

replace bmicateg_female=2 if gender==2 & age ==15 & bmi >=26 & bmi<=30.4

replace bmicateg_female=2 if gender==2 & age ==16 & bmi >= 27 & bmi<=30.9

replace bmicateg_female=2 if gender==2 & age ==17 & bmi >=27 & bmi<=31.9

replace bmicateg_female=2 if gender==2 & age ==18 & bmi >=27.5 & bmi<=31.9

replace bmicateg_female=2 if gender==2 & age ==19 & bmi >=27.5 & bmi <=31.9


and below is the dataex that did not match

input byte age float(height weight bmi gender bmicateg_female bmicateb_male BMI_cat)
10 120 33.1 22.98611 2 . . .
18 167 89 31.912226 1 . . .
12 140 50.8 25.91837 1 . . .
12 152 61 26.402355 2 . . .
17 159 79.4 31.406984 1 . . .
15 168 86 30.470524 2 . . .
18 166 88 31.93497 2 . . .
18 176 99 31.96023 1 . . .
17 171 92 31.46267 1 . . .
3 89 14.6 18.432016 1 . . .
3 93 16 18.499249 1 . . .
12 157 64 25.96454 1 . . .
12 158 66 26.43807 2 . . .
13 139 52 26.91372 1 . . .
3 87 14 18.496498 1 . . .
11 143 50 24.451075 1 . . .
14 139 55 28.466436 1 . . .
17 170 91 31.48789 1 . . .
12 157 65.2 26.451374 2 . . .
3 84 13 18.424038 1 . . .
15 156 74 30.40763 2 . . .
11 159 63 24.9199 2 . . .
17 154 75.7 31.91938 2 . . .
16 159 77 30.457655 1 . . .
10 140 44 22.44898 1 . . .
14 148 64.6 29.49233 2 . . .
16 165 83 30.486687 1 . . .
18 177 100 31.91931 1 . . .
18 167 89 31.912226 1 . . .
17 171 92 31.46267 1 . . .
15 156 74 30.40763 2 . . .
3 87 14 18.496498 2 . . .
3 84 13 18.424038 2 . . .
10 170 65 22.49135 1 . . .
13 135 51 27.98354 2 . . .
12 157 64 25.96454 1 . . .
14 174 86 28.405336 1 . . .
16 168 86 30.470524 1 . . .
15 160 78 30.46875 2 . . .
12 152 61 26.402355 2 . . .
17 165 87 31.955923 2 . . .
10 135 41 22.49657 1 . . .
16 156 74 30.40763 1 . . .
10 140 44 22.44898 1 . . .
13 154 64 26.986 1 . . .
18 166 88 31.93497 1 . . .
11 147 54 24.989586 2 . . .
18 164 86 31.975016 2 . . .
18 164 86 31.975016 1 . . .
16 170 88 30.449825 1 . . .
16 159 77 30.457655 1 . . .
11 140 48 24.4898 1 . . .
3 84 13 18.424038 2 . . .
12 143 54 26.40716 2 . . .
12 158 66 26.43807 2 . . .
16 164 82 30.487806 1 . . .
17 148 70 31.95763 2 . . .
11 143 50 24.451075 1 . . .
14 159 72 28.479885 1 . . .
11 153 57.2 24.43505 1 . . .
18 147 69 31.93114 2 . . .
14 144 61 29.417437 2 . . .
15 158 76 30.44384 2 . . .
10 143 46 22.49499 1 . . .
16 164 82 30.487806 1 . . .
3 96 17 18.446182 1 . . .
16 167 85 30.47797 1 . . .
15 170 85 29.41176 1 . . .
12 152 61 26.402355 2 . . .
10 155 54 22.47659 1 . . .
12 153 62 26.48554 2 . . .
13 157 68.8 27.91188 2 . . .
3 93 16 18.499249 2 . . .
16 165 83 30.486687 1 . . .
12 152 61 26.402355 2 . . .
12 152 61 26.402355 2 . . .
3 93 16 18.499249 2 . . .
11 154 58 24.456064 1 . . .
14 179 91 28.40111 1 . . .
15 160 78 30.46875 2 . . .
18 167 89 31.912226 2 . . .
10 140 44 22.44898 1 . . .
15 146 65 30.493525 2 . . .
10 146 49 22.987427 2 . . .
15 157 75 30.427197 2 . . .
16 170 88 30.449825 1 . . .
18 177 100 31.91931 1 . . .
10 130 38 22.48521 1 . . .
11 151 57 24.998903 2 . . .
11 155 60 24.97399 2 . . .
16 169 87 30.461117 1 . . .
10 140 44 22.44898 1 . . .
10 143 46 22.49499 1 . . .
16 145 65 30.915575 2 . . .
13 148 59 26.93572 1 . . .
17 170 91 31.48789 1 . . .
16 145 65 30.915575 2 . . .
12 141 52.5 26.407124 2 . . .
16 172 90 30.42185 1 . . .
17 166 88 31.93497 2 . . .
end
label values gender gn
label def gn 1 "Male", modify
label def gn 2 "Female", modify
label values BMI_cat bm1






kindly for your assistance please

[Clyde Schechter]

Look at, for example, the first observation in your example data. This is a female, age 10, with BMI = 22.98611. Now look at your code. For bmicateg_female = 1, at age 10, the threshold is >= 23.0, which she does not meet. And for bmicateg_female = 2, at age 10, the threshold is <= 22.9, which she also does not meet. Her actual BMI falls between those two thresholds, so she never gets assigned a bmi category.

You are probably thinking of bmi's as if they are exact 1-decimal place numbers, because we typically report them that way. But your calculated BMIs have greater precision than that and that is tripping you up.

You need to decide how you want to resolve this problem. One possibility is to round the BMIs to the nearest 0.1. But using Stata's -round()- function for this will fail because internally all of Stata's numbers are binary, and most decimals have no exact finite binary representation. So what you would have to do is actually round 10*bmi to the nearest integer, and rephrase your criteria in terms of this rounded 10*bmi instead of the bmi itself.

So, like this:

gen bmi10 = round(10*bmi, 1)

gen bmicateg_female= .

replace bmicateg_female=1 if gender==2 & age ==2 & bmi10 >=195

replace bmicateg_female=1 if gender==2 & age ==3 & bmi10 >=185

replace bmicateg_female=1 if gender==2 & age ==4 & bmi10 >=18

replace bmicateg_female=1 if gender==2 & age ==5 & bmi10 >=18

replace bmicateg_female=1 if gender==2 & age ==6 & bmi10 >=185

replace bmicateg_female=1 if gender==2 & age ==7 & bmi10 >=19

replace bmicateg_female=1 if gender==2 & age ==8 & bmi10 >=20

replace bmicateg_female=1 if gender==2 & age ==9 & bmi10 >=215

replace bmicateg_female=1 if gender==2 & age ==10 & bmi10 >=230

replace bmicateg_female=1 if gender==2 & age ==11 & bmi10 >=25

replace bmicateg_female=1 if gender==2 & age ==12 & bmi10 >=265

replace bmicateg_female=1 if gender==2 & age ==13 & bmi10 >=28

replace bmicateg_female=1 if gender==2 & age ==14 & bmi10 >=295

replace bmicateg_female=1 if gender==2 & age ==15 & bmi10 >=305

replace bmicateg_female=1 if gender==2 & age ==16 & bmi10 >=31

replace bmicateg_female=1 if gender==2 & age ==17 & bmi10 >=32

replace bmicateg_female=1 if gender==2 & age ==18 & bmi10 >=32

replace bmicateg_female=1 if gender==2 & age ==19 & bmi10 >=32

//bmi10 Categ for OVERIEGHT children for GILRS//

replace bmicateg_female=2 if gender==2 & age ==2 & bmi10 >=18 & bmi10 <=194


replace bmicateg_female=2 if gender==2 & age ==3 & bmi10 >=17 & bmi10 <=184

replace bmicateg_female=2 if gender==2 & age ==4 & bmi10 >=17 & bmi10 <=179

replace bmicateg_female=2 if gender==2 & age ==5 & bmi10 >=165 & bmi10 <=179

replace bmicateg_female=2 if gender==2 & age ==6 & bmi10 >=17 & bmi10 <=184

replace bmicateg_female=2 if gender==2 & age ==7 & bmi10 >=17 & bmi10 <=189

replace bmicateg_female=2 if gender==2 & age ==8 & bmi10 >=18 & bmi10 <=1999

replace bmicateg_female=2 if gender==2 & age ==9 & bmi10 >=19 & bmi10 <=214

replace bmicateg_female=2 if gender==2 & age ==10 & bmi10 >=20 & bmi10 <=229

replace bmicateg_female=2 if gender==2 & age ==11 & bmi10 >=21 & bmi10 <=249

replace bmicateg_female=2 if gender==2 & age ==12 & bmi10 >=225 &bmi10 <=264

replace bmicateg_female=2 if gender==2 & age ==13 & bmi10 >=24 & bmi10 <=279

replace bmicateg_female=2 if gender==2 & age ==14 & bmi10 >=25 & bmi10 <=294

replace bmicateg_female=2 if gender==2 & age ==15 & bmi10 >=26 & bmi10 <=304

replace bmicateg_female=2 if gender==2 & age ==16 & bmi10 >= 27 & bmi10 <=309

replace bmicateg_female=2 if gender==2 & age ==17 & bmi10 >=27 & bmi10 <=319

replace bmicateg_female=2 if gender==2 & age ==18 & bmi10 >=275 & bmi10 <=319

replace bmicateg_female=2 if gender==2 & age ==19 & bmi10 >=275 & bmi10 <=319

[Meshal AlQhtani]

Many thanks, Mr.Clyde, it worked!


I really appreciated it