Creating variable that captures a range of time

[Galenda Nagudi]

Hello all,

I need assistance with creating a variable that captures a range of days. Participants come for monthly visits over a period of 6 months.Exact days are the number of days between enrolment to study and a specific date. We are carrying out multiple evaluations however, we are giving room of ±7 days. Hence, I went ahead to create a variable that subtracts 7 days from exact date (days_minus7) and another that adds 7 days (days_plus7).
The range of days per visit are:

Month 1	21 - 35
Month 2	49 - 63
Month 3	77 - 91
Month 4	105 - 119
Month 5	133 - 147
Month 6	161 - 175

How can I create a variable that captures the range of days as shown in the above? Please see sample data below.

* Example generated by -dataex-. For more info, type help dataex
clear
input float PID long Visit float(exact_days days_minus7 days_plus7)
 1 1  97  90 104
 1 2  97  90 104
 1 3  97  90 104
 2 1  39  32  46
 2 1  39  32  46
 2 1  39  32  46
 2 1  39  32  46
 3 1  32  25  39
 3 1  32  25  39
 3 1  32  25  39
 3 1  32  25  39
 4 8  21  14  28
 5 1 133 126 140
 5 2 133 126 140
 5 4 133 126 140
 5 3 133 126 140
 6 1 134 127 141
 6 1 134 127 141
 6 3 134 127 141
 6 4 134 127 141
 6 1 134 127 141
 6 1 134 127 141
 6 2 134 127 141
 7 3 113 106 120
 7 1 113 106 120
 7 4 113 106 120
 7 1 113 106 120
 7 1 113 106 120
 7 1 113 106 120
 7 2 113 106 120
 8 6 327 320 334
 8 1 327 320 334
 8 2 327 320 334
 8 4 327 320 334
 9 2 110 103 117
 9 1 110 103 117
 9 3 110 103 117
10 1 110 103 117
10 4 110 103 117
10 2 110 103 117
10 3 110 103 117
11 1  40  33  47
12 3 137 130 144
12 4 137 130 144
12 2 137 130 144
12 1 137 130 144
13 8   8   1  15
14 1  40  33  47
15 3 159 152 166
15 2 159 152 166
15 5 159 152 166
15 1 159 152 166
15 4 159 152 166
16 6 173 166 180
16 3 173 166 180
16 2 173 166 180
16 1 173 166 180
16 1 173 166 180
16 5 173 166 180
16 4 173 166 180
17 5 166 159 173
17 4 166 159 173
17 1 166 159 173
17 1 166 159 173
17 2 166 159 173
17 3 166 159 173
18 1  54  47  61
19 3  61  54  68
19 2  61  54  68
19 1  61  54  68
20 2 166 159 173
20 4 166 159 173
20 3 166 159 173
20 1 166 159 173
20 5 166 159 173
20 1 166 159 173
21 4 117 110 124
21 2 117 110 124
21 3 117 110 124
21 1 117 110 124
21 1 117 110 124
21 1 117 110 124
22 1 104  97 111
22 2 104  97 111
22 3 104  97 111
23 2  75  68  82
23 1  75  68  82
24 3 169 162 176
24 2 169 162 176
24 4 169 162 176
24 1 169 162 176
24 1 169 162 176
24 5 169 162 176
24 1 169 162 176
25 1  97  90 104
25 3  97  90 104
25 2  97  90 104
26 1  75  68  82
27 1  34  27  41
28 1 179 172 186
end
label values Visit Visit
label def Visit 1 "Month1", modify
label def Visit 2 "Month2", modify
label def Visit 3 "Month3", modify
label def Visit 4 "Month4", modify
label def Visit 5 "Month5", modify
label def Visit 6 "Month6", modify

[Clyde Schechter]

I'm not sure I understand what you want. I think it is this: create a variable, let's call it range, taking on values 1 through 6. It's value will be n if the interval from days_minus7 to days_plus7 overlaps the interval specified as range of days for Month n in your original post. If I have that correct, the following code does it:

gen range = .
forvalues i = 1/6 {
    local lb = 28*`i'-7
    local ub = 28*`i'+7
    replace range = `i' if min(`ub', days_plus7) >= max(`lb', days_minus7)
}

A couple of caveats. The variable in your example data, Visit, is labeled as Month1 to Month 8, but it does not correspond well at all with the values of exact_days and the ranges defined. I really don't understand what that variable Visit represents.

Perhaps more importantly, there are some instances in your data where the values of exact_days, days_minus7, and days_plus7 do not fall anywhere at all within any of the ranges of dates you specified. The code above returns missing values for range in those observations.

[Galenda Nagudi]

Thank you so much Clyde Schechter. This is definitely what I needed. And you are absolutely right that I do have missing values. However, this is helpful in easily identifying outliers. For clarity, the "Visit" variable is one I created to indicate the participant visit and limited it to Months 1 to 6, so the 8 in the data was the screening visit.