how to use an asterisk(*) for multiple variables with an "if" condition

[Deogratious Sebuwufu]

Dear all, i have an issues which iam trying to figure out but i am not getting anywhere and i need ur help. For example i have several individuals in a household with different sex and i need to use for example "count sex* if age*<13 & age*>6 & sex*=="male" " this seems not to work. any advice on how to run this and it picks all the numbers required?

[Nick Cox]

That's illegal, which explains why it doesn't work.

What are the names of your variables? Do you want just

the number of males who are aged 7-12

because typically that would depend on an age variable and a sex variable?

[Deogratious Sebuwufu]

I have a dataset which has household members and in that dataset, it records all the people with different sex [person1(sex1), person2(sex2), person3(sex3).....person10(sex10)] and different age (groups [person1(age1), person2(age2), person3(age3).....person10(age10)]) as well. that means, if in the household i have ten(10) members, i will have males and females. Secondary, since i am interested in knowning the sex of the age group ( 7 -12yrs) of interest and for a particular sex for example "males", that would be of importance to me. I did it in excel and i get results but i don't see reasons why i kept going to excel. i want to use start fully with syntaxes for anyone who wants to proof read my analysis.

[Nick Cox]

I think you may need to give a data example. Do look at https://www.statalist.org/forums/help#stata

Otherwise it sounds as if you may want to use one of the functions in egen whose names start with any or row.

Alternatively, your data structure may be better reshaped so that each observation is a person not a household.

[Deogratious Sebuwufu]

below is an example which i have pulled from my dataset

uniqueid	hhrostersex1	hhrosteryr1	hhrostersex2	hhrosteryr2	hhrostersex3	hhrosteryr3	hhrostersex4	hhrosteryr4	hhrostersex5	hhrosteryr5	hhrostersex6	hhrosteryr6	hhrostersex7	hhrosteryr7	hhrostersex8	hhrosteryr8	hhrostersex9	hhrosteryr9	hhrostersex10	hhrosteryr10
11036174101
1103617411	Female	49	Male	54	Male	20	Male	19	Male	17	Male	15	Female	10	Male	8	Male	16
11036174111	Female	58	Male	63	Male	18	Male	15	Male	4
11036174121	Female	23	Male	23	Female	43	Female	14	Female	7	Male	15	Male	12	Female	1
11036174131	Female	30	Male	54	Female	45	Female	14	Female	13	Male	16	Male	21	Female	19	Female	5
11036174132	Female	23	Male	30	Male	4	Female	2
11036174141
11036174151	Male	46	Female	41	Male	20	Male	18	Female	17	Male	16	Female	14	Female	12
11036174161	Female	38	Male	43	Female	6	Female	3	Female	1
11036174171	Male	63	Female	38	Male	14	Male	10	Female	3
11036174181
11036174191	Female	20	Male	23	Female	12	Male	2
11036174201	Female	22	Male	22
1103617421
11036174211	Female	66	Male	71	Female	15	Male	9
11036174221	Female	20	Male	30	Female	12
11036174231	Female	15	Male	43	Female	38	Female	13	Male	11	Female	10	Male	2
11036174241	Female	35	Male	44	Female	20	Male	13	Female	11	Male	10	Male	7	Female	3
11036174251	Male	38	Female	29	Male	12	Male	10	Male	10	Female	6	Female	5
11036174261	Male	48	Female	35	Female	2
11036174271	Male	27	Female	22	Male	3	Male	2
11036174281	Male	76	Female	70	Male	5
11036174291	Female	65

[Nick Cox]

Unless you use dataex as requested I have no way of working out whether your sex variables are string or numeric with value labels.

[Deogratious Sebuwufu]

the sex variable is coded and its numeric (1"Male" 2"Female")

[Nick Cox]

Still not using dataex. A little progress is possible but hard work for the reader. Let's show the principles by addressing the syntax that wouldn't work

count sex* if age*<13 & age*>6 & sex*=="male"

This syntax is wrong (1) because your variable names don't match (2) because wildcards can't appear in this command (3) because your sex variables are numeric.

I think this needs to be a loop

gen age7_12male = 0 

forval j = 1/10 { 
       replace age7_12male = age7_12male + (hhrostersex`j' == 1  & inrange(hhrosteryr`j', 7, 12))  
}

You need to look at each pair of variables and check whether you have suitable sex and age to include it in the count.

Your present data structure obliges this. A reshape long would help for this purpose, but perhaps not other purposes.

[Deogratious Sebuwufu]

Hello Cox, i have a pressing question. How can i run cross tabulations, create dummy variables and run a regression with a dataset in stata in wide format. The solutions you gave me last time were the best and i expanded on that.

[Nick Cox]

Please start a new thread. However, you’re asking for code for a entire task, which is optimistic.

I am travelling now and restricted to short answers.

[Deogratious Sebuwufu]

Thanks. I will be honoured by your replay. I have also posted the thread "how to create a cross-tabulations, dummy variables and regression with wide data"

[William Lisowski]

See

https://www.statalist.org/forums/for...with-wide-data

for the thread mentioned in post #11.

[Deogratious Sebuwufu]

can't i just do the analysis without using the reshape command?