Harmonisation of variables names prior to conducting statistical analyses with multiple datasets: an automated approach

Table 3 Partial view of Stata table ('_Structure.dta') containing examples of variables of interest.

Obser-vation	Items	DOBcmc	DTP3	Urb
1	Key terms	Date Birth	DTP 3	Area Residence
2	0	NA	Not vaccinated
3	1	NA	Vaccinated recall	Rural
4	2	NA	Vaccinated mark	Urban
5	3	NA	Vaccinated date
6	4	NA
7	9	Inconsistent	Inconsistent	Inconsistent
...	...	...	...	...
11(+)	dc-MICS_2-ch.dta	cdob	dtcoq3	hi6
12(+)	ke-DHS_41-ch.dta	v011	h7	v025
13(+)	mo-MICS_2-ch.dta	NA	dpt3	hi6
	(rest of the datasets)

Rows headings (observation numbers) and columns headings are used by Stata to define the configuration of this table. From rows 1 to 7: values labels of the variables of interest defined by the user. From row 11 onwards, datasets files names and variables as named in those datasets corresponding to each variable of interest.
'DOBcmc', 'DTP3' and 'Urb' are the harmonised names for the variables date of birth DTP3 and area of residence respectively. The terms in each column, in rows 11 to 13, refer to the names of those variables in the datasets.
The file may contain additional fields as needed (e.g. additional information of the datasets included in the analyses) and fields can be renamed to be accommodated to other programming practices.
NA: not available (a numerical variable usually does not have codes).
(*) There are as many additional columns as variables of interest. (+) There are as many additional rows as datasets included in the project or analyses.

ISSN: 1472-6947