Microsoft Word research method fm doc

(3) Computer compilation
Before you decide to use a computer, you have to be sure that it will save time or that the 
quality of the analysis will benefit from it. Note that feeding data into a computer costs time 
and money. The computer should not be used if your sample is small and the data is mainly 
generated by open questions (qualitative data), unless there is a resource person who is 
competent in using a program for qualitative data analysis. The larger the sample, the more 
beneficial in general the use of a computer will be. 
Computer compilation consists of the following steps: 
1. Choosing an appropriate computer program
2. Data entry
3. Verification or validation of the data
4. Programming (if necessary)
5. Computer outputs/prints
 
i. Choosing an appropriate computer program 
The identification of an appropriate statistical package is the first step in using a computer. 
Some examples of packages commonly used are: Epi Info, SPSS, STATA, etc. 

---

Research methodology 
74
ii. Data entry
To enter data into the computer you have to develop a data entry format, depending on the 
program you are using. After deciding on a data entry format, the information on the data 
collection instrument will have to be coded (e.g., Male: M or 1, Female: F or 2). During data 
entry, the information relating to each subject in the study is keyed into the computer in the 
form of the relevant code. 
iii. Verification
During data entry, mistakes will definitely creep in. The computer can print out the data 
exactly as it has been entered, so the printout can be checked visually for obvious errors, 
(e.g., exceptionally long or short lines, blanks that should not be there, alphabetic codes 
where numbers are expected, obviously wrong codes). 
Example:
• 
Codes 3-8 in the column for sex where only 1(F) and 2 (M) are possible
• 
Codes above 250 when you had only 250 subjects
If possible, computer verification should be built in. This involves giving the appropriate 
commands to identify errors. 
Example:
The computer can be instructed to identify and print out all subjects where the ‘sex’ 
column has a code different from 1 (F) or 2 (M).