The data which has been collected from various primary/secondary sources is RAW in nature, this means that there are likely chances of errors and inconsistencies in it.

Since data collected is of pivotal importance to policy and decision makers everywhere be it governmental departments, business organizations, health or educational institutions etc it would be better to have a team of experts at hand who know how to scrutinize, review and edit this data before it is finally fed into the data bases and the required statistics are generated.
So what is data editing and what does it involve?

Read: the impact of automated teller machine topic

How to define data editing
The process through which the data is reviewed to check for consistency, adequacy, detect errors and outliers (values that are either too big or too small from the rest of the data) and the correction of errors within the data in order to maximize its usefulness for the purpose for which it was collected is called data editing.

Purposes and objectives of data editing
The basic purpose served by data editing is that it improves the quality, accuracy and adequacy of the collected data thereby making it more suitable for the purpose for which the data was collected. The following can therefore be identified as the main objectives of the data editing process :

Detection of errors in the data that otherwise affect the validity of outputs.
Validation of data for the purposes it was collected.
Provision of information that that would help access the overall level of accuracy of the data.
Detection and identification of any inconsistencies in the data and outliers and to make adjustments for them.
Types of data editing
There are different types of data editing

Validity and completeness of data: refers to correctness and completeness of obtained responses. This helps ensure that there are no missing values or empty fields in the data bases.

Range: verifies that data within a field fall between the boundaries specified for the particular field.
Duplicate data entry: this helps ensure that there is no repetition or duplication of data and each unit on the data base or register was filled only once.
Logical consistency: through this type of editing connections between data fields or variables are taken into account.

Outliers: this type of editing helps detect values that are too extreme or unusual so that they can be verified and checked.
Stages of data editing
The data editing process can be divided into three parts or stages

The first stage is the stage where you set the rules for editing. This stage is further subdivided into two steps.
In step one, you provide instructions to desk editors who then check the data for coherence and consistency.

In step two, you set the rules by establishing logical relations between the variables according to various criteria. This set of rules is called automated validation rules and this type of editing seeks to detect errors during data entry and to screen them.

The manual desk editing stage is a traditional method that is put into effect by a specialized editing team. The data,(if) on paper is checked after the data has been collected and before it is fed into the data bases. If however, electronic means have been used to collect the data, the forms entered into the database are revised individually.
The automated data editing method makes use of computer programs and systems for checking the data all at once after it has been entered electronically. These programs and systems contain Audit rules which validate the data, detect errors and determine unacceptable responses.
Limitations to data editing
There are certain factors that can influence or pose limitations to the data editing process, these can be summed up as follows;

Data editing can be influenced by the amount of time available, the budget, the presence or absence of other resources and also by the group of people involved in the editing process.
The available computer software programs.
Follow up with the respondents is of critical importance in the data editing process because they are often the best source of information in many cases. However, the respondents might feel this to be stressful and burdensome thereby causing limitations to the data editing process.

Some types of data do not require extensive editing, therefore it would be better to keep in mind the intended uses of data and make sure that the more important part of data iz kept free from all errors. In this way, the intended use of data does play an important role in influencing the data editing process.
What you need to do is to establish the methods and procedures that must be followed while correcting or handling the data errors, in the survey plan, right at the start of the project otherwise the process would be of no or little use.
Also remember that if you plan to edit your data manually, you must develop and document the methods that are to followed. Your team must be trained, a method must be established to check their work progress and the impact of the edits on the original data must also be assessed.
In case of automated editing, you would need to develop and document the rules for editing. You might also need to develop a software or customize an existing computer program as per your data editing demands or requirements.
General guidelines for data editing
There are certain things that must be kept in mind when editing the data, they are as follows;

Who should make or set the editing rules? The answer to this question would be that such rules should be made by professionals who are experts in data collection, questionnaire design and analysis.

The editing rules need to be consistent and free from any contradictions.
When setting the editing rules, it must be established whether the variable is qualitative or quantitative because the rules for editing either one are different from the other.

Give enough time to each of the various stages of the process, that is, data collection entry and analysis and at the end of each make a quick check to see that all the necessary edits have been made and that there are no empty places within the questionnaire form.

The questionnaire must be edited in full during the early stages of editing. If however it is found that some errors remain, a sample of forms should be subjected to re editing. The size of the sample is determined according to the expected number of the remaining errors.
You also need to re run the desk editing stage to ensure that the data is almost free from all errors.
The questionnaire must be subjected to desk editing and also to automated rules within the built in data editing/computer software programs.