Data processing is crucial to any research program; even the best-designed questionnaire cannot make up for data that is incorrect or incomplete. The amount of data that is generated in the various industries could be overwhelming for the organisations to manage. This is where data processing comes into the picture. By implementing data processing techniques, the data that are generated can be organised, presented and made readily accessible.
- Data processing is the act of handling or manipulating data in some fashion. Regardless of the activities involved in it, processing tries to assign meaning to data. Thus, the ultimate goal of processing is to transform data into information. Data processing is the process through which facts and figures are collected, assigned meaning, communicated to others and retained for future use. Hence we can define data processing as a series of actions or operations that converts data into useful information. We use the term ‘data processing system’ to include the resources that are used to accomplish the processing of data.
- Market research data processing can refer to different aspects of the entire market research analysis process. Most often, data processing and data cleaning are used interchangeably.
Occasionally, a respondent to your survey does not really know the answer to a survey question (and just guesses) or simply makes a mistake in answering, or would just instead not answer the question (make sure you put in a “don’t know” category for the respondent to default to).
For example, suppose a survey question asks how many children live in the household. Instead of typing in a “2”, the respondent (or the interviewer) slips and types in “22” accidentally. Now, it might be possible for 22 children to be living in a house (not a home we would want to live in!), but we can probably assume that this was a mistake. Since we don’t know the correct response for sure, we would change the response for that respondent from “22” to “missing.” That way, this answer is not counted as part of the statistics generated for this question.
If we do not “clean” this market research data, when we calculated the average number of children living in respondents’ houses, the number would be inflated. That could easily cause the researcher to make an incorrect conclusion based on the data. By “cleaning” the data, those responses would be “corrected,” and the statistical software you are using (or spreadsheet or whatever) would not include that response in the analysis.
- The good thing is that you don’t have to go through every response from every respondent to clean the data. You can generate (or have a professional analyst produce for you) what is called a frequency table or “freqs.” A freq is simply a count of each response category for each question in the survey. So, in the example above, the freq would list the response category of 0 and the number of respondents who gave that answer. It would do so for all responses to the question. You can easily go through the freq to see what answers just out as mistakes. You “clean” them in the dataset and presto, you are ready to analyse your market research data.
DATA PROCESSING ACTIVITIES
Man has in course of time devised certain tools to help him in processing data. These include;
- Manual tools such as pencil and paper
- mechanical tools such as filing cabinets
- Electromechanical tools such as adding machines and typewriters, and
- Electronic tools such as calculators and computers.
Many people immediately associate data processing with computers. As stated above, a computer is not the only tool used for data processing; it can be done without computers also. However, computers have outperformed people for certain tasks. There are some other tasks for which computers are a poor substitute for human skill and intelligence.
Regardless to the type of equipment/tool used, various functions and activities which need to be performed for data processing can be grouped under five basic categories as shown in the Diagram below