Data Collection


As we had said in our earlier discussion, there are different methods that can be used to obtain data. These methods ensure reliability and adequacy of the data. The method used will depend on the degree of accuracy desired, nature of the study, objectives and scope of the study.

 Sources of Data
We have two main sources of data; primary data and secondary data.

Primary data is the data which is collected for the first time either directly or indirectly. It is usually authentic and raw.

Secondary data is which is obtained from other individuals for a specific objective. It is not collected by the researcher. Secondary data, in most cases, might have gone through statistical operations in order to achieve the objectives for which it was first collected.

There are four methods which we can use to obtain secondary data; Sampling, Observations, Questionnaires and Interviews. We shall discuss these methods in our subsequent discussions.

Sources of secondary data are;
1. Journals and periodicals
2. Research publications by Universities and research institutions
3. Business reports
4. Official government publications
5. Individual researchers publication

Let’s have a look at one of the common tool of data collection briefly; a questionnaire.

A questionnaire is a data collection tool which has questions that will be asked to the relevant people.
The following points should be considered when constructing a questionnaire:
1. Relevancy
Only questions that are in line with your objectives must be asked.
2. Clarity
The questions should be short and easy to understand. In case you want to use technical terms, they should be defined to help the respondent in understanding.
3. Definiteness
The potential responses to the questions asked should not overlap.
4. Logic
The questions should follow logical sequence. You should know which question to put first and which one to put last.
5. The questions should not be many as they may discourage a respondent from completing the questionnaire.


Comments

Post a Comment

Popular posts from this blog

Microsoft Excel for Data Analysis

Image
For the purpose of this tutorial, we shall use Excel 2013 and 2010. However, most formulas and methods discussed can be applicable to Excel 2007. Many people outside the field of data analysis understand the power of MS Excel in data analysis. Perhaps this is because they are little aware of the many statistical programs we have for the purpose of data analysis. However, those in the field of statistics tend to forget or ignore the power of excel in data cleaning, manipulation and even analysis and visualization. One of the advantages of using Excel is that it is easy to use. Its functions are easy to apply and you don’t need to memorize all of them. Now, let’s walk together step by step and I will show you however we can enter raw data into excel, clean it, perform some basic statistical procedures and visualize our data using charts and graphs. NOTE: For some later part of this series of tutorials, you must have Analysis ToolPak. The Analysis ToolPak is an Excel add-in program t...
Read more

Data Cleaning Using MS Excel

Image
Last time we discussed briefly about data collection methods. Once we have collected the data, we need to edit it with the aim of making it clean for the next step of analysis. We shall assume that after collection, our data was entered into MS Excel. Editing is the process of examining or scrutinizing data in order to identify any errors, mistakes or omissions. Under this process, we focus on the completeness of data (no missing values), distribution of the data and extreme or outliers in the data. Some of the errors that that we may have to deal with under this process are sampling errors, Non-sampling errors, biased errors, non-biased errors, and positive or negative errors. Sampling errors occur due to the type of sampling method used during the collection of data. Sampling errors mean the difference between the estimate of a value as obtained from the sample and the actual value. Non-sampling errors take place if randomization was not used during sample selection. Biased errors...
Read more