Dear Amit,

Data obtained via a third party website is fine provided that it is reliable, of a sufficient amount and enables you to answer your research question. If it isn't then collecting your own data is probably a better option because you will have more control over these issues.

Dear Amit,

Data obtained via a third party website is fine provided that it is reliable, of a sufficient amount and enables you to answer your research question. If it isn't then collecting your own data is probably a better option because you will have more control over these issues.

I too agree with Mr. Peter Samuels. you get a glance of the the total picture from the national or international level data. but to specifically address a research question in a particular area in-depth manner, you need to take data as per your requirements. such studies acts as supporting evidence to the findings of larger studies and vice versa.

Data drawn from nationally representative samples are the gold standard for external validity, but their value declines as rudimentary proxy variables are often drawn from those data sets in place of reliable and valid measures. Secondary data will also err in aggregating data to the detriment of hidden populations. Collecting your own data can improve the interval validity of the findings by disaggregating the data and measuring the variables of interest in a more reliable and valid manner.

The comment about preference for primary data is occasionally overstretched. There are occasions when a researcher cannot directly obtain primary data.  Use of National or Regional or International data does not necessarily render the research unreliable. 

The preference for primary data is for specific studies for specific purposes. I vividly remember a situation when a postgraduate student planned to collect population data for a very large country that already had reliable national census data.

Dear Amit,

There is no speculation that secondary data is a great source of information to address research objectives. Some of the disciplines in social sciences, especially Economic research widely utilizes secondary data for answering the research questions. Although it is very useful, sometimes the available secondary data may not be enough capacity to answer your specific research question. To conclude, the selection of sources of data is highly depended on the RESEARCH QUESTION. It is the rationale of the researcher to look into the available data and decide its capability of answering the questions.

Government-sponsored surveys often are performed with resources you are unlikely to be able to match. If such surveys offer the data you need, there is no reason to try to duplicate such data. You'll waste time and money, and probably get inferior data. Of course, if you need data that they don't offer, then you need alternative data. You may also be asking about administrative data and other types of data besides survey data, often classified with big data. There are advantages and disadvantages to both survey data and big data, and that would be a complex discussion in itself.

The data in those repositories tend to be from larger samples, more representative samples, and from questions that have been previously field tested. The advantage to doing your own questionnaire is that you can ask questions not on those larger surveys. Another advantage, is that you can focus on a specific sample that you know to differ from the larger population in certain ways, and you want to see if it differs on other variables as well.

If the data from government web-sites answers your questions, I do not see the need of interviewing/surveying individuals working for the government in order to obtain the data that is already publicly available.

For me it is always interesting to compare "official" secondary data with results of well created primary research. Sometimes you get really interesting results.

Documented data or secondary data normally is not the most recent data as compared with primary one. Both has its own strength. You may also compare and contrast the trend of  respondent' behaviour having both data in hand.  Next, if data provided by reliable n trusted institutions and the data is sufficient to answer your objectives, look no further; used it.

Longitudinal studies may best be undertaken using secondary data. It is also useful where regional or international comparative studies are being undertaken. However, check whether the available secondary data meets the needs of your study. If there is some limitation, then you mention in your research report.

Amit -

Note that government websites generally use questionnaires themselves.  I think the first thing you need to do is to contact that government agency or agencies, and ask them their methods of data collection, analysis, and inference, including estimated standard errors.  Any metadata they can report may be helpful.  What are the definitions of the populations studied?  What were the goals?  Any questions you have that are not answered on the government websites may possibly be answered by contacting the agency/agencies responsible for creating the websites and their numerical content.  You won't know if you need another data collection if you don't know exactly what you are being given on the website.  Some agencies may give you the data, in some cases, if you want to aggregate it a different way, or perhaps even possibly redefine target populations.  I think it best to find out what is available, because, as others have said, you won't be able to match the resources and expertise brought to the task that a government agency can. 

Cheers - Jim