One of the most popular ways to collect data is through a survey. Think about how often you are asked to take a survey…when you leave a store, when you purchase something online, after talking to your cell phone carrier on the phone they ask if you are willing to take a survey, etc. They are everywhere. Why? The general features are easy to work with.
· They are versatile – Surveys can be designed to study almost any social issue and we can use them to learn about people, organizations, criminal behavior, etc.
· They are efficient. Data can easily be collected from a large number of people relatively inexpensively and quickly.
· Often times they can be generalizable as we are usually able to get information from a representative sample of a large population (Bachman and Schutt, 2012).
Students should understand the basic terminology involved in survey research including the following: Survey Research – Collection of information from a sample of individuals through their responses to questions.
Survey instrument containing the questions in a self-administered survey.
A person who answers questions on a survey.
Percentage of persons surveyed who actually complete a survey.
(Bachman and Schutt, 2012)
Please note that probably the most important thing to know about a survey is the development of the survey questions is key. This will affect the reliability and validity of any data collected. Second to this is how the interview or survey is administered; this too is important and relates directly to the reliability and validity of the survey results. Keep in mind that survey questions are answered as part of a questionnaire, not in isolation from other questions. The context created by the questionnaire and how the questionnaire is administered impact how individual questions are interpreted; and whether they are even answered. Finally, we must give very careful attention to design of the questionnaire as a whole, as well as to each individual question that it includes. Remember, to do this type of work at APUS you will need IRB approval!
Another method that we can pull from to carry out quantitative research is through secondary data analysis. There is a tremendous amount of research done each year that makes use of this method. This type of research uses pre-existing data to perform the study in question rather than going out and collecting new data for the inquiry. In effect, the data have already been collected but they will be analyzed in a different way or can be used to answer a different research question than was originally intended by those who collected the data in the first place. This happens all of the time and it’s a great way to do research.
The major types of secondary data analysis used in social science research include:
· official statistics
· official records
· historical documents
The most common sources of secondary data are social science survey data collected in studies funded by federal and state government. Other common sources include official records maintained by government agencies for administrative, rather than research, purposes – for example, police department arrest data. (Data such as this may also still need IRB approval!) Often what happens is a government agency – like the National Institute of Justice – will fund data collection and research. After the data is collected and several reports are written, the government agency then asks for a clean set of the data collected; they then hold this for other researchers that may want to analyze the data. Other researchers can request a portion of the data to answer certain research questions. In fact, the government agency encourages use of the data to help with our understanding of various social science and social policy questions. They will even sometimes fund researchers’ projects that analyze secondary data.
Validity and Reliability
You should know by now that validity and reliability are two very important research terms. To assess validity in a study we are essentially asking if we are studying what we think we are studying? Are we measuring what we think we are and do our measures really represent the concepts that we think they do?
Reliability is achieved when a measure or study yields consistent scores or observations on different occasions and/or different locations (city to city for example). Basically we’re concerned with the consistency of the measures that are in use (Bryman, 2012). A study and or measure must be reliable in order to be valid.
Some options for testing reliability include:
TEST-RETEST METHOD (STABILITY)
Where a study or measure obtains the same results at two different times.
INTER-ITEM (INTERNAL RELIABILITY)
When we use multiple items to measure a single concept.
Compare slightly different versions of measures.
Use more than one observer to measure the same thing.
Validity refers to whether or not the test measures what we think it measures. There are a number of ways to determine measurement validity including face validity, concurrent validity, predictive validity, construct validity and convergent validity (Bryman, 2012). Face validity is what researchers need to establish within their work as this basically shows that you are “getting at the concept that is the focus of attention.” (Bryman, 2012, p. 171).
In this lesson, we moved into quantitative analysis which involves the techniques by which researchers convert data into a numerical form and subject it to statistical analyses. In the next lesson, we will start thinking about how we’ll be working with our data.
Bachman, Robert and Russell K. Schutt. (2012). Fundamentals of Research in Criminology and Criminal Justice. Thousand Oaks, California: Sage.
Bryman, Alan. (2012). Social Research Methods 4th ed. New York: Oxford University Press.