Experimental Strategies of Inquiry. Quantitative Methods: Surveys

¶ … experimental strategies of inquiry.

Quantitative methods: Surveys vs. experiments

Surveys are primarily used to compile data on a particular phenomenon. Cross-sectional studies attempt to encapsulate a phenomenon that is occurring at a particular point in time while longitudinal studies track a phenomenon for an extended duration of time (Survey methods, 2013, IS School). Regardless, the purpose of surveys is informative. The method is not designed come to a conclusion about why something is taking place and how we should address it. For example, a cross-sectional survey might attempt to discern the patterns of Internet use of low-income students within a particular district. A longitudinal study might track those patterns of Internet use over time. Surveys can be useful to gather data for experiments even though they rarely establish causal relationships themselves.

An experiment, in contrast, poses a hypothesis that can be proven or disproven. The experimenter understands that many different factors can affect outcomes, so a control group will be used that is similar to the experimental group, except for the presence of the independent variable that is being tested. There is "constancy of conditions across groups (aside from experimental manipulation" and assignments to control groups are randomized (Data collection, n.d., Research PO 390W: 2). The experimenter manipulates the introduction of the independent variable. The dependent variable is the result that is produced that may or may not prove the initial hypothesis (Independent vs. dependent variables, 2013, UNCP). For example, the experimenter might hypothesize that a lack of Internet access is associated with low grades for low-income students, because of lack of informational resources at home. "The experiment is the most systematic and fruitful form of data collection in the social sciences" (Data collection, n.d., Research PO 390W: 3). However, ruling out extraneous external variables that can affect the experiment (external validity) and generalizing the results to other scenarios (internal validity) can be challenging although this is required for a truly accurate experiment (Data collection, n.d., Research PO 390W: 3).