torsdag 15 november 2012

Theme 3: Post-reflections


Deductive and inductive research
During her lecture, Marta Cleveland-Innes (2012) presented the terms deductive and inductive research, which I hadn’t been acquainted with before and that I found interesting. By doing a deductive research you start from general ideas, like theory and general studies, and work your way toward a more specific part, with more specific data. According to Cleveland-Innes, deductive research is more common for quantitative researchers since they often work with lager sampling sizes and more general data. Inductive research is the opposite, when you start from a more specific point and data and move toward more general data and analysis (SocialResearchMethods.net, 2012). Inductive research is more often used by qualitative researchers, as they use more detailed and specific data and a smaller sampling size.

The image Cleveland-Innes showed at her lecture:

Image reference: https://toknow-11.wikispaces.com/file/view/induction-deduction-1.png/284914014/induction-deduction-1.png

By using a mixed research I guess that it is hard to say if it is deductive or inductive, but it might be beneficial to do the research both ways. That way you work somewhat iteratively and can compare the starting theories from the first method with the end results from the second.


Validity and reliability
Cleveland-Innes (2012) spoke shortly about validity and reliability during her lecture. She stressed that researchers must strive for both of these to be relevant and accurate. I believe that these two terms are often mixed up, and I have at least previously been doing it. I really liked the image she used to present it, which was:

Image reference: http://explorable.com/images/validity-and-reliability.jpg

Reliability is when you get similar results for every time you do a test or research, and when the different people you test on get similar results. Validity is when the results are close to reality and when the methods to gather data have been conducted in a correct and controlled way. One example Cleveland-Innes (2012) mentioned, and that is stated on www.exporable.com (2012), is that researches should have randomly selected sample persons for the research to be valid. If some sample persons drop out or are not randomly chosen, the validity immediately gets lowered.

I believe that there are different obstacles and difficulties with conducting valid and reliable researches depending on if the research is more quantitative or qualitative. In the case of quantitative research it is important to have a big sample size for the data to be valid, while in a qualitative research it is more important to have randomized and reliable persons because of the lower sample size.


References:
Cleveland-Innes, Martha (2012). Lecture, 2012-11-12.
SocialResearchMethods.net (2012). http://www.socialresearchmethods.net/kb/dedind.php.
Explorable.com (2012). http://explorable.com/validity-and-reliability.html.

1 kommentar:

  1. I don't really see why a person dropping out of a test would make the results less valid; you as a researcher can't predict what will happen. And if the dropout is replaced with another random sampled person from the same target group; will the study still have less validity?

    SvaraRadera