Limitations and Delimitations

Limitations are elements the researcher cannot control.  Delimitations are controllable.  Delimitations are endless, but it seems like most of us could encounter these limitations.  

1.     Points on the Likert scale:  With five points (or any odd number), your evaluator can choose the “NA” or “No Opinion” option, limiting your data.  Although I suppose no opinion tells you something without telling you something.  An even number of points on the Likert scale forces your evaluator to choose a side and could polarize your data.  Either way, a limitation, right?  

2.     Survey respondents:  The people responding to your survey may (probably) already have some familiarity with you, your program, or at least your educational hopes and dreams.  Doesn’t this skew things?  My respondents are people who are getting free stuff from me, and I’m not the only one who has done this.  Even the National Science Foundation has to determine not only the incentive, but also the best timing and whether to award monetary or nonmonetary incentives to increase survey response rates. 

3.     Correlation versus causation:  No matter what, your data will be limited in the simple fact that despite how highly variables correlate, no causation between said variables can be determined.  This must have driven Joseph Lister really crazy.

4.     Researcher bias:  By now, we have all written, published, or taught others about the area we are researching.  It is hard to imagine that we don’t have a stake in how our efforts will all pan out at the end of the research study.  If my program measures “no change”, can I be okay with that?  Well, further research would be needed… simple as that.  



This entry was posted in Blog and tagged , , , . Bookmark the permalink.

One Response to Limitations and Delimitations

  1. Paula Dawidowicz says:

    Survey respondents don’t always know about your hopes and dreams. Here are a couple of ways you can avoid this kind of limitation. If you’ve designed your study well and are getting the appropriate number of responses, you can plan to go outside a circle of people who will know you. If you conduct the study as having blind responses (Survey Monkey can really help with that), people can feel free to respond in any way they believe is accurate. If you’ve designed your questions well, people will not be able to tell what your bias is, either. Equally important, depending on your population and the length of the survey, you may not need to offer an incentive. The more convenient you can make responding (again, Survey Monkey) and the larger your population, the less that will become an issue. Even if you do need to, for a cup of coffee at Starbucks (you’re going to keep your incentive small/very small if you want to follow NIH ethics guidelines), would you change your responses? Offering an incentive like that to get people to participate is not normally considered much of a limitation.

    Further, researcher bias is also not that clearcut. These answers appear biased in themselves, perhaps because they’re designed to be snippets about topics that are much more complex.

Leave a Reply