Kaelyne Chavez
RCR Week 11
COLLAPSETop of Form
1.According to the Course Skills Notes from Week 11, what are the four main types of data that one can collect (think NOIR)? What are two examples of each of these four types? And which type is most often considered the most valid (producing the most accurate results) in a study and why?
The four main types of data that once can collect are: nominal, ordinal, interval, and ratio. Two examples of nominal data are “yes” or “no” questions and names that can be listed. Two examples of ordinal data would be names that can be ranked in order and any measuring questions like “agree” or “disagree.” Regarding interval data, two examples are centigrade temperatures and the calendar in that this data does not contain a true zero point. When looking at ratio variables, examples of this are figures which contain a true zero, so age is an example, as well as income. According to the Course Skill Notes, the data which yields the most valid results are ratio data because it is exhaustive, mutually exclusive and contains meaningful measurements between the attributes withing the data that is able to also measure the degree of absence of what is being measured.
2. According to the Course Skills Notes from Week 11, what is an “instrument” in a study, and what is its primary purpose? What are surveys and questionnaires, and what kinds of data are they usually the most useful to collect from subjects? What are “scales”/”scaling questions” in surveys and questionnaires, and what are their main purpose? Also, please choose an example of a scale from the Course Skills Notes from this week and provide its name here, as well as the type of data it usually collects, and a kind of research question it might be good to use to help answer.
The course skill notes state that an “instrument” in a study is the method for which the data is being collected and its purpose is to provide the most practical and effective data for the research study. Surveys and questionnaires are methods of collecting data in which a set of questions is presented to a number of subjects. In this type of data collection, the questions are typically most formed with purpose of defining the perspective or opinion of the subject on the designated topic. According to the Course Skill Notes, “scales” and “scaling questions” are used to assign units of analysis to categories of a variable. The purpose is to assist researchers in ordering or measuring things that would otherwise not be ordered or measured. An example of a scale from the Course Skill Notes is the “Faces Scale.” The “Faces Scale” collects data which usually elicits a person’s feelings, typically how they feel physically. This scale would be useful in a research project in which the data collection involved asking children how they felt about a given topic or how a given topic made them feel. This scale could also be useful if one was conduction research in how people, more specifically again, young children, felt who were dealing with a specific illness or disease.
3. According to the Course Skills Notes for the week, what is “validity”? What is “internal validity” and what is “external validity”? And what makes these two forms of validity different from one another? Also, what is a “confound”? Please provide an example here of one form of internal confound and one form of external confound in such a way that it demonstrates to the professor that you understand why each of these examples could conceivably negatively impact the validity of a research study. Be please as specific as you can.
According to the Course Skill Notes, validity is an indication of how sound your research is, in other words, how trustworthy it is and how much truth it contains. Internal validity is when a researcher tries to stop the results of a study from being affected by the flaws within the study itself by controlling major variables or by using standard equipment which has already been tested previously. External validity is the extent to which you can generalize your findings to a larger group or other contexts and the results can be applied to contexts other than the one the research was conducted within. The difference between internal and external validity is that internal validity, to some degree, can be controlled by the researcher; whereas external validity involves the results of the study being transferred to another group of people or situation and applying or working for that other group too. Confounds are flaws within research. An example of an internal confound is an instrumental confound in which a researcher changes the instrument or data collection method mid-way through the experiment. This could compromise the data and result in flaws because it would create inconsistencies in the data that is collected. This would then affect the results of the study as the data could not be accurately compared or represented as the data was not collected consistently throughout. An example of an external confound is population characteristics confounds in which the result of the study is likely only applicable to a certain population of people containing the characteristics necessary to participate in the study. This could be for example a study in which pregnant women are surveyed about how their pregnancy is going and how they felt day-to-day during their pregnancy. This study heavily based on the characteristics of the population it is intended for, pregnant women, and it would be difficult, nearly impossible to apply this study to women who had never been pregnant and men.
4. Finally, as usual, please take a few moments to think about your own proposed research question in relation to what we’ve learned this week in class, and please be sure to remind everyone here again what your research question is. What types of data do you think you might propose to collect (choose from NOIR)? It can, of course, be more than one type. What kind of instrument(s) do you think you might propose to use, and why? Do you think you might use a survey/questionnaire, and if so, will you use scaling questions? Again, why or why not? [It is ok if you do not want to use a scale of any kind – the point is to think on using a scale or not]. And last, what are two confounds that you may specifically need to be wary of when putting together a proposed research design (this may be based on the nature of your question, the subjects you will need to solicit for the study, whether you plan to conduct the study with human subjects vs. a data set, etc).
My proposed research question is “Why do women who previously had gestational diabetes go to develop diabetes after pregnancy?” In answering this the data, I plan to use nominal and ratio. I would use the nominal data in determining the age of the person and the age at which they were both pregnant with gestational diabetes and the age in which they then went on to develop diabetes after pregnancy. This would assist me in determining if the ages the women developed both types of diabetes line up with studies that suggest the ages and time frames in which most women develop both types of diabetes. I would then use the ratio data in, if the women were willing to share it, the results of lab tests done which showed the degree to which they had diabetes. This would assist in showing the actual results which determined the diabetes. The instruments I would use to collect my data would be surveys/questionnaires and potentially in-person or virtual interviews where I asked the subjects the same set of specific questions in order to collect the results pertaining to my research question. In the survey/questionnaire I may possibly use a Likert scale if I could formulate questions that I felt subjects in the study could potentially answer based on that scale. If not then I would like not use a scale. Two confounds that I specifically need to be wary of in regard to my research question and the study I conduct based on it would be potentially a selection bias, an internal confound, in which I am not able to randomly select participants of the study and in turn simply pick friends and family who I know have been pregnant. Other potential confounds, external confounds, could be population characteristics and interaction of subject selection and research. Population characteristics in that naturally the research conducted specifically applies to pregnant women who had gestational diabetes and then develop diabetes postpartum and not simply all pregnant women. And interaction of subject selection and research, in that if I am unable to select random participants, I would be selecting people whom I know, and this could result in invalid results.
Bottom of Form
