Leslie Gale's Blog

Just another WordPress.com weblog

Standard 11 Meta Reflection: Inquiry/Research August 23, 2013

Filed under: Standard 11. Inquiry/Research — lkgale @ 2:39 am
Tags: ,

Taking two courses pertaining to educational research, while a bit daunting, was instrumental in supplying me with the skills I need to understand the role of research in education as well as apply the concepts, procedures and vocabulary to my educational practice. The appropriate use of educational research enables educators collect and analyze data to make make sound instructional decisions, evaluate research findings for validity and effectively communicate research findings to others.

Research methods can be either quantitative or qualitative. Quantitative studies are objective and seek to bring forth unambiguous findings. Such research emphasizes numbers and measurements. Qualitative research involves the interpretation of verbal narratives and observations rather than numbers. Historically, educational research was primarily quantitative but today it is common to find useful qualitative studies or mixed-method designs. (McMillan, 2012)

Researchers must also decide whether the question or hypothesis requires an experimental or non-experimental design (McMillan, 2012) . An experimental design is necessary if the investigator wishes to directly manipulate a variable and establish cause and effect. For example, experimental design would be appropriate if a researcher wanted to determine which group of students achieved higher fluency scores, those that practiced with a parent volunteer or those who practiced with an older buddy student. An investigator who does not plan to manipulate a variable will employ a non-experimental research design. This would be appropriate if there was a desire to investigate whether or not there is a correlation between teacher attitudes towards a particular math curriculum and student achievement. Observations would be made and data collected but no variable would be manipulated.

Action research is another research design that is used to solve specific problems at one site such as a classroom or school. There are three types of action research which are individual teacher research, collaborative action research, and school wide action research.  Veteran teacher Wendi Pillars conducted an individual action research study to address the problem of minimal reading growth in her ELL third graders. Although she did not find the all-inclusive answer to her question, she found that “AR can be a way to practice what I preach, take my own learning deeper, and improve my effectiveness as a teacher “(Pillar, 2012).

Having an understanding of the processing and analysis of research data allows one to organize data in a meaningful way as well as critically review and interpret the research of others. Researchers are able to summarize and compare scores and values by understanding the nature of percentages, measures of central tendency, the normal curve and probability.

The ability to determine the validity of a research study is valuable both for the researcher and the consumer. Validity can be external or internal. External validity addresses how well the study results can be generalized to different subjects within the same defined population parameters. Internal validity refers to “the extent to which the intervention, and not extraneous or confounding variables, produced the observed effect” (McMillan, 2012). In order for research studies to be considered credible consumers and researchers alike must evaluate factors that contribute to or detract from validity. As an illustration, validity is questioned in a study by Stanford University’s Center for Research on Education Outcomes (CREDO) which concluded that there is steady progress in student achievement level in charter schools. In a letter to the editor of Education Week (Center for Education Reform…, August  2013), the writer claims that randomized control trials were not used to measure progress. The writer also stated that CREDO employed “statistical gymnastics to compare student achievement in charter schools across state lines while adjusting data to ensure that all students start at the same level” (Center for Education Reform…, 2013). Such factors would threaten the external validity of the study.

The value of sound educational research cannot be understated. Equally important is being an informed consumer of research so that decision making and problem solving can be based on the reliable and accurate interpretation of data. It is true that data shown through statistics can be intentionally misleading. As Sprinthall states (2012), “To the uninitiated, liars can figure plausibly.” I am thankful for the tools and knowledge I received in EDU 6975 and 6976 that have enabled and inspired me to be initiated.

I am including a link to a description of a research hypothesis and outline I prepared here:  Instructional Grouping Strategies that Affect Reading Fluency in Kindergarten  that includes notes regarding the subject relevancy, related literature references, sampling procedures, a description of variables and measures, possible threats to validity with plans for correction, and operational definitions.

Resources:

Center for Education Reform ‘Skeptical’ of Charter Study. (2013, August 5). In Education Week. Retrieved August 23, 2013, from http://www.edweek.org/ew/articles/2013/08/07/37letter-3.h32.html?qs=research+validity+in+education

McMillan, J. H. (2012). Educational Research: Fundamentals for the Consumer (6th ed.).             Boston, MA: Pearson.

Pillars, W. (2012, September 5). What Teachers Need to Know About Action Research. In Education Week. Retrieved August 23, 2013, from http://www.edweek.org/tm/articles/2012/09/05/tln_pillars_actionresearch.html?qs=action+research

Sprinthall, R. C. (2012). Statistical Analysis (9th ed.). Boston, MA: Pearson.

Advertisements
 

Linear Regression

Filed under: Standard 11. Inquiry/Research — lkgale @ 1:45 am
Tags:

Module 8 Reflection

During this module I learned about linear regression and how it allows us to predict a value of y based on a given value of x if there is a significant correlation between x and y. I also learned that the regression line helps us determine how much y changes as a result of x changing as well as helps to determine the “intercept” or the value of y when x is 0. I learned that the confidence interval refers to the range of data in which there is a 95% chance that the value for Y can be found.

Linear regression relates to previous content in that it is another way to analyze data. In this case cause and effect cannot be determined but analysts can make predictions about how one variable will change in a population when another variable has a significant correlation with it.

I am finding that at my first time through in the readings and the lecture I tend to be pretty lost, but I persist anyway. After reading the clear and concise discussion posts of classmates and answering the discussion questions presented by Dr. M, I am able to straighten most of it out.

 

Sprinthall, R. C. (2012). Statistical Analysis (9th ed.). Boston, MA: Pearson.

 

Reflection on p value

Filed under: Standard 11. Inquiry/Research — lkgale @ 1:38 am
Tags:

Module 3 Reflection

Module 3 reinforced to me that if a study did not use random selection or random assignment there cannot be a cause and effect relationship to the variables. I also learned that the p value of a study that assumes a null hypothesis indicates what percentage chance there is of drawing a random sample with a given difference in value between the control group and the group that had the treatment. The p value relates directly to how strongly the evidence either supports or rejects the null hypothesis. Further, I realize that if a p value does not support the null hypothesis, it only means that the results obtained are not likely to be due to chance. It does not necessarily mean that the alternate hypothesis is supported. Therefore, both studies that we reviewed in this module only showed that the results were probably not due to chance. They did not show that the independent variable actually caused the difference.

Previous content introduced and reviewed topics such as the null hypothesis, randomization in research, and characteristics of the normal curve. This module showed us how these concepts can be used to evaluate research studies.

This module pulled several things together for me and gave me quite a few “aha” moments, especially regarding the significance of p value.

Sprinthall, R. C. (2012). Statistical Analysis (9th ed.). Boston, MA: Pearson.