Sunday, January 10, 2010

Discussion of Yin's Ch. 5 2010

Explain the five analytic techniques and how you can "press for a high-quality analysis." Connect this to your own case study plan/idea too.


  1. The five anyaltic techniques outlined by Yin to enable the researcher to draw conclusions from evidence include pattern matching, explanation building, time-series analysis, logic models, and cross-case synthesis. Pattern matching compares a pattern which has been established by experience with a predicted pattern. The goal is to see if the two patterns match. If matching occurs, the internal validlity of the case study is stronger. Explanation building, the second analytic method, is a special type of pattern matching. This involves analyzing case study data by building an explanation about the case. Yin cautions that explanation building is extremely difficult, and requires a great deal of analytic insight on the part of the researcher. Conducting a time-series analysis is another analytic techique used to analyze data. In this technique, the researcher matches the observed trend using either a theoretically significant trend or a rival trend. The more intricate and precise the pattern, the more the time-study analysis will support the conclusions of the study. The important thing is to examine the how and why questions about the connection of events over time, not just to chronicle events. Logic models, an additional form of analysis, establishes events over an extended time period, and sets up a cause and effect relationship pattern, whereby an event at an earlier time causes another event to occur.Yin suggests establishing a logic model before collecting data and then testing the model by seeing how well the data supports it. Cross-case synthesis, Yin's final technique, applies to the analysis of two or more cases. Each indivual case in the cross-case synthesis is treated as a separate case, but synthesis of the data of each case strengthens your own case study data. Pressing for a high-quality analysis includes four principles. First, you should attend to all evidence. Yin uses the word "exhaustive" when covering evidence applying to key research questions. Second, all major rival interpretations must be addressed. Examining rivals is the best way to strengthen your own research. Third, analysis should address the most significant aspect of your case study. Otherwords, focus on the most important issue. Finally, you should use your own prior knowledge in your study. Know your study inside and out and show awareness of current thinking and discussion concerning your topic.

  2. The five analytic techniques are pattern-matching, explanation building, time-series analysis, logic methods, and cross-case synthesis. Pattern matching logic can be considered a high-quality analysis because it not only can compare researched based patterns with predictable outcomes along with patterns associated with alternative pattern predications. If patterns show a relation results can strength internal validity. When looking at my own research, when analyzing survey results I can use pattern matching to find similar answers between students and teachers and how they feel about PBS. Explanation building is a high-quality analysis for explanatory case studies. Through this analysis, researches follow a process to provide details. Yin explains that when using explanation building, researches makes an initial theoretical statement or initial proposition about policy or behavior, compare the findings of the initial case, revise the statement or proposition, compare other details of case against revision, comparing revision to the faces of multiple cases, and will repeat the process as many times as needed. When looking at my research I could use explanation building after conducting my surveys, questionnaires, and focus group interviews. After collecting all this data, I may be able to explain “how” and “why” PBS has made a positive or negative effect on the school environment in my study. A time-series analysis can be considered high-quality because it allows researchers to answer “how” and “why” and patterns over time. By finding results that consist of a pattern or unexpected results, researches can determine if results are legit. In my research, I could use time-series analysis by seeing how students and teachers feelings about PBS become more positive or negative with the implementation of PBS can completing the multiple steps to the program. Logic models are similar to patterning. Logic models allow the researcher to observe events to theoretically predicted events. A researcher should first define to logic model before collecting data and then collect data to see if logic model is accurate. Using logic model in my research would be explaining the purpose of my research and what I would like to find out and then collecting my data to see if my prediction is correct. I hope this is correct. The last analysis is cross-case synthesis focuses mostly on multiple case studies. This is where an analysis is done consisting of two case studies. This is where each case study can be performed as an individual study. I again, was confused with this analysis. When looking at my research, I could each method of data collection as a different research report. I could look at the survey, questionnaire, and focus groups as different “case studies” when looking at my results.

    Question: Yin explains that “none of the analytic techniques should be considered easy to use, and all will need much practice to be used powerfully.” Page 126. Why is he determined to scare all of using a case study method of research? Creswell was more encouraging then Yin has been. It seems that every chapter he is explaining how time consuming and hard a case study is to complete.

    Fact: To find analytic conclusions in your case study research there are four strategies; relying on theoretical propositions, developing a case description, using both qualitative and quantitative data, and explaining rival explanations. Page 131-133
    Quote: Unlike statistical analysis, there are few fixed formulas or cookbook recipes to guide the novice. Instead, much depends on an investigator’s own style of rigorous empirical thinking, along with the sufficient presentation of evidence and careful consideration of alternative interpretations. Page 127

  3. Even though Yin cites using any oa all analytic techniques when examining data, I am not sure if all of the five would apply to by research. Explanation building is the method of data analysis which might be the best fit for my research question. My research asks the question, "Do negative educator attitudes toward children of poverty promote lowered expectations for student achievement?" In other words, are educator attitudes regarding poverty a barrier to student achievement? I am attempting to explain if (in at least one rural district in Missouri), educator attitudes regarding poverty have a direct relation to instruction of children of poverty. I am making an initial proposition about social behavior and attitudes. It is my opinion, based upon sixteen years in a rural elementary classroom, that educators set lowered academic expectations for students of poverty. Public education is the opportunity for every individual, regardless of economic status, to achieve according to their abilities. Educators must understand the critical nature of their role in student achievement. As I conduct this research, hopefully the gradual building of an explanation will occur. It will be necessary to conduct an extremely thorough literature review, and to examine rival explanations as they occur throughout the project. Yin's warnings about explanation building as an analytic technique are noted. I'm certain my anylytic insights as a researcher are feeble, but as an educator, I feel this is one of my strengths.

  4. Question: How often does a researcher need to pause and analyze data?

    Fact: Page 130: Yin states to be aware of your analysis strategies and choices before collecting data; so data will be analyzable.

    Quote: Page 136: "None of the analytic techniques should be considered easy to use, and all will need much practice to be used powerfully."

  5. Christina - I appreciate your question as that was my quote. My thought after reading that was well crap, how motivating....NOT!

  6. Christina - In response to your fact. I am glad Yin has provided us with these for strategies. However after reading this chapter I am thinking, what the heck. Seems hard for the novice researcher...I think I will have to read this chapter again to fully understand the strategies.

  7. Christina - In response to your quote. I am glad we as novice researchers are provided with guidelines and guidance. However, if there are so many strategies we are to incorporate, I guess I am confused as to why there is any room for alternative interpretations.

  8. The five analytic techniques Yin shares in chapter 5 are as follows: Pattern Matching, Explanation Building, Time-Series Analysis, Logic Models and Cross-Case Synthesis.
    1.Pattern Matching: This technique is the most desirable of the five analytic techniques. Pattern mating is used when there is a need to compare logic in a empirical based pattern with a predicted one. Also pattern matching coincides results and can help a case study to strengthen its internal validity. If a case study is exploratory then patterns may be related to the independent or dependent variables or even both. This technique is also used if a case study is descriptive and relevant as long as the predicted pattern of specific variables are define prior to the collection of data.
    For my research: This technique could be used when I compare notes from teacher and student conferences. The conferences with the topic of how can direct instruction of the basics of reading and writing be beneficial.
    2.Explanation Building: This technique is a special type of pattern matching. The procedure however, is much more difficult and deserves undivided attention. The goal of this technique is to analyze the case study data by building an explanation about the case. Yin explains that the elements of the explanation are to explain is to stipulate a presumed set of casual links about the case study, or the "how" or "why" something happened. He also mentions within this technique it is important to make an initial theoretical statement or initial proposition about policies or social behavior. He also mentions that it is important to compare findings of an initial case against such a statement or proposition. Next researcher is to revise statement or proposition, compare details of case against the revision and to compare the revision to the facts of a second, third or more causes. Repeating this process may be necessary.
    For my research: I could get the positive as well as the negative feelings students and teachers have towards the use of direct instruction when teaching the basics of reading and writing.
    3. Time-Series Analysis: For this technique there may be only one independent variable or dependent variable. There is a large number of data points that are relevant and available as well as statistical tests can be used to analyze data. There are appropriate start and end points for single variables may not be clear. Yin also discusses a more complex time series. This technique is used when trends within the case are more complex. The example he uses is that when there is not just a rising or declining (or flat) trend, but some rise followed by some followed by some decline within the same case.
    For my research: This technique would be used to analyze test, checkpoint and homework assignments.

  9. 4. Logic Models: This technique is useful in doing case study evaluations. The Logic Model deliberately stipulates a complex chain of events over an extended period of time. Such as when events are staged in repeated cause-effect-cause-effect patterns. More specifically a dependent variable (event) may be in an early stage and later turn into the case study independent variable. This technique can be used in a variety of circumstances.
    For my research: I could use this in analyzing tests, assignments and checkpoints. To to understand why students performed a certain way. Was it the type of question presented? What kind of day was the student having? What time of day it was.
    5. Cross-Case Synthesis: This technique applies specifically to the analysis of multiple cases. It is relevant if the case study consists of at least two cases. The examination of word tables for cross-case patterns will rely strongly on argumentative interpretation.
    For my research: I am a little fuzzy on how this would have a place in my research project. However, if I chose to continue my research with next my class next year. Would this technique be used since I would be comparing two different classes of students?

  10. Quote: (Page 130) (referring to analytical techniques) "These strategies or techniques are not mutually exclusive. You can use any number of them in combination. A continued alert is to be aware of these choices before collecting your data, so that you can be sure your data wil be analyzable."

    Fact: (Page 129) However, even under the best circumstances, nearly all scholars express strong caveats about any use of computer-assisted tools: You must be prepared to be the main analyst and to direct the tools; they are the assistant, not you.

    Question: In the Abstract to Chapter 5, Yin says a researcher must produce "high-quality analyses, which require attending to all the evidence collected" (Page 126). How does a researcher ever feel he or she as collected enough evidence? How would you know when enough is enough?

  11. Yin lists the five analytic techniques as pattern matching, explanation building, time-series analysis, logic models and cross-case synthesis. Yin writes that pattern matching is one of the most desirable techniques because it compares an empirically based pattern with one or more predicted patterns. Internal validity of the case study is strengthened if these patterns coincide. Explanation building is a type of pattern matching but Yin gives it a separate distinction because it is more difficult. Explanation building is what its name implies, it involves analyzing study data to build an explanation about the case. The basic logic in a time-series design is matching observed trends and either a theoretical trend or a rival trend occuring before the investigation. Logic models are a chain of cause-and-effect patterns in which a dependent variable becomes an independent variable in the next stage. The first four techniques can be used with either single- or multiple-case studies, but the fifth technique, cross-case synthesis is specifically used in the analysis of multiple cases. The researcher can use information from previous case studies to compare with their study data or actually conduct more than one study in their own case study research.

    For my research project, designed to see if self-recorded rereading of text improves reading fluency and comprehension scores, I plan to use a quantitative research method. I understand that my research findings could be better enhanced if I included possible "why" data from doing case study research. However, because of time restraints and shear fear from reading Yin's book, I admit I'd rather stick to my project as planned. If I were to choose one of Yin's techniques that best fit with my research, it would seem to be pattern matching because in essence I theorize my intervention will work and predict that the data will support that theory.

    Question: Previously, when I read the phrase "manipulating data", I thought the author was talking about changing data. However, after reading this chapter, I think it actually means putting the data in different formats to look for trends. Am I on the right track now with the meaning?

    Quote: (pg 133) "... qualitative data may be critical in explaining or otherwise testing your case study's key propositions."

    Fact: My fact goes along with my question. It seems the more I learn the more I realize how little I know!

    Barb, I have the same question as to how a researcher knows when they've hit that magical "enough" level of evidence.

    Ann D., in response to your quote, when Yin writes that all of the analytical techniques will need much practice to be used "powerfully" I find myself wondering exactly what constitutes powerfully?

    Christina, in response to the quote you chose, I sure wish there was a fixed formula for us novices!

  12. Christina, Yin's comment about using the analytical techniques "powerfully" is true, and I do like his choice of words, it's just that we don't know how to do that yet. Hopefully, we'll learn!

    ADotson, it appears that "pausing and analyzing data" must be a continual process during data analysis-if we use the five techniques, that's all we would be doing, isn't it? Do researchers have a life?

    Ann, your quote about qualitative data being critical in explaining your study's key proposition seems right to me. I plan on using a mixed method which is mostly quantitative, but that ends with qualitative questions that will explore the issue in greater detail.

  13. The five analytical techniques are:
    1. Pattern matching- This is the most desirable, if a pattern arises after comparing an empirically based pattern with a predicted one the cases internal validity will be strengthened.
    2. Explanation building- This is a more difficult, special type of matching. This is completed in narrative form to explain a phenomenon.
    3. Time-series analysis- This follows intricate patterns and can lead to a firm foundation at the conclusion of the research.
    4. Logic models- This useful method specifies a complex chain of events over an extended period of time.
    5. Cross-case analysis- This specifically analyzes multiple cases mentioning more than 2 cases could strengthen your findings.

    Pressing for higher quality analysis- 4 principles underlie all good social research:
    1. (Attend to all the evidence) While following protocol during my research I need to make sure I look closely at the pre and post test data. Using my survey and information gained from it will also be important when answer my research question- How does technology-enhanced classroom instruction compare with traditional classroom instruction in student academic achievement, student motivation to learn, and student attitude towards technology?

    2. (Address all major rival interpretations) Be aware if someone else has a different interpretation of the research and see if the have anyway way of supporting it, if they do I then need to investigate it later.

    3. (Address the most significant aspect of your case study) I need to stay focused on the most important part of my study and not be side tracked by other, less vital information that may or may not arise.

    4. (Use your own prior expert knowledge) I know technology can be significant in the classroom…now to have the proof!

    Question-I ask again, “Am I doing a case study?” I may be, but need confidence in knowing the answer!!! How does technology-enhanced classroom instruction compare with traditional classroom instruction in student academic achievement, student motivation to learn, and student attitude towards technology?

    Quote Pg 162- “In the absence of such strategies, you may have to “play with the data” in a preliminary sense, as a prelude to developing a systematic sense of what is worth analyzing and how it should be analyzed.”

    Fact: Research is not easy, take it one step at a time and do not begin too difficult.

  14. Chris and Ann in response to the quote Chris mentioned about few fixed formulas or cookbook recipes to guide the novice--sometimes have a recipe is a good thing!!!! It's comfortable and although our own flair is exciting and different, how do we know it's right?

  15. Barb in response to your question about how would you know when enough is enough when collecting evidence? I think back to last semester and the word "saturation" comes to mind. Could we use this some idea when collecting evidence?

  16. Quote: (p.128) "...developing a rich and full explanation or even a good description of your case, in response to your initial "how" or "why" questions, will require much post-computer thinking and analysis on your part."

    Question: How is "developing a case description" an analytic technique? To me, developing a case description is like telling a story and not analyzing data.

    Fact: Yin describes four general analytic strategies for case study analysis: relying on theoretical propositions, case descriptions, a dual use of quantitative and qualitative data, and rival explanations.

  17. Corey, I like your fact. I am all for things being "not too difficult."

    Barb, your fact highlights how Yin stresses the importance of researchers accepting the fact that they are the main analyst and not the computer software. How grand would it be if we could let computers do the main analysis? Perhaps some day.

  18. Quote: (160) “Your analysis should show how it sought to use as much evidence as was available, and your interpretations should account for all of this evidence and leave no loose ends. Without achieving this standard, your analysis may be vulnerable to alternative interpretations based on the evidence you had (inadvertently) left out.”
    Question: (151-152)I am asking myself…” How can I use the Individual Logic Model in my classroom with my students who are at-risk?” I really liked this model. (I have a behavior flowchart that I use with one of my more challenging students. I hadn’t thought of putting in the numbers for the opportunities for intervention. I think I’ll go back and add those into my flowchart! I am seeing opportunities for implementing some of these ideas into my teaching, not necessarily just in a research project…light bulb moment!)
    Fact: (141) Better case studies reflect theoretical propositions in their explanations. Causal links may lead to changes in public policy and change how and where public funds are distributed. Social science propositions may lead to theory building (as in the study of how farm societies in various countries became industrialized).

  19. Yin describes the use of five analytic techniques to ensure that case study research is of high quality.
    1. Pattern Matching Logic- this technique is used to compare empirically based patterns with predicted patterns. The matching of these patterns increases the case study’s internal validity; thereby increasing the quality of the study.
    2. Explanation Building- although this technique is a specific type of pattern matching, it is more difficult. The better explanations result in theoretically significant propositions that may result in public policy changes, or in the social science propositions, may lead to theory building.
    3. Time-Series Analysis- is used to track changes over time. The match between what is observed and either a theoretically significant trend prior to the study or a rival trend prior to the study.
    4. Logic Models-explains a complex chain of events over a period of time.
    5. Cross-Case Synthesis- unlike the other four methods, can be only used with a multiple cases study. This method attempts to aggregate findings across multiple case studies.
    In order to “press for high-quality analysis” Yin describes four main principles that a researcher must employ. First, attend to all the evidence. Second, address all major rival interpretations. Third, address the most important aspect of your case study. Finally, use your own prior, expert knowledge in your case study.

  20. Mrs. Dalrymple: Your Yin quote caught my eye, as well. It seems a bit daughting, doesn't it? It would be nice to be able to input data and out would come our perfect little research project!

    Ann Dotson: I had to chuckle when I read your quote again. What makes Yin think that we would find any of this "easy"? :)

  21. The five analytic techniques Yin discusses are pattern matching, explanation building, time-series analysis, logic models, and cross-case synthesis.

    Yin describes pattern matching as the most desirable technique to use for case study analysis. This technique involves the researcher in comparing an empirically based pattern with a predicted one. According to Yin, if the patterns coincide, a case study's internal validity is strengthen. With my research, I need to keep clear and precise records of the data from the pre- and post-tests, as well as the self-efficacy and attitudes toward science questionnaires in order to observe any patterns. I predict that the higher the self-efficacy and the attitudes toward science, the higher the student achievement in science.

    Explanation building is an analytic technique that Yin describes as a "special type of pattern matching." In most case studies, explanation building has occurred in narrative form. To explain a phenomenon, a researcher must tell the "how" or "why" something happened. Yin warns that explanation building is fraught with dangers. For instance, an investigator may slowly diverge from the original topic of interest. With my research, I could use explanation building by making an initial theoretical statement about the relationship between self-efficacy, attitudes toward science, and science achievement. I then could use my review of literature to compare findings of an initial case against my theoretical statement. I would then revise my statement and continue to compare with additional cases.

    A time-series analysis is the third technique Yin explains. Simply put, a time-series analysis monitors data collected over a period of time. Important to note that establishing the appropriate starting or ending points for a single variable may not be clear. Despite this fact, the ability to trace changes over time is a major strength of case studies. With my research, a time-series analysis may not be feasible sense I am collecting data over a relatively short period of time and will not have the same students I had for the pilot study compared to the final study.

    Yin describes logic models as becoming increasingly useful in recent years. Logic models involve the use of cause-effect-cause-effect patterns in which a dependent variable in an early stage becomes an independent variable for the next stage. The use of this analytic technique consists of matching empirically observed events to theoretically predicted events. With my research, a logic model would be appropriate if I chose to delve deeper. For now, I am looking at the effect of self-efficacy and attitudes toward science on science achievement. Upon completion of this research, I could choose to investigate how self-efficacy beliefs and attitudes toward science affect the educational performance of students in advanced science courses in high school and beyond. In this case, a logic model would be handy.

    Cross-case synthesis is an analytic technique that applies to the analysis of multiple case studies. This technique involves a researcher in examining data collected from multiple cases and not necessarily by the same investigators. This information is put in table form and analyzed. With my research, this method does not apply since I am essentially researching one group of students.

    To press for a high-quality analysis, Yin suggests four principles to uphold: 1) attend to all the evidence, 2) address all major rival interpretations, 3) address the most significant aspect of your case, and 4) use your own prior, expert knowledge in your case study.

  22. Corey, You go girl!!! You are my hero! I applaud you for taking the challenge of completing a case study. It seems way over my head at this time.

    Tanya and Dana, I agree that completing quantitative data is looking more appealing at this moment.

  23. Class,

    Remember there are pros and cons for each research method! Also, researchers like to argue about the weaknesses of these methods too. No method is perfect!

    Yin wants you to value case studies and take them seriously; he also wants them done the right way too (as real research).

    No worries!!!

  24. According to Yin exemplary case study reports include these five topics: 1) target the case study reports; 2) case study reports as part of larger, mixed methods studies; 3) illustrative structures for case study compositions; 4) procedures to be followed in doing a case study report; 5) in conclusion, speculations on the characteristics of an exemplary case study covering the design and content of the case.

  25. Question: Yin makes an interesting point that our case studies have the possibility of reaching and educating a wider variety of people (not just specialists on our topic) than a traditional research report. I am intrigued by this. With so much press and attention on autism spectrum disorders, I am wondering how to make my research more accessible so that it can help to educate people about the social benefits of the co-teaching model for children with autism spectrum disorders. My question is how can we gain wider audience for our research?
    Quote: 179 “From nearly the beginning of your investigation, certain sections of your report will always be draftable, and this drafting should proceed even before data collection and analysis have been completed.”
    Fact: In order to report the case study in the case study in the most effective way, the researcher should take into account the preferences of the audience and examine prior case study reports that were successful.

  26. Fact: pg. 134 “…the typical hypothesis in an evaluation is that the observed outcomes were the result of an intervention supported by public or foundation funds.”
    As a classroom teacher, I think we all take the liberty of knowing that an intervention method was helpful to a student without funding of the process. I found this interesting that the rival hypothesis is one that shows no funding.
    Quote:pg. 128 “, unlike statistical analyses, you cannot use the software’s outputs themselves as if they were the end of your analysis. Instead, you will need to study the outputs to determine whether any meaningful patterns are emerging.” With that in mind, I am happy to be sticking with quantative data only!
    Question: Dr. Hendrix, Do you think a research study is “easier” to analyze than a case study? Have you done both?

    The first analytical technique is pattern matching. Here, one is using a pattern matching logic. A researcher would compare an empirical based pattern with a predicted pattern. If the pattern coincides the results would be considered to have a higher internal validity. If the case study is an explanatory one the patterns may be related to the two variables, or if it is a descriptive one, the matching of the pattern is still relevant as long as the specific variable pattern is defined prior to data collection. The second technique is explanation building a form of pattern matching. Here the researcher is trying to analyze the data by building an explanation about the case. To explain a phenomenon is to stipulate a presumed set of casual links about it or how or why something happened. Explanation building is used not to conclude a study but to develop ideas for further study. In the third technique, time-series analysis, can be much easier to use because there may only be one single independent or dependent variable. The researcher conducts a time-series analysis in direct analogy to the time-series analysis conducted in experiments. A sort of one-on-one correspondence between data sets. This also helps solidify the validity of the research. The fourth technique is logic models a more recent useful analogy. This model deliberately stipulates a complex chain of events over an extended period of time. The events are staged in a repeated cause and effect pattern, where the a dependent variable becomes the independent variable. The last technique is called cross-case synthesis specifically used to analyze multiple cases. When a researcher is analyzing the results of two or more case studies at the same time, using this technique may make the analysis easier and more believable to the intended audience. I see myself using the time-series analysis because the research is set up based on one intervention method. However, I wonder if as the research gets started, that I might also be able to use logic models through my study. If the intervention strategy starts a cause and effect chain, then I might be able to use both analysis.
    There are four ways a researcher can press for a high quality analysis. The first is to show that you the researcher used as much evidence as one could find on your topic. It’s imperative not to have loose ends in your analysis of the data. Next, is for the researcher to address all rival interpretations of your study. The researcher must decide if a rival explanation should be addressed in the analysis, or used in a future project. Third, is to make sure that your analysis covers the main issue and doesn’t veer off into minor data results. In other words, if you are studying an intervention strategy, then the analysis should match that question, not other minor questions. To insure validity of your study, you need to stick to the pre-designed road map of data analysis. Last, a researcher should use their prior background knowledge in the case study. Keep current of the topic at hand.

  27. Ann G. I agree it is hard to keep up and track of all the info in both texts. I just keep thinking one day at a time and ask questions. It seems as if there is toooo much to learn in the amount of time that we have classes.
    Ann D. I see where Yin says we need lot of practice to use the analytical technique. Do we have a lot of time?