The Research Proposal - Advanced Research Methods | FYC 6802, Papers of Introduction to Sociology

Download The Research Proposal - Advanced Research Methods | FYC 6802 and more Papers Introduction to Sociology in PDF only on Docsity! Organizing the Thesis Mickie Swisher Associate Professor Family, Youth & Community Sciences There are many guides to writing the thesis. This one is based on my own experience. It focuses on organizing the thesis and highlights the problems that my students have encountered over the years. You should consult closely with your supervisory committee before you start your thesis work. They have the final say. However, I think you may find these suggestions helpful. I give some suggested page lengths for the different chapters in the thesis. These are rough approximations, based on a 100-page thesis. These numbers are designed to give you an ideas of the proportion of your thesis that each chapter normally comprises. Your own thesis may vary. For example, if your methodology is complicated it may require more space and/or a greater proportion of the thesis than I suggest here. The Research Proposal Chapters I, II and III of the thesis comprise the research proposal. I prefer that my students actually complete these chapters prior to conducting the research. Many faculty members do not feel this way and often a 10 to 15 page proposal is sufficient. That is, the proposal may consist of a shortened or abbrevi- ated version of Chapters I, II and III. You can reduce the introduction in the proposal. It must, however, contain a justification and a clear statement of the research question. You can shorten the literature review by focusing on the research literature that played a pivotal role in the development of your hypoth- eses. Notice, however, that you still have to read all of the relevant literature before you prepare the proposal. Otherwise, how will you define your research question and hypotheses? Later, in the thesis, you will need to expand the literature review if you decide to prepare a shortened proposal. There is no room for shortening the methodology chapter. This is your “recipe” for conducting the research and failure to explicitly state what you are going to do will almost undoubtedly result in problems in completing the research in a timely and satisfactory fashion. Chapter I. Introduction The introduction to the thesis justifies your research. It explains why the research is important and valu- able. The introduction puts your research in context. A normal length would be about five to ten pages. You should include four key elements in the introduction. 1. Identify the general or over-arching problem that you address in your research. First and fore- most, select a problem that interests you. You will live with it for months. Document the importance of the problem. The documentation should include quantitative or qualitative evidence that the problem is impor- tant, such as the number of people affected, the severity of the problem for those who are affected, or the economic or social impacts caused by the problem. Government reports and data bases like the census are excellent sources of documentation of the importance and implications of the problem. Examples of problems • Inadequacy of elders as surrogate parents for children in single-parent households • Decreasing involvement in governance by the population in rapidly growing communities Examples of evidence • Number of single-parent households over the past 30 years in US and/or Florida communities; num- ber of single-parent households where elders play a major role in parenting Organizing the Thesis - 2 • Voter turnout; number of volunteer hours per thousand of the population; number of local civic orga- nizations supported by volunteers 2. Narrow the focus of the research and explicitly state your research question. Rarely will your research directly address the entirety of the broad, over-arching problem you have identified. Narrow the focus to something manageable. Provide a succinct statement of your specific research question. Explain how the question that you want to answer fits into or relates to the larger problem that you have identified. Explain how your research will contribute to solving the problem you have identified. The research ques- tion should be posed in terms of the totality of the problem you have identified, but normally will focus on one specific aspect of the problem. Examples of narrowing the focus • Adequacy of parenting by male elders in single-parent households headed by females raising male children • Citizen participation in community based civic organizations Examples of relationship to the larger problem • Percentage of all single-parent households that are headed by women raising male children; evi- dence that male children in these households are more apt to commit violent acts, take drugs, acquire AIDS, be arrested, etc.; discussion of reasons why male elders may not provide adequate parenting for male children in these households • Historical data showing declining rates of participation in local civic organizations; data showing the relationship between participation in civic organizations and savings in public expenditures, improved quality of life in the community, etc. Examples of research questions • Examine the degree to which male elders have the knowledge, experience, and parenting skills needed to help prevent male children between the ages of 15 and 19 in female headed single-parent households from experimenting with drug use • Evaluate the relative importance of six selected factors in reducing the participation of recent immi- grants in civic organizations Examples of contribution to solving the problem • Provide suggestions for improving the ability of male elders to reduce risk-taking and anti-social behavior by teenage males in female-headed single-parent households • Suggest mechanisms for increasing the participation of recent immigrants to Florida in community governance 3. Define any terms that you use that do not have well established definitions and/or delimit the terms you use. Some terms have very clear, universally accepted definitions. Many do not. If you use any terms that do not have a generally accepted definition, it is your responsibility to explain exactly what you mean. In many cases, you will want to delimit a term – set explicit limits on how you will use the term that are at variance with a more normally accepted definition. Examples of definitions • Male elder – this could mean a grandfather, a great-grandfather, an uncle, etc; you might want to limit this to a male grandparent for the purposes of your study • Participation in civic organization – this could mean anything from occasionally coming to a public event to holding elected office; you might define this as attending at least 50% of the meetings and Organizing the Thesis - 5 men. In short, you have to develop a theory based, testable hypothesis that you believe (based on other research and the literature) could really be wrong. One-tailed hypotheses (ones that predict a direction of change or that specific the relationship such as “positively correlated with”) are superior to two-tailed hypotheses (hypotheses that just say there will be a difference without specifying greater, positive, nega- tive, etc.). The best hypotheses, if disproved, will strike a serious blow at the theory on which they are based. Einstein once predicted that light rays would “bend” when they passed near the sun – something totally outrageous, never heard of, etc. He was very honest. This was a logical, inevitable prediction that flowed from his theory of relativity. So, no matter how ridiculous it seemed, it was a good test of the theory. Had the hypothesis been disproved, Einstein’s “famous” theory of relativity would be Einstein’s “never heard of again” theory of relativity because it would have been bashed! That was a good hypothesis. Examples of hypotheses • The relationship between knowledge about the kinds of drugs commonly offered to youth, prevalence of their use by youth, and risks associated with their use and an elder male’s ability to help young men he parents avoid drug use will be positive (e.g., more knowledge about the problem – drugs – equals better success avoiding drug use). • The relationship between life experiences with drug use and an elder male’s ability to help young men he parents avoid drug use will be positive (e.g., direct experience with drug use by the elder equals better success at avoiding drug use) • The relationship between parenting skill and an elder male’s ability to help young men he parents avoid drug use will be positive (e.g., greater parenting skills equals better success at avoiding drug use). • Each of these three components (knowledge about drugs, experience with drug use and parenting skills) will contribute equally to success in helping young men avoid experimenting with drug use. Chapter III. Methodology The overall purpose of the methodology chapter is to provide enough information for another researcher to repeat your study. It is really that simple. You can normally do this in about 10 pages in a 100-page thesis. There are several key questions and issues. 1. Describe the research design. There are three broad groups of research designs: (1) experimental and quasi-experimental, (2) inferential and (3) case studies. There are requisites for each. If you do not understand the difference between these, you need to take a research design course (we offer one) or read about the three types of designs. They are all equally valid and useful. None is “better” or “stronger” than the other. The nature of your research question determines which type of design you will use. If you will manipulate your test subjects you will use an experimental design. If they have already experienced something that you think may differentiate them (exposure to parenting training, for example) you will probably use a quasi-experimental design. If you want to reach conclusions about one or more popula- tions as a whole (newcomers to Florida versus people born here) you will probably use an inferential design. If you want to explore many aspects of a phenomena – many characteristics of a group of people and how those characteristics interact – you will probably use a case study. At any rate, you need to decide and state which type of design you will use. Then you need to be specific. How will you define the population? What specific experimental design (such as two group, no control), will you use? If you are conducting a case study you will need to explain how you selected the case (because it was easy is not the answer) and describe the case (a group, an organization, a community) in terms of its salient features. 2. Explain how you will collect your data. This usually includes three components. Sample selection. In the social (and biological) sciences, we almost always have to select a sample. The sample must be representative of the group (or population) as a whole for experimental and quasi-experi- Organizing the Thesis - 6 mental and inferential designs. The sample does not have to be representative for case studies, but in many situations even in case studies you will need to select a representative sample. You need to explain how you will select your sample. This involves two separate steps. First, define the sampling frame. Usually we do not really have access to the entire population that we want to study – there is no list of the homeless in Florida, for example. So what will you use – people who eat at soup kitchens, people who sleep at homeless shelters? Second, once you have your equivalent of a list – the sampling frame – how will you select the individual sampling units – people, families, communities, groups or whatever? Instrumentation. The second component is your instrumentation. In the social sciences this is often a survey instrument, some sort of test of knowledge or attitude or some sort of outcome test. You must describe the instruments you will use. If you are using an instrument that has already been described in the literature, give the citation and move on. The next researcher can consult the original author. If you are altering a standard instrument, explain the changes you are making and why. If you are developing your own instrument, explain in detail how you arrived at each component. If there are several instruments available (say different measures of leadership), you must indicate which one you will use and justify your selection. Somewhere, usually the appendix, you need to provide a copy of the instruments themselves. Procedure. Finally, you will usually have to explain the data collection procedure. Will you conduct per- sonal interviews or use a mail-out or telephone contact? When will you conduct the interviews? 3. Explain how you will analyze the data. This is straightforward. Simply explain each analytic proce- dure that you will use, statistical or other. Define the independent and dependent variables. Remember that the objective here is to provide enough detail for another researcher to replicate your work. Do not provide unneeded detail. On the other hand, make sure that someone else really could replicate your work – that they would not have to guess about anything you did. Use citations and refer- ences. Often, M.S. these appear with almost no references in the methodology section. This is not ac- ceptable in most cases. You can, for example, say something like: “I will use a two-group, no control design (Leedy and Ormrod 1998). Group one will include people who have lived in Florida five or fewer years. Group two will include people who have lived in Florida for twenty or more years.” Notice that by citing Leedy and Ormrod you do not have to explain what “two group, no control design” means. You could also say: “I will use the Likert-like scale of personal esteem developed by Watson (1994) as modified by Jansen and others (1996). I selected the modified version because previous researchers (Jansen et al. 1996; Robbins 1998, Welch 2000) found that the internal reliability of the modified version is greater than that of the original version used by Watson (1994). I have included a copy of this instrument in Appendix B.” You have provided the references and justified your selection. 4. Explain any limitations in your study and problems that you encountered. Limitations include everything from very mundane occurrences to major problems. For example, let’s say that you decided to use the tax rolls in County X as your sampling frame for landowners. You conducted personal interviews. The tax rolls included several absentee land owners. Obviously, you could not go to New York, Boston, etc. to interview these people. In this case you would have to state a limitation in your thesis something like the following. “I could not interview the absentee landowners listed on the tax rolls. My sample is not representative of this portion of the population. My findings about the attitudes of taxpayers about the quality of life in X County may or may not apply to absentee landowners.” Your proposal should contain a limitations section in the methodology, too. This is where you list foreseeable limitations – such as the tax roll example. Almost every researcher also encounters unforeseen problems. For example, perhaps you had a high non-respondent rate – people who would not let you interview them. You should indicate that this occurred. It does not mean that your data is worthless, your effort wasted, or any other disaster. However, you are being less than honest if you do not explain the problem. (During my doctoral research Organizing the Thesis - 7 the same pig ate my corn plants on one farm three times! Needless to say, given that there is an optimal date for planing corn, the corn plants on this farm never did produce much corn.) You should offer expla- nations for the problem (I could not guard the field 24 hours a day), explain what you did about it (tried to buy the pig), explain the implications for your results (unrepresentative corn yields in all treatments on that farm which probably reduced the difference between treatments), and offer suggestions about how other researchers could avoid the problem (stay away from farms with pigs whose owners commonly spend all weekend in the local cantina, thereby guaranteeing that the pig is looking for food). 4. Organization. There are many ways to organize the methodology chapter. If your methodology is easy to explain, you may present it as one discussion. In many cases you will have multiple hypotheses. In this case you may want to explain the instrumentation and analyses for each hypothesis. I prefer this ap- proach. At any rate, you must clearly indicate how you will use each piece of data that you collect and what data you will use to test each hypothesis – quantitative or qualitative. I prefer the simple approach. First, explain the independent variables. “My independent variables include demographic data. These data include age, race, place of resi- dence, annual income by category, .... I used logistic regression analysis (Evans, 1992) to determine the degree to which these demographic characteristics explain differences in success in parenting.“ ”My first hypothesis is that knowledge about drugs contributes to successful parenting by the male elder. I developed a test instrument to evaluate the subject’s knowledge about drugs, which I treated as an independent variable used to predict successful parenting. The instrument measured knowl- edge in three areas: (1) the kinds of drugs commonly offered to youth by drug dealers or friends, (2) the prevalence of usage of four drugs among youth in Florida, and (3) the health risks associated with the use of these four drugs.” “To test knowledge about the kinds of drugs that young people encounter, I asked each respondent elder to name the four drugs that he thought were most likely to be offered to a young person attend- ing the local high school. I scored the individual’s response by comparing the list he provided with a list of the four drugs that are most prevalent in high schools in Florida. These are alcohol, marijuana, Ecstacy and amphetamines or speed. I awarded one point for each of the four drugs that the respon- dent correctly identified. No points were subtracted for incorrect responses.” “I also asked each respondent to pick the four drugs that he believes are the most commonly used by Florida high school students from a list of ten drugs. The list included the four drugs mentioned previously plus barbiturates, cocaine, crack cocaine, heroin, opium and laughing gas. The four cor- rect answers were distributed throughout the list. I awarded one point for each of the four drugs that the respondent identified correctly. No points were subtracted for incorrect responses.” “Finally, I asked the respondent to select the health risks associated with the use of each of the four drugs most commonly used by Florida high school students. I provided a list of six possible health risks for each drug. The lists are provided in Appendix A in the complete instrument. In each case, I included some correct and some incorrect risks. I varied the number of correct answers to prevent respondents from identifying a pattern of correct number of responses. I awarded one point for each correctly identified risk and did not deduct points for incorrect answers. The correct responses are indicated by an asterisk in the example in the Appendix.” “I used Worthen’s (2000) instruments as a model for my own instruments. Worthen developed and tested his instruments in a study of the contribution of knowledge, experience and skills in successful outcomes in drug counseling. My instruments are based on those he used to evaluate the knowledge, Organizing the Thesis - 10 3. The case of “insignificant” results. First, if your results do not corroborate your hypotheses, think hard about your methodology. Could something you did have caused the result? Can you identify any weakness in your instruments? In your sample selection? Hopefully, the answers to these questions will be “no” because then you have a very exciting situation. The most exciting theses are those where your data (assuming you did a good job on methodology and analysis), do not corroborate your hypothesis. This is exciting because it implies that something is going on that we, as a community of researchers, have not thought about. Let’s go back to our elder parenting and drug use example. You have built a good strong, logical, well-based argument for why the elder’s knowledge, experience and parenting skills should have a strong outcome on experimentation with drugs. But .. your results end up negative! There’s just no relationship there. Now that has some strong implica- tions. What in the world is going on here? If these are not the predictors – what are they? What have we missed? What is wrong with all of this knowledge that we have accumulated? After all, you have a pile of research results that you covered in your literature review that show that these things ought to predict success. But remember, everyone else looked at either just one of these predictors or they did not look at male elder mentors for male youth in female single-parent families. Aha! Is there something different about the elder male to young male relationship that is important? Or are we looking at the wrong predic- tors? Or should we develop a new model that incorporates more factors into explaining success as we have defined it? Negative results are not insignificant. They are exciting. If you did a good job on the research and had good hypotheses (well based in the current knowledge, cutting edge), you are opening up whole new arenas of research. Tell us about them. Chapter VI. Conclusions This is a short chapter. Often ten pages is enough. It is also a fun chapter. The conclusion is not a summary of the results and discussion. This is a very common mistake. Do not, repeat not repeat what we already know. 1. Go back to the big problem you identified in the introduction. First, tell us what your findings and your interpretation of them (the discussion) mean for the problem you were trying to solve in the first place. What are the implications of your research for the “real world.” Should we be giving elders parenting training? Should we have elders out there experimenting with drugs just so they’ll know what it’s all about (well, okay, maybe not)? Should we provide young males in female-headed households with “drug coun- selor” types – men nearer their own age who have more knowledge and experience with drug abuse? Go back to that problem and explain what your research suggests we should do about it. 2. Future directions for research. You answered some questions and filled some of the gaps in our knowledge. What questions should the next researcher address? Why? How should they go about it? Did you uncover any new questions or mysteries? What are they? How could other researchers attack those questions? 3. Go ahead – get wild. Up until now you have had to stick with published research findings and your data — no going out on a limb, no speculation, no dreaming about what this all possibly, just maybe could mean. This is where you get to do all that. No dream is too big for the conclusions chapter. You do not have to stick with carefully constructed, totally reasoning, well-founded in my findings ideas here. You can speculate. Go ahead – give us the really big picture. Hypothesize some more. Your research probably gave you some “inklings” of questions that should be asked, research that should be done. Maybe your results do not say enough for you to be willing to call it a “conclusion” or a “recommendation” or even a “next step in the research.” But you probably do have some idea now of a much bigger picture, maybe a whole radical new approach to tackling this problem, a very “wild” set of questions that should be asked. Organizing the Thesis - 11 Well, go ahead. Do it here. End with a bang, not a whimper. Albert Einstein is one of my favorite role models. He had the audacity (as a 28-year old patent clerk who could not get a university job, no less) to propose that, contrary to common experience with everything else in the world, the speed of light is constant and that nothing ever, not ever, can exceed or even equal the speed of light. Hey – what a wild and crazy idea! Name one other thing in the whole universe, one thing that you have ever seen, that has a constant speed; one thing that you couldn’t catch if you just go fast enough. Couldn’t be right – crazy man. Well, of course, he was right and he (more or less) proved it. So fear not in the conclusions. Propose those ideas that seem “just maybe, just barely” possible.